Gymnoxenisthmus tigrellus, new genus and species of gobioid fish from the Red Sea (Gobioidei: Xenisthmidae) Author Gill, Anthony C. Author Bogorodsky, Sergey V. Author Mal, Ahmad O. text Zootaxa 2014 3755 5 491 495 journal article 46570 10.11646/zootaxa.3755.5.9 4c3ba205-d1d6-4a9e-87c7-eaeb0d829ca7 1175-5326 226656 96AA169A-0FB4-4475-BB99-559FC2EBA860 Gymnoxenisthmus , new genus Type species. Gymnoxenisthmus tigrellus , new species . Diagnosis. The following combination of characters distinguishes Gymnoxenisthmus from other xenisthmid genera: head pores absent; scales lacking; first dorsal fin with five spines; dorsal- and anal-fin rays branched; at least some pectoral-fin rays branched; pelvic fin with a spine and five unbranched rays. Comparisons. Characters distinguishing xenisthmid genera are summarised in Table 1 . The genus is most similar to Paraxenisthmus and Xenisthmus in general body form, in most meristic details, and in having dorsal-and anal-fin rays and at least some pectoral-fin rays branched. It differs from Paraxenisthmus in lacking laterosensory pores on the head, vomerine and palatine teeth, and scales, and in having only five (versus six) spines in the first dorsal fin and a relatively broad (versus narrow) proximal head on the third branchiostegal ray. It differs from Xenisthmus in lacking scales, laterosensory pores on the head, and developed gill rakers (though tiny rudiments are present), and in having only five (versus six) spines in the first dorsal fin, a narrower proximal head on the third branchiostegal ray (see Remarks below), and in having only unbranched rays in the pelvic fins (versus anterior four segmented rays branched in Xenisthmus ). Gymnoxenisthmus might also be confused with Rotuma , with which it shares the following characters: first dorsal fin with five spines; pelvic fins with a spine and five unbranched segmented rays, the inner of which is vestigial; no head pores; and no scales. It is readily distinguished from Rotuma in having branched (versus unbranched) dorsal-, anal- and pectoral-fin rays, a relatively broad (versus narrow) proximal head on the third branchiostegal ray; autogenous anterior and posterior ceratohyals (versus a single ceratohyal ossification present); and more segmented rays in second dorsal and anal fins (13 and 12, respectively, versus 9 in both fins). Remarks. Gill and Hoese (1993) hypothesised the following relationships among xenisthmid genera: ( Paraxenisthmus ( Xenisthmus ( Rotuma + Tyson + Allomicrodesmus ))). Character evidence for these relationships was in part from osteological characters, which have not yet been verified in Allomicrodesmus . We similarly lack detailed osteological information for Gymnoxenisthmus . Considering those characters that can be observed in the two genera, evidence supports placement of Gymnoxenisthmus in a clade that also includes Rotuma , Tyson and Allomicrodesmus (synapomorphies: five or fewer spines in first dorsal fin; sensory pores absent; scales absent; fifth segmented pelvic-fin ray vestigial or absent). Within this clade, Rotuma , Tyson and Allomicrodesmus in turn form a clade that excludes Gymnoxenisthmus (synapomorphies: all segmented second dorsal-fin rays unbranched; all segmented anal-fin rays unbranched; all pectoral-fin rays unbranched; ceratohyals represented by a single ossification). Contrary to these relationships, Gymnoxenisthmus shares a possible synapomorphy with Xenisthmus : enlarged proximal head on third branchiostegal ray. Gill and Hoese (1993) noted that Xenisthmus species differ from other xenisthmids in having a very broad proximal head on the third branchiostegal ray (see Springer 1983 : fig. 9), which they interpreted as an autapomorphy of the genus. The proximal head of the third branchiostegal ray is also enlarged in Gymnoxenisthmus , but not to the same extent as in Xenisthmus species (slightly broader than the proximal head on the fourth ray in Gymnoxenisthmus , versus about twice the width of the fourth ray in Xenisthmus species). More detailed phylogenetic analysis must await more complete osteological studies of Gymnoxenisthmus and Allomicrodesmus (which in turn are dependent on the discovery of additional specimens of both genera for osteological preparation). Etymology. The generic name is a combination of the Greek gymnos , meaning bare or naked, and the gobioid genus Xenisthmus , and alludes to the absence of scales on the body. Gender is masculine. Gymnoxenisthmus tigrellus ͵ new species Figure 1 , Table 1 Holotype . SMF 34903, 15.2 mm SL, gravid female, Red Sea, Saudi Arabia , Farasan Archipelago, unnamed island, 16°47.451’N 042°11.838’E , coll. S.V. Bogorodsky & T. Alpermann, 22 February 2012 . Diagnosis. As for genus. Description. Dorsal-fin rays V + I,13, all segmented rays branched; first dorsal-fin pterygiophore formula 3- 2210; anal-fin rays I,12, all segmented rays branched; pectoral-fin rays 15/15, upper 2/2 and lower 1/1 rays unbranched; pelvic-fin rays I,5, all segmented rays unbranched, inner ray vestigial; segmented caudal-fin rays 9 + 8; branched caudal-fin rays 6 + 6; upper unsegmented caudal-fin rays 6; lower unsegmented caudal-fin rays 6; total caudal-fin rays 29; no developed gill rakers (about 6 tiny rudiments present on upper part of ceratobranchial 1); vertebrae 10 + 16; epurals 2. TABLE 1. Comparison of selected characters of xenisthmid genera. * indicates where data are included from undescribed species. Paraxenisthmus Xenisthmus * Gymnoxenisthmus Rotuma Tyson Allomicrodesmus * Pelvic I, 5 I , 5 I ,5, inner ray I,5, spine and 1 3 or absent vestigial inner ray vestigial

