Redescription and new records of Jansen’s goatfish, Parupeneus jansenii (Mullidae), from the Western Pacific and Eastern Indian Ocean

For Jansen’s goatfish, Parupeneus jansenii (Mullidae), taxonomic knowledge has been rather poor and occurrence information restricted to a few localities only, with verified records from northern Sulawesi and Sunda Street (Indonesia), and Luzon (Philippines). This species is here included in the so-called “ heptacanthus ” group together with the Indo-Pacific cinnabar goatfish, P. heptacanthus , and five Indian Ocean congenerics, based on an evenly, symmetrically rounded posterior maxilla margin and similarities in meristic characters and preserved colour. In total 29 P. jansenii and 53 P. heptacanthus consisting of types, recently collected material from south-central Vietnam, southern Indonesia, and NE Australia, and geographically related reference material were studied along with single types of the other five heptacanthus -group species and P. angulatus , a morphologically similar Western Indian Ocean species. In total 62 quantitative meristic and morphometric characters including standard length (SL) were analysed after splitting the data into two size classes (small-sized fish, < 110 mm SL, vs. large-sized fish > 110 mm SL) to account for size-related differences due to allometry. Diagnoses for P. jansenii and P. heptacanthus and a redescription for P. jansenii were prepared. Fresh-fish colour descriptions for both large- and small-sized voucher specimens of P. jansenii are provided and the effects of freshness status on colour patterns is documented. New records for P. jansenii for Vietnam, southern Indonesia and NE Australia (= first verified record for Australia) are reported and depth information (24–100 m depth range) is given for the first time. P. jansenii differs from the co-occurring P. heptacanthus - independently of size - in shallower body and caudal peduncle, smaller maximum head depth, shorter barbels, caudal and pelvic fins, shallower first dorsal fin, shallower first ray of second dorsal fin, and lack of small dark red or reddish-brown spot on or just below the eighth lateral-line scale. The populations of P. jansenii from Vietnam and Australia differ from each other in a single character, the height of the eighth dorsal-fin ray, but both overlap with the intermediate population from Indonesia. This comprehensive alpha-taxonomic approach provides a useful tool for uncovering the diversity of Parupeneus goatfishes.


Introduction
The goatfish genus Parupeneus (Mullidae) is the second largest of the family Mullidae consisting of 32 valid species (Randall & Heemstra 2009) which are characterized by a dorsal fin with 8 spines, 14-18 pectoral-fin rays, 24-42 total gill rakers, bluntly conical teeth in a single row on both jaws, the snout longer than postorbital length in adults, and lack of dark-brown or black markings on fins. All Parupeneus species occur in the Indo-Pacific (Randall 2004). Recently, one species, P. forsskali (Fourmanoir & Guézé, 1976), has been also encountered in several areas of the southeastern Mediterranean as a Lessepsian migrant (Mehanna et al. 2016).
In a taxonomic review of the goatfishes of the Western Indian Ocean (WIO) 19 Parupeneus species were listed and an identification key with important diagnostic characters and overall distribution information for each species was provided by Randall & Heemstra (2009). For the other 13 species information on distribution and diagnostic characters was given by Randall (2004). The 32 Parupeneus species can be divided into three groups according to the form of the rear of the maxilla (Randall 2004;Randall & Heemstra 2009): one group of seven species, here termed the "heptacanthus" group, has an even, symmetrically rounded posterior margin; it consists of P. fraserorum Randall & King, 2009, P. heptacanthus (Lacepède, 1802, P. jansenii (Bleeker, 1856), P. minys Randall & Heemstra 2009, P. nansen Randall & Heemstra, 2009, P. procerigena Kim & Amaoka, 2001, and P. seychellensis (Smith & Smith, 1963; a single species, P. angulatus Randall & Heemstra, 2009, has a symmetrical but angular rear maxilla end; in the remaining 24 species the maxilla ends in a broad dorsoposterior extension. All seven heptacanthus-group species and P. angulatus overlap considerably in meristic characters and show uniform pale colour in preservative, but can be distinguished by body form characters and colour patterns in fresh condition (Randall & Heemstra 2009;Randall & King 2009). Three of these species were described and one species, P. seychellensis, was resurrected in the review by Randall & Heemstra (2009).
