Apogon fugax Gon, Bogorodsky, Mal & Alpermann 2020, new species

Apogon fugax Gon, Bogorodsky, Mal & Alpermann, new species

Elusive Cardinalfish

Figure 1, Table 3

Apogon sp. 1— Kuiter & Kozawa 2019: 34 (Red Sea and Myanmar).

Apogon sp. —Psomadakis et al. 2020: 404 (Myanmar).

Holotype. SMF 35884 [tissue sample KAU14-542], 46.15 mm SL, Red Sea, Saudi Arabia, off Jizan, 16°53.052’ N 42°21.122’ E, station 18, 60– 67 m, coll. T. J. Alpermann, S. V. Bogorodsky, A.O. Mal & M.H. Gabr, 5 November 2014.

2 Ivanova et al. 2007

3 Mabuchi et al. 2014

4 This study (modified after Mabuchi et al. 2014)

5 Miya & Nishida 2000

6 Li et al. 2007

7 Amplification was carried out at 52 °C for 10 cycles followed by 30 cycles with a reduction of -0.34 °C each cycle

...continued on the next page

...continued on the next page

...continued on the next page

1) Based on unpublished data (T. Fraser, pers. comm.).

2) Specimen identification was changed due to phylogenetic placement in one lineage with the specimen of Apogon crassiceps 1 from Mabuchi et al. 2014.

3) Identification based on re-examination during the present study.

4) COI barcode was added from BOLD entry Apogon dovii RDFCA 313-05.

5) Specimens from Western Australia listed as Apogon sp. were found in BOLD with the identification tool and identified as A. fugax due to their phylogenetic position in the present study. 6) COI barcode was added from BOLD entry Apogon imberbis BIM 752-19.

7) GenBank species ID is A. aurolineatus (tissue KUT 5231, see Mabuchi et al. 2014), but COI sequencing of A. pseudomaculatus (tissue KUT 213, also used in Mabuchi et al. 2014) showed that the resulting COI sequence is identical (see GenBank KF929614 or BOLD UKFBJ048-08). We therefore conclude that the GenBank sequence AB890009 is incorrectly treated as A. aurolineatus and we treat it here as A. pseudomaculatus. This is confirmed by the position of the COI sequences in a single gene phylogeny of the COI sequences used herein (not shown). 8) GenBank species ID is Nectamia fusca, but see Table 2 in Mabuchi et al. 2014.

9) The specimen is not related to A. deetsie species group, as shown in the present study.

10) The specimen is not conspecific to A. coccineus examined in the present study.

11) The specimen is not related to A. erythrinus, Hawaiian endemic species.

Paratypes. SAIAB 203703, 59.65 mm SL, Myanmar, off Ayeyarwady Delta, 13.0535 N, 96.6937 E, bottom trawl, 129 m, RV Dr Fridtjof Nansen, survey 2015404, station 123, coll. P. Psomadakis, 21 May 2015; SAIAB 205028, 42.5 mm SL, Myanmar, southwest of Ayeyarwady Delta, 15.6415 N, 93.8695 E, bottom trawl, 54.5 m, RV Dr. Fridtjof Nansen, survey 2015404, station 62, coll. P. Psomadakis, 9 May 2015; CSIRO H 6378-16 [BW-A4835], 52.9 mm SL, Western Australia, 21.9833 N, 113.8167 E, 166 m, CSIRO, 11 December 2015; CSIRO H6378-17 [BW-A4836], 42.7 mm SL, collected with previous paratype.

Diagnosis. A member of the Apogonini with two supraneurals, pectoral-fin rays 12, large membranous preopercular skin flap, head length 2.2–2.4 in SL (2.4–2.8 in SL in the other members of the ‘ talboti look-alikes’ and A. deetsie groups), lower jaw length 3.85–3.9 in SL (vs. 4.1–4.5 in SL), length of first dorsal-fin spine 9.5–11.0 in SL (vs. 12.5–17.85 in SL), total gill rakers 4+10–11 (vs. 3–5+10–18), and uniformly reddish orange body.