Scales	present	present	absent	absent	absent	absent
D1	VI	VI	V	V	absent	II
D2	I,11–12	I,11–15	I,13	I,9	I,8–9	29–33

Segmented D2 yes rays branched	yes	yes	no	no	no
A I,10	I,10–14	I,12	I,9	I,8–9	23–26
A rays branched yes	yes	yes	no	no	no
Pectoral 15–18	15–18	15	16	17–21	10–12
Pectoral rays yes branched	yes	yes	no	no	no

Pelvic rays no yes, outer 4 no no no no branched

Segmented C rays 9 + 8	9 + 8	9 + 8	9 + 8	8 + 7	8 + 7
Upper 7–8 unsegmented C rays	6–9	6	6	8–9	7
Lower 7–8 unsegmented C rays	5–9	6	6	9	6–7

Vertebrae	10 + 16	10 + 16–17	10 + 16	11 + 15	13 + 13	17 + 26–28
Epurals	2	2	2	1	1	?
Head pores	present	present	absent	absent	absent	absent
Vomerine teeth	present	absent	absent	absent	present	absent
Palatine teeth	present	absent	absent	absent	absent	absent

Gill rakers	absent	at least some no developed developed, 2– rakers, about 6 4 + 8–14 = tiny rudiments on 10–17 upper part of ceratobranchial 1	absent, but tiny teeth present	absent	absent
Anterior and posterior ceratohyals	autogenous	autogenous autogenous	undifferen- tiated	undifferen- tiated	undifferentiated
Proximal head on 3rd branchiostegal	narrow	very broad broad	narrow	narrow	narrow