For the large Indo-Pacific area extending from the South China Sea to the Philippines and Indonesia and further south to Australia no recent taxonomic review of the genus Parupeneus exists. Among the above-listed species, the cinnabar goatfish P. heptacanthus and Jansen's goatfish P. jansenii occur in this area. Parupeneus heptacanthus is widely distributed from the WIO to southern Japan, the Caroline and Marshall Islands, New Guinea, Australia, New Caledonia, Lord Howe Island and Fiji. Its depth range is 15-88 m (Randall 2004). Parupeneus jansenii has been reported from Indonesia, Philippines, New Guinea, and northern Australia. Depth information for this species is completely lacking (Randall 2004).
Reference material of P. jansenii has been available only from Indonesia and Philippines (Randall 2004). The record of P. jansenii for New Guinea is that of Munro (1958), who listed Mansinam Island, New Britain, and Port Moresby as localities for this species. The record for northern Australia is derived from an annotated list of marine fish species of the Northern Territory by Larsen et al. (2013) who refer to a fish-species list by Wei et al. (1972) that was based on collected material. As Larsen et al. (2013) however state, no reference material is available for verification of that earlier record. Both Randall (2004) and Allen & Erdman (2012) emphasized that P. jansenii is a poorly known species with very few records, no recent sightings, and no habitat information. Furthermore, no photographs of fresh specimens have been available so far and only a painting provided by Bleeker (1877-78) was referred to in species diagnoses (Randall 2004;Allen & Erdman 2012). Because of the morphological similarities, confusion of P. jansenii with P. heptacanthus has occurred at least for the juvenile stage (Randall 2004;Randall & Heemstra 2009).
Based on collections made at fish markets and landing sites in central Vietnam and during a cruise in the area of the Great Barrier Reef to the Torres Strait, eastern Australia (Pitcher et al. 2007), the first two authors examined several samples of preserved goatfishes, partly associated with voucher photographs of fresh fish, at the Vietnam National Museum of Nature (VNMN), Hanoi, the Australian National Fish Collection (ANFC) at CSIRO, Hobart, and the fish reference collection at the Institute of Marine Research (HIFIRE), Bergen, Norway. A considerable number of the examined specimens was assigned to P. heptacanthus and P. jansenii using the available diagnostic information and identification keys (Randall 2004;Allen & Erdman 2012). Reliable distinction between the two species was however sometimes difficult, because relatively few diagnostic characters were listed in the previous studies, and colour patterns of. P. jansenii were documented only by a painting provided originally by Bleeker, reproduced in both Randall (2004) and Allen & Erdman (2012; see also further above).
Making use of the wealth of new study material of P. jansenii and P. heptacanthus now available, we adopted a comprehensive alpha-taxonomic approach (Uiblein et al. 2016(Uiblein et al. , 2017. This approach features comparative interand intraspecific studies of large numbers of quantitative morphometric and meristic characters as well as colour patterns, considering size-related and population differences, and possible colour changes related to freshness status. Diagnoses for both species and a redescription and updated occurrence information for P. jansenii are provided. Comparisons of P. jansenii were made with type material of the other five species of the heptacanthus group and P. angulatus. Finally, size classes and populations of P. jansenii are compared. The results are discussed with respect to the need for more taxonomic studies of the genus Parupeneus in insufficiently explored areas of the Indo-Pacific.

Material and methods
In total 29 specimens of Parupeneus jansenii, 53 specimens of P. heptacanthus from an area overlapping largely with the distribution of P. jansenii, and single types of each of the other five species of the heptacanthus group as well as P. angulatus were examined. Standard length (SL) and 62 quantitative morphometric and meristic characters were studied, taking the most recent taxonomic studies of relevant Parupeneus species into consideration (Randall 2004;Randall & Heemstra 2009;Randall & King 2009). The quantitative characters are listed in Table 1, partly accompanied by additional descriptions. Because the anal-fin spine is considerably reduced and sometimes hidden below the skin, measurements of this character were taken only when the spine was visible. All morphometric data are given as percentage of SL or as a ratio of SL or other characters as specified. Ranges of the resulting values below 10 were rounded to one decimal place or to the nearest 0.05. The means of all measurements and counts were rounded to one or two decimals. All Parupeneus species have eight dorsal-fin spines, and hence this meristic character is only listed in Table 1.
Size-related changes in morphometric characters were accounted for by splitting the data into two size classes, one of specimens > 110 mm SL ("large-sized fish") and the other specimens < 110 mm SL ("small-sized fish"). The size class threshold was established by plotting and subsequent inspection of the relationships between SL and all quantitative characters. Several morphometric characters showed strong positive or negative allometry in both species including several head characters and pelvic-fin length (examples for P. jansenii shown in Figure 1). Except for a slightly increasing number of developed vs. a slightly decreasing number of rudimentary gill rakers in P. heptacanthus (Figure 2), meristic characters did not vary with size.
