A new microendemic species of the deep-water catshark genus Bythaelurus (Carcharhiniformes, Pentanchidae) from the northwestern Indian Ocean, with investigations of its feeding ecology, generic review and identification key

A new deep-water catshark, Bythaelurus stewarti, is described based on 121 examined specimens caught on the Error Seamount (Mount Error Guyot) in the northwestern Indian Ocean. The new species differs from all congeners in the restricted distribution, a higher spiral valve turn count and in the morphology of the dermal denticles. It is distinguished from its morphologically and geographically closest congener, B. hispidus (Alcock), by the larger size (maximum size 44 vs. 39 cm TL, maturity size of males 35–39 vs. 21–28 cm TL), darker fresh coloration and dark grayish-brown mottling of the ventral head (vs. ventral head typically uniformly yellowish or whitish). Furthermore, it has a strongly different morphology of dermal denticles, in particular smaller and less elongate branchial, trunk and lateral caudal denticles that are set much less densely and have a surface that is very strongly and fully structured by reticulations (vs. structured by reticulations only in basal fourth). In addition, the new species differs from B. hispidus in having more slender claspers that are gradually narrowing to the bluntly pointed tip without knob-like apex (vs. claspers broader and with distinct knob-like apex), more spiral valve turns (11–12 vs. 8–10) and numerous statistical differences in morphometrics. A review of and a key to the species of Bythaelurus are given.


Introduction
The family Scyliorhinidae sensu lato, comprising all catsharks of the order Carcharhiniformes, is the largest family of sharks with currently 158 described and valid species (number updated from Weigmann [1,2]). Members of the family reach maximum total lengths from 27 to 162 cm and live in shallow and deep water on continental and insular shelves and slopes in a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 interorbital space (INO) between anterior ends of orbits, caudal-fin length (CL) from ventral caudal-fin origin to the tip, caudal-fin height (CH) as greatest height from caudal-fin dorsal margin perpendicularly to apex of the ventral lobe, caudal-fin postventral margin (CPoV) from apex of caudal-fin ventral lobe to subterminal notch, and caudal-fin terminal lobe height (CTH) at subterminal notch. Additional measurements according to Kaschner et al. [9]: head width at level of lateral indention of head (slightly before anterior margin of nostrils) (HLIW), head width at level of maximum outer extent of nostrils (HONW) and head width at posterior edge of nostrils (HPNW). Stage of maturity was determined following Stehmann [11]. Terminology of glans clasper components is after Séret [12], Compagno [3], Kaschner et al. [9], and Weigmann et al. [4]. Vertebral counts and terminology follow Springer & Garrick [13]. Vertebrae and tooth rows were counted from radiographs, vertebrae were counted from all 121 type specimens, tooth rows from a representative selection of 50 specimens. Teeth and dermal denticles were examined using light microscopy and a scanning electron microscope (SEM). The feeding ecology was quantitatively and qualitatively examined from radiographs of all 121 type specimens. Stomach fullness was scored 0 = stomach empty, 1 = stomach partially filled, food remains detectable, 2 = stomach apparently full, extensively filled with food remains. The stomach of one female paratype (295 mm TL, ZMH 26252) was dissected for examination of food remains not identifiable from the radiographs. Tissue samples were taken internally from 10 paratypes of the new species and 14 specimens (13 specimens collected during cruise 17 of RV 'Vityaz' plus specimen BMNH 1898.7.13.21) of B. hispidus for molecular analyses. The catch locations of the 121 type specimens of the new species and verified occurrences of all other species of Bythaelurus are shown in the map in the Discussion section, which was generated based on the Global Relief Model ETOPO1 by the NOAA, the National Oceanic and Atmospheric Administration [14]. Country borders, lakes, and rivers were visualized by means of the shapefiles supplied by ESRI for the ArcExplorer-Java Edition for Education 2.3.2 (AEJEE). For a map with all stations of cruise 17 of R.V. 'Vityaz' see Weigmann et al. [15] or Weigmann et al. [16]. None of the specimens analyzed in this study were collected specifically for the purposes of this study. All specimens examined are preserved in museum collections. All type specimens of the new species were caught during cruise 17 of the Russian RV 'Vityaz' in 1988 and 1989, an official and authorized expedition exploring the western Indian Ocean from the Gulf of Aden to the southern end of the Madagascar Ridge at Walters Shoals. No special permissions were required to obtain specimens as the areas trawled were not protected.
The type specimens were deposited in the Zoological Museum Hamburg (ZMH).