As percentage of SL: head length 28.3; predorsal length 39.5; prepelvic length 29.6; preanal length 60.5; first dorsal-fin origin to second dorsal-fin origin 17.1; second dorsal-fin base length 30.9; anal-fin base length 25.7; pectoral-fin base depth 6.6; first dorsal-fin origin to pelvic-fin origin 17.1; second dorsal-fin origin to anal-fin origin 13.2; snout length 5.9; orbit diameter 7.2; head width 14.5; body width 13.2; bony interorbital width 1.3; snout tip to retroarticular tip 13.2; caudal-peduncle length 15.1; caudal-peduncle depth 9.2; length of first spine of first dorsal fin 12.5; length of third spine of first dorsal fin 14.5; length of spine of second dorsal fin 9.2; length of first segmented ray of second dorsal fin 10.5; length of last segmented ray of second dorsal fin 11.8; anal-fin spine length 7.2; length of first segmented anal-fin ray 8.6; length of last segmented anal-fin ray 13.2; pectoral-fin length 21.7; fourth segmented pelvic-fin ray length 23.0; caudal-fin length 18.4, ray possibly regrown (longest mid-ray on lower hypural 23.7). Scales absent; laterosensory head pores absent; lower lip fleshy and protruding, with uninterrupted, free ventral margin; anterior naris in short tube (abnormally branched on left side); posterior naris without raised rim or membranous flap; tongue tip rounded; gill opening extending anteriorly to vertical through about midpoint between preopercle edge and eye. FIGURE 1. Gymnoxenisthmus tigrellus , holotype, SMF 34903, 15.2 mm SL, gravid female, Farasan Archipelago, Red Sea. Photo by Sven Traenkner. Upper jaw with 2 or 3 (anteriorly) to 2 (posteriorly) rows of slightly curved conical teeth; lower jaw with 2 or 3 (anteriorly) to 2 (posteriorly) rows of slightly curved conical teeth; vomer and palatine edentate. Live coloration (based on colour photograph of holotype when freshly dead; Figure 1 ): head and body translucent bluish grey; body with fourteen equally spaced, mid-lateral orange markings (first a spot just anterior to pectoral-fin base; second a short bar just behind pectoral-fin base; third a chevron below anterior part of first dorsal fin; fourth a chevron below posterior edge of first dorsal fin; fifth a chevron through space between dorsal fins; sixth a chevron through first segmented ray of second dorsal fin; seventh a bar through third segmented ray of second dorsal fin; eighth a bar through fifth segmented ray of second dorsal fin; ninth a bar through seventh and eighth segmented rays of second dorsal; tenth a bar through tenth segmented ray of second dorsal fin; eleventh through base of last second dorsal-fin ray; twelfth and thirteenth bars slightly oblique and less distinct, through caudal peduncle; fourteenth a narrow, short bar along posterior edge of hypurals); small, indistinct pale orange or yellow spots present between most midlateral orange bars and spots; scattered melanophores present on head and body, mostly confined to orange areas; head with orange stripe extending from mid-upper part of upper lip to upper half of eye, then from behind eye to point above upper edge of preopercle, with large (almost pupil-sized) orange spot near edge of opercle; second, oblique orange stripe extending from just behind and below eye to middle of operculum; isolated orange spot on anterior part of operculum, between two orange stripes; lower lip orange; orange “L”-shaped marking on cheek extending from just below anterior margin of eye, with bottom of “L” extending along lower cheek edge; iris mostly orange on dorsal two-thirds, remainder pale yellow to pale gold; pectoral-fin base orange anteriorly and dorsally, with two pale orange spots, one on mid-upper and the other on mid-lower part of fin base; first dorsal fin with first spine base orange; remainder of fin translucent on outer third, followed proximally with silvery white stripe, orange stripe, and silvery white basal stripe; scattered melanophores on first dorsal, these densest over orange stripe; second dorsal fin orange (basally) to dusky orange (distally), with narrow (anteriorly) to broad (posteriorly) distal margin translucent; orange bars from body extending on to fin base; silvery white stripe through middle of fin, with a basal series of short, silvery white oblique bars that extend anterodorsally from between orange bars from body; a few scattered silvery white spots present on translucent part of fin; anal fin translucent with narrow indistinct dusky orange stripe through basal third to half of fin, indistinctly bordered basally with silvery white spots (anteriorly) and stripe (posteriorly); caudal fin mostly translucent, with two large indistinct pale grey-orange spots on basal part of fin, one dorsal and the other ventral; pectoral and pelvic fins translucent. Preserved coloration: head and body generally pale beige, greyish brown on lower abdomen; melanophores within orange bars on body remain, though indistinct and confined to midside and upper half of body; melanophores within orange areas on lips, upper stripe on head and upper half of pectoral-fin base remain; orange markings on first dorsal fin become dark grey; orange markings on second dorsal and anal fins remain, becoming dark grey-brown. Etymology. The specific epithet is from the Latin, meaning a little tiger, alludes to the orange bars on the body. The name was selected by school children at the Australian Museum Science Festival Expo in August 2013 . Habitat. The holotype was collected from an unnamed rocky island with a narrow reef flat, and a slope with patches of corals and a rocky wall of about 3m with small caves and shelters. The sandy slope began at depths of 8– 10 m ; the holotype was collected on sand at the base of coral in 8 m .