For the diagnoses and comparisons of data generated by this study, all morphometric data are expressed in % SL. For the redescription and for additional comparisons of P. jansenii and P. heptacanthus with earlier published morphological data (Randall 2004), selected morphological characters were given as ratios of SL or other characters.
Colour patterns of fresh and preserved specimens were described and compared taking into consideration possible population differences and possible colour changes during life, after capture, or during preservation. Colour was identified using photographic documentation. Care was taken to consider artificial light effects when inspecting the photographs and applying colour corrections. Freshness status was determined by distinguishing photographs of specimens taken while still on board of a fishing vessel or at landing sites (= freshly caught) from those taken on fish markets (= still relatively fresh) or later in the laboratory (= less fresh). In-situ photographs or direct observations of P. jansenii colour patterns were not available for this study. In the comparisons, attention was focused on the most diagnostic characters only, considering both sample size and intraspecific variation in P. jansenii and P. heptacanthus. In comparisons with the other six species, all represented by single type specimens, only characters showing values lying clearly outside the ranges for P. jansenii are given. In addition, meristic data of the two species were statistically compared using Chi-square test for trends for larger tables (one degree of freedom in all cases) and Fisher's exact test for 2x2 tables (GraphPad Prism version 5.00 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com). Significance level was at p<0.05.

Characters Descriptions
Size and morphometric characters Standard length (SL; in mm) distance from snout tip to caudal fin base at mid-body (posterior end of hypural plate) Body depth at first dorsal-fin origin Body depth at anal-fin origin Half body depth at first dorsal fin origin depth from lateral line downwards Half body depth at anal fin origin depth from lateral line downwards Caudal-peduncle depth minimum depth anterior to caudal-fin dorsal origin Caudal-peduncle width width at position of caudal-peduncle depth measurement Maximum head depth vertical distance at ventral edge of operculum Head depth through eye distance along a vertical midline through eye Suborbital depth distance between lower edge of orbit to ventral midline of head Interorbital length least distance between upper bony edges of orbits Head length distance from snout tip to posterior-most margin of operculum Snout length distance from snout tip to anterior margin of orbit Postorbital length distance from posterior edge of orbit to posterior margin of operculum Orbit length horizontal bony orbit diameter Orbit depth vertical bony orbit diameter Upper-jaw length distance from symphysis to posterior end of upper jaw Lower-jaw length distance from symphysis to posterior end of upper jaw Maxilla depth maximum vertical depth of posterior end of upper jaw Snout width least distance between hinder margins of upper jaw, with closed mouth Barbel length Maximum barbel width horizontal width measured at base of soft part of barbel First pre-dorsal length distance from snout tip to origin of first dorsal fin Second pre-dorsal length distance from snout tip to origin of second dorsal fin Interdorsal distance distance from last spine of first dorsal fin to first ray of second dorsal fin Caudal-peduncle length distance from last anal-fin ray to ventral origin of caudal fin Pre-anal length distance from snout tip to origin of anal fin Pre-pelvic length distance from snout tip to origin of pelvic fin Pre-pectoral length distance from snout tip to dorsal origin of pectoral fin Second dorsal-fin depth distance from origin of second dorsal fin to origin of anal fin Pelvic-fin depth distance from origin of first dorsal fin to origin of pelvic fin Pectoral-fin depth distance from origin of first dorsal fin to dorsal origin of pectoral fin Length of first dorsal-fin base Length of second dorsal-fin base Caudal-fin length distance from dorsal caudal-fin origin to upper caudal-lobe tip Length of anal-fin base Anal-fin height distance from origin to anterior tip of anal fin (= to tip of second anal-fin ray) Height of anal-fin spine Height of first anal-fin ray Height of seventh anal-fin ray Length of pelvic-fin spine Pelvic-fin length distance from pelvic-fin origin to pelvic-fin tip Pectoral-fin length distance from pectoral-fin dorsal origin to pectoral-fin tip Pectoral-fin width width of pectoral-fin base Height of minute first dorsal-fin spine First dorsal fin height height of first long dorsal-fin spine (= second dorsal-fin spine in Parupeneus) Height of third dorsal-fin spine Height of fourth dorsal-fin spine Height of first dorsal-fin ray Second dorsal-fin height height of second dorsal-fin ray Height of eighth dorsal-fin ray Height of ninth dorsal-fin ray