Description. Dorsal-fin rays VI+I,9; anal-fin rays II,8; pectoral-fin rays 12 (both sides). Lateral-line scales about 24+3 and predorsal scales about 6–7 (both counts of scale pockets); total number of gill rakers 4+10 (4+10– 11); developed gill rakers 2+9 (2+9–10); gill rakers on ceratobranchial 7. Scales transforming ctenoid, one full size scale between lateral line and dorsal-fin origin, and 7 scales between lateral line and anal-fin origin; scales around caudal peduncle 12; lateral-line canal simple, narrow, and apparently lacking dorsal and ventral pores.

Body relatively deep and compressed, its depth 3.1 (2.9–3.3) in SL, and its width 1.7 in depth and 5.5 (5.1–5.6) in SL. Head large, 2.2 (2.4) in SL; snout shorter than eye, 5.1 (5.3–5.6), eye large, 3.4 (3.1–3.7), and interorbital width 5.9 (5.3–5.7), all in head length. Predorsal distance 2.3 (2.4), preanal distance 1.6 and prepelvic distance 2.6 (2.5–2.7), all in SL; distance between pelvic-fin insertion and anal-fin origin 4.0 (3.4–3.8) in SL. Posterior edge of opercular bones thin and weakly ossified; posterior preopercular edge with 15 and 18 (left and right sides, respectively) serrations spread along most of its length; ventral part of preopercle with large, very thin membranous skin flap extending posteriorly to edge of gill cover, its ventral edge seems slightly crenulate, but could be damaged; preopercular ridge, posttemporal and infraorbitals smooth.

Mouth large, somewhat oblique, upper jaw 2.0 (1.9) and lower jaw 1.75 (1.6–1.65), both in head length; maxilla reaching posteriorly to below posterior edge of eye, its vertical edge slightly concave, and its depth 3.85 (3.8–4.0) in upper-jaw length. Ventral edge of infraorbitals membranous and indented at second and third infraorbital pore positions, extending ventrally over part of upper jaw. Both jaws with broad band of villiform teeth, 7–8 teeth across widest place; vomer and palatines with similar teeth, the former with single series and latter with two series tapering to one posteriorly; upper-jaw teeth exposed when mouth closed. Free tip of tongue rounded, short, about half pupil diameter.

Anterior nostril round, with elevated rim, closer to tip of snout (excluding upper lip) than to posterior nostril and lower than posterior nostril; posterior nostril narrow, oval, about 4.5 in eye diameter, at level of middle of eye and close in front of its anterior edge. First supraorbital pore next to tip of snout (excluding upper lip), medial to ventral edge of anterior nostril and of similar size, its short skin flap not reaching tip of snout; second supraorbital pore medial to posterior nostril and about half its vertical length; third supraorbital pore next to dorsal edge of orbit and in front of vertical through centre of eye, its size about two-thirds of posterior nostril. First infraorbital pore below ventral edge of posterior nostril, with no free skin flap and of similar size to second supraorbital pore; second and third infraorbital pores on ventral edge of infraorbital bone, of similar size, lack a free skin flap and separated by space more or less equal to their diameter.

Dorsal profile of head straight, moderately sloping from dorsal-fin origin to snout; first dorsal-fin origin at vertical through upper pectoral-fin base; anal-fin origin below base of second dorsal-fin ray; second dorsal and anal fins reach backwards to about two-thirds of caudal-peduncle length when pressed down against peduncle; first dorsal-fin spine 1.7 (1.8–2.0) in second spine and 4.3 (4.6–5.3) in head length; second spine longest, 2.5 (2.4–2.6) and third spine 2.65 (2.6–2.7) in head length; spine of second dorsal fin 3.2 (2.9–3.1) and longest dorsal-fin ray (second) 1.8 (1.6) in head length; first soft dorsal-fin ray branched. First anal-fin spine 4.3 (4.05–4.3) in second spine; second spine 3.6 (3.3) and longest ray (second) 1.9 (1.75–1.8) in head length; first soft anal-fin ray branched. Pectoral fin inserted relatively low on body, its upper base at level of mid-distance between ventral edge of eye and upper jaw, and distance from its upper base to first dorsal-fin origin 1.5 in body depth (4.6 in SL); length of pectoral fin 3.3 (3.4) in SL, its middle rays longest, reaching posteriorly over base of 6 th anal-fin ray; uppermost pectoral-fin ray short, but not rudimentary, its length about one-third of ray immediately below it; lowermost ray and uppermost two pectoral-fin rays unbranched. Pelvic-fin length 4.15 (4.2–4.3) in SL, inserted in front of vertical through dorsal-fin origin, almost reaching anal-fin origin; pelvic-fin spine 1.6 (1.6–1.7) in fin length and 3.0 (2.9–3.0) in head length. Caudal-fin rays 9+8, upper- and lowermost unbranched; upper and lower procurrent caudal-fin rays 6; caudal-peduncle length 3.9 in SL, depth 1.9 (1.9–2.1) in its length. Caudal-fin length 2.8 (2.95–3.2) in SL.