Description. Morphometric data as ratios of SL for large-sized fish, data for small-sized fish (including syntypes) in squared brackets: body elongated, its depth at first dorsal-fin origin 3.8-4.3 [3.9-4.5]; body depth at anal-fin origin 4. 8-5.3 [4.8-5.4]; caudal peduncle depth 11-12 [11-13], subequal to interorbital length (10-12 [10-12]); maximum head depth 4. 1-4.5 [4.2-4.9], subequal to barbel length (4.0-4.6 [3.9-4.6]); head depth through eye 4.7-5.2 [4.9-5.6]; head length 2.9-3.2 [3.0-3.2], larger than maximum depth of body and caudal-fin length  Fresh colour: freshly landed large-sized fish ( Figure 3A, VNMN-I 4, 159 mm SL): body and head vermillion dorsally, becoming paler ventrally, starting from below eye and from scale row above lateral line; head with four slightly bending, thin, pink stripes: the uppermost stripe starting from behind dorsal third of eye, interrupted by orbit, to head margin in front of mid-eye, paralleled by a stripe of similar length right below; a longer pink stripe starting anterior to lower eye margin and ending close to nose tip; and the lowermost pink stripe from behind eye to below eye and ending anteriorly at half distance of snout length; jaws pink; further below eye the specimen has a large vertically oriented vermillion broad band that touches the lowermost rose stripe (see above) dorsally and ends well above ventral head margin; the preopercle is pale rose with vermillion margin; the black pupil of the eye is surrounded by an almost entirely red iris; all scales on body, apart from the ventral-most scale row, have dark reddish margins that contrast with the lighter scale colour; the lateral-line scales along the anterior half of the body show small red, mid-posteriorly placed spots; a broad, pale orange-yellow stripe in width of about scale vertical length runs along lateral line from behind preopercle to caudal-fin base; belly silvery white; caudal fin pale pink at base, with vermillion bands reaching along exterior margins of lobes to tips, covering in the anterior halves of lobes 5 to 6 of the distalmost rays; the remaining inner part of the caudal fin is covered by brighter yellow-orange colour reaching from distal-most part of fin base to close to lobe tips; dorsal fins vermillion, pelvic fins with some rose traces, and pectoral fins hyaline; anal-fin colour not identified, because folded in photograph; barbels white to pale rose, tips white.
Large-sized fish encountered at fish market ( Figure 3C, CSIRO H 5972-19, 146 mm SL): body and head mostly entirely vermillion, dorsally darker; ventral underside of head white; eye with black pupil surrounded by red iris that contrasts weakly with the darker vermillion head colour; body colour changing laterally towards ventral side from dark to bright vermillion to beige and rose and pale pink below lateral line; belly pale pink; scales with dark red or reddish margins that contrast with the lighter colour of scale surface, apart from the scales adjacent to the anterior dorsal body margin where body colour is darkest and the contrast is diminished; caudal fin entirely vermillion at base, the colour spreading along exterior margins of lobes to tips, covering at anterior halves of lobes 5 and 6 of the distalmost rays, then continuing along the two distalmost rays until tips, becoming much weaker close to tip, especially on upper lobe; the remaining posterior part of the caudal fin is yellow-orange to beige in colour; dorsal fins are mostly vermillion or rose, spines and anterior base of first dorsal fin most intensely coloured, rays of second dorsal fin proximally vermillion and pale on tips, the beige interspaces partly covered by one to three rounded, partly connected, vertically arranged vermillion patches; pelvic fins hyaline with pale rose spine and all rays except for innermost ray with three to four beige to pale brown patches proximally; pectoral fins hyaline; anal fin hyaline with three thin pale green to pale brown stripes covering entire length of fin, one close to base, the distalmost stripe thinnest, positioned at about two thirds of fin height. Barbels vermillion at base, distal two thirds white including tip.
Freshly caught small-sized fish ( Figure 4A, VNMN-I 135, 68 mm SL): body and head dark orange dorsally, becoming ventrally paler, starting from below eye and lateral line; head yellowish-orange below and behind ventral half of eye, with two weakly indicated pink dashes anterior to lower half of eye; snout tip and jaws dark orange; the black pupil of the eye is surrounded by a red iris; nearly all scales on body have reddish, orange or yellowish margins that contrast with the lighter scale-surface colour; the lateral-line scales form a thin bright orange to yellow stripe from posterior of the preopercle to behind second-dorsal fin base; this stripe is surrounded by a broader yellow pigmentation starting below first dorsal-fin base that widens from ca. mid-body to caudal fin base, almost entirely covering the dorsal part of the caudal peduncle; belly silvery white; caudal fin and dorsal fins dark orange; pelvic fins hyaline with pale orange spine and first three rays; pectoral fins hyaline with orange basis; anal fin hyaline; barbels pale rose along proximal two thirds and white distally including tip.
Parupeneus jansenii differs from P. angulatus and P. nansen in shorter caudal fin, shallower third dorsal-fin spine, and fewer lateral-line scales; furthermore, it differs from P. angulatus in round vs. angular rear end of maxilla, in shorter postorbital, longer caudal peduncle, shorter anal-fin base, slightly higher first anal-fin ray and first ray on second dorsal fin; and from P. nansen it differs in shallower maximum body depth, wider barbel, and shorter pelvic fin.
Parupeneus jansenii is morphologically rather similar to P. minys, especially, when comparing the 104 mm SL paratype of P. minys with P. jansenii small-sized fish. Accordingly, similar-sized P. jansenii differs in larger eyes, wider barbels, longer and wider pectoral fins, and shallower first dorsal-fin spine. Parupeneus jansenii reaches at least 159 mm SL (this study), while for P. minys a maximum size of only 106.5 mm SL has been reported (Randall & Heemstra 2009).
Parupeneus jansenii shows intraspecific differences among size classes and populations ( Figure 1; Table 3). The single large-sized fish from Indonesia differs from co-occurring small-sized fish in several morphometric characters. Small-sized fish of the NE Australian population differ from the other populations in having a higher eighth ray on the second dorsal fin. This difference applies also for the large-sized fish of the NE Australian population when compared with similar-sized specimens from Vietnam. The values for the geographically intermediate population from Indonesia largely overlap with those of the two other populations.

Discussion
This study highlights a goatfish species that to date has been poorly known and/or viewed as rare, having a much wider distribution and of more frequent occurrence than previously assumed. We report new records of Parupeneus jansenii for Vietnam, southern Indonesia (SE Indian Ocean), and NE Australia (Queensland), as well as the first verified record of this species for Australia. Earlier occurrence reports from the Australian Northern Territory (Wei et al. 1972) and from New Guinea (Munro 1958) have remained unverified.
Most of the specimens of Parupeneus jansenii studied (22 of 29 specimens) derive from collections made at fish markets and landing sites in south-central Vietnam and southern Indonesia during the last ten years, and from several trawling stations in the areas of the Torres Strait and Great Barrier Reef, NE Australia during a fairly recently conducted research cruise (Pitcher et al. 2007). Hence, one can assume that this species is not uncommon, but may have been misidentified or overlooked. Evidence for wrong or incomplete identification comes from labels attached to lots of specimens reading e.g. "Parupeneus sp." or "Parupeneus heptacanthus". In a recently published book featuring short diagnoses and colour photographs of market fishes of Indonesia (White et al. 2013) the specimen of P. jansenii from southern Bali (CSIRO H 5972-19, Figure 3C) has been wrongly referred to as Mulloidichthys pfluegeri (Steindachner, 1900). Another identification error has been discovered by Randall (2004) who stated that "the record of P. jansenii from the Seychelles by Randall and van Egmond (1994) is based on a misidentification of juvenile specimens of P. heptacanthus.". During later examinations, however, these specimens turned out to belong to P. minys, described by Randall and Heemstra (2009).
Misidentifications may happen especially when similar species cannot be sufficiently well distinguished based on available diagnostic information and identification keys. Table 4 provides a direct comparison of quantitative character data obtained in the current study with those listed in the diagnoses for Parupeneus jansenii and P. heptacanthus by Randall (2004). The two data sets largely overlap for individual characters and species (thus confirming consistency among authors) and both reveal clear interspecific differences in body depth and barbel length. They disagree however regarding pectoral-fin length which does not allow species distinction with our data set, possibly as a result of the larger number of specimens, wider size range, and more populations studied by us. On the other hand, many of the characters we studied that show high diagnostic value for correct identification of P. jansenii were not listed in earlier taxonomic studies (e.g., Randall 2004, Allen & Erdman 2012. For example, caudal-peduncle depth is introduced as a diagnostic character in this study, because it allows clear distinction of P. jansenii from P. heptacanthus if expressed as a ratio of head length (Table 4), or as a proportion of SL (Table 2). 4. Quantitative morphological characters expressed as ratios resulting from the present study (size classes merged for comparability) and from Randall (2004