Colour after death (Fig. 1A & B): reddish orange all over, scale edges enhanced by larger chromatophores; five black dots orientated vertically on membrane between 2 nd and 3 rd dorsal-fin spines; faint dark dots of various sizes across caudal peduncle immediately in front of caudal-fin base.

Colour in alcohol (Fig. 1C): body creamy white with no other markings; fins mostly clear, but tips of branched principal caudal-fin rays with dark dots which extend anteriorly over two thirds of fin’s middle rays; several dark dots scattered on distal third of pelvic-fin rays; 5 dark dots in a vertical line on membrane between second and third dorsal-fin spines; dark dots scattered on occiput and on area between upper end of gill cover and first dorsal-fin origin; faint dark dots of various sizes loosely clustered posteriorly across caudal peduncle.

Etymology. This species is named fugax (Latin for elusive) for avoiding discovery despite intensive collecting efforts in the Red Sea and Indian Ocean.

Distribution and habitat. Apogon fugax is currently known from five specimens, including the holotype trawled off Jizan, southern Saudi Arabia, Red Sea, between 60–67 m; two paratypes trawled off Myanmar at depths of 54 and 129 m; and two paratypes trawled off Western Australia at a depth of 166 m (Fig. 2). We assume that this species is associated with a moderately deep rocky or coral reef habitat and therefore difficult to collect by diving or trawling; the Red Sea specimen was trawled together with two species that inhabit such habitats, namely Apogonichthyoides pseudotaeniatus (Gon) and Lutjanus argentimaculatus (Forsskål), and the two Myanmar specimens were captured in the early evening (between 1650 and 1830 hrs), possibly when they emerge from their daytime cover to feed.

.

*Data in parentheses are for non-types from Greenfield (2007) and Gon (unpublished data)

**Data transformed from Greenfield & Randall (2004)

Comparisons. The phylogenetic analysis grouped Apogon fugax with A. deetsie and A. rubrifuscus in one clade, albeit with very high support(Apogon - 4in Fig.3).This clade was recovered as the sister group of a larger clade(Zapogon to Apogon -3) of mostly Atlantic and Mediterranean species that together formed a well-supported clade (Apogon -4 to Apogon -3). By contrast, the two species of the ‘ talboti look-alikes’ species group (A. caudicinctus, A. dianthus) for which sequences were available were recovered as sister taxa and closest to A. talboti in a highly supported clade that also included A. unicolor Steindachner & Döderlein 1883.This A. unicolor clade, named after its basal species, was recovered as the sister clade of the remaining Indo-Pacific Apogonini species, together forming the large Apogon -1 clade (clade terminology follows Mabuchi et al.2014).The lower nodes forming the main clades in the phylogram all received high bootstrap support (Fig. 3). The topology described above implies that the enlarged ventral preopercular skin flap had evolved more than once independently in Apogonini species (it is also present in a number of Atlantic species).

Although similar in its general appearance, Apogon fugax differs from its sister species, A. deetsie (Fig. 4A) and A. rubrifuscus (Fig. 4B & C), and from the two remaining species of the ‘ talboti look-alikes’ species group, A. caudicinctus (Fig. 5A & B) and A. dianthus (Fig. 5C), in having fewer gill rakers, a larger head and a longer first dorsal-fin spine (Table 3). In addition, A. deetsie has an indistinct blackish bar below the second dorsal fin, and both, A. caudicinctus and A. deetsie, have a blackish bar posteriorly on the caudal peduncle. Moreover, A. caudicinctus has two scales between the lateral line and dorsal-fin origin and a smaller ventral preopercular skin flap. One specimen of A. rubrifuscus from the Society Islands (MNHN-IC-2008-0997) has a faint pattern of melanophores below the rear of the second dorsal fin and posteriorly on the caudal peduncle, whereas no melanophores are visible in the holotype; more specimens are needed to define this character in A. rubrifuscus. Apogon soloriens, a newly described species from Japan similar to A. caudicinctus and A. deetsie in having a black bar at the end of caudal peduncle (Fig. 5D) and a preopercular skin flap, has been added by Yoshida & Motomura (2020) to the ‘ talboti look-alikes’ group. It differs from A. fugax in having 8–9 developed gill rakers, including a single raker on the upper limb, and 13 pectoral-fin rays. Apogon fugax differs from A. talboti and A. unicolor (Fig. 5E & F, respectively) in having two supraneurals (vs. three), a developed membranous preopercular flap (vs. no flap), and in having 12 pectoral-fin rays (vs. 13 and 14, respectively). See Table 3 for other potential differences between A. fugax and some of these species, but these differences need to be confirmed when additional specimens of this species group are available. Other Apogonini species with uniformly red body colour distributed in the Indian Ocean (e.g., A. campbelli Smith, A. coccineus Rüppell, A. crassiceps Garman, A. doryssa Jordan & Seale, A. erythrosoma Gon & Randall, and A. indicus Greenfield), have a higher number of pectoral-fin rays and/or no enlarged flap on preopercular ventral edge.

Remarks. In their description of Apogon rubrifuscus, Greenfield & Randall (2004) related their species to A. caudicinctus and A. deetsie. They pointed out that due to their similarity to A. talboti these species were found in collections under that name. Greenfield (2007) added A. dianthus to this group saying that he also found them misidentified as A. unicolor and, following Greenfield & Randall (2004), he referred to these four species as the ‘ talboti look-alikes’. Apogon soloriens is the latest addition to this ‘species group’. Apogon talboti and A. unicolor differ from this species group in having three supraneurals (vs. two), no membranous preopercular flap (vs. developed), and in having 13 and 14 pectoral-fin rays, respectively (vs. 12 rays, 13 in A. soloriens). Apogon fugax would then have been the sixth species of this ‘species group’ and the first to be found in the Red Sea, bringing the number of apogonid species in the Red Sea to 60 (see e.g. Gon & Randall 2003; Gon & Allen 2012; Gon et al. 2013).

Of the three species now comprising the ‘ talboti look-alikes’ group, A. caudicinctus has a wide Indo-West Pacific distribution (see references in Eschmeyer et al. 2019) and a shallow depth range at 0–40 m (usually less than 12 m) (Randall & Smith 1988; Kuiter & Kozawa 2019); A. dianthus has an Indo-West Pacific distribution at a depth range of 7–27 m (Greenfield 2007) and A. soloriens is known only from the Bonin Islands, Japan, at depths of 4–45 m (Yoshida & Motomura 2020). Apogon deetsie is known from Oahu, Hawai’i, and Rangiroa Atoll (Tuamotu Archipelago) at a depth of 25–53 m (Randall 1998, 2007). Apogon rubrifuscus is known from the type specimen collected in Easter Island at a depth of 39 m and from specimens from the Society Islands at depths of 6–9 m (Greenfield & Randall 2004; present study). At present A. fugax has the deepest records among other Indo-West Pacific species of Apogon, its specimens trawled at depths of 54– 166 m.

Despite earlier phylogenetic analyses confirming that the genus Apogon is a paraphyletic group (see e.g. Mabuchi et al. 2014), we assigned the new species to Apogon sensu stricto pending a revision of the group (O. Gon, in preparation).