Statistical treatment of morphological data
All morphometric and meristic data of the new species were statistically processed, involving ranges, means, and standard deviations. Morphometric and meristic data from 120 type specimens of the new species from the Error Seamount (one paratype with deformed caudal fin was excluded) and from 100 specimens of B. hispidus, caught off the Socotra Islands (see map in the Discussion section), were used to conduct three discriminant function analyses (DFAs). The DFAs were performed to determine if these two morphologically and geographically closest Bythaelurus species could be differentiated based on morphological parameters using XLSTAT (version 19.7.48622, 13.12.2017, Addinsoft), a statistical analysis add-in for Microsoft Excel.
DFA was used to demonstrate the degree of separation in multivariate space defined by the main patterns of morphological variation among species, which is described via the discriminant functions. It also shows which character contributes more to the differentiation. The standardized discriminant function coefficients represent the contributions of every variable to the discriminatory power of the function. Hence, the larger the standardized coefficient, the larger the weight of the variable in the function.
The first discriminant analysis was conducted for 120 individuals of the new Bythaelurus species and 100 specimens of B. hispidus. The second discriminant analysis was performed for 57 males and 63 females of the new species and 70 males and 30 females of B. hispidus. The third discriminant analysis was conducted for 7 and 42 adult males of the new species and of B. hispidus, respectively. Morphological variables with multicolinearity, variables including other variables, and variables not available for all specimens were excluded from the DFA procedures. The first two DFAs were performed for the following 67 morphological characters (for abbreviations see Table 1

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank.org/". The LSID for this publication is: urn:lsid:zoobank. org:pub:B24567BF-F872-4059-99EB-7EF32E1306A2. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

Systematic account
Bythaelurus stewarti Weigmann  Diagnosis. A medium-sized Bythaelurus species with the following characteristics: body firm and slender; snout long (preorbital length 4.9-7.4% TL) and broad, bell-shaped in dorsoventral view with distinct lateral indention; pre-outer nostril length 0.6-1.4 times internarial space; preorbital snout length 0.7-1.1 times interorbital space; preoral snout length 0.8-1.7 times in mouth width; eye length 10.2-15.5 times in predorsal distance, 4.9-7.7 times in head length and 1.2-2.3 times eye height; head length 2.2-2.6 times width at level of maximum outer extent of anterior nostrils; head width at level of maximum outer extent of anterior nostrils 1.1-1.3 times width at level of lateral indention of head, 1.2-1.6 times preorbital length, and 8.1-10.1% TL; tongue and roof of mouth densely set with knob-like oral papillae; pelvicfin anterior margin 1.6-3.5 times in pectoral-fin anterior margin; first dorsal-fin base 1.3-2.3 times in interdorsal space; length of second dorsal-fin inner margin 0.8-2.3 times in second dorsal-fin height; second dorsal-fin base length 5.1-8.9% TL; anal-fin base 0.7-1.9 times interdorsal space. Coloration: dorsally dark grayish-brown with rather indistinct dark blotches at nape, on flank, below both dorsal fins, and across caudal fin; ventral side grayish-white, usually with dark grayish-brown mottling on head. Upper jaw with 64-85 and lower jaw with 64-88 rows of small tricuspidate teeth with outer surface of crown furrowed by strong longitudinal ridges and strongly structured by reticulations; monospondylous trunk vertebrae centra 37-42, diplospondylous precaudal centra 37-45, total centra 125-140. Branchial, trunk and lateral caudal-fin dermal denticles loosely set, their surface very strongly and fully structured by reticulations. Claspers rather long and very slender, gradually narrowing to bluntly pointed tip without knob-like apex, inner margin length 10.1-11.3% TL, base width 1.4-1.5% TL; clasper hooks present along inner edge of large exorhipidion, large envelope overlapping part of clasper groove, inner lobe with rhipidion, cover rhipidion, pseudopera and pseudosiphon. The reproductive mode is yolk-sac viviparous. Bythaelurus stewarti n. sp. differs from all congeners in the distribution, which is apparently restricted to the Error Seamount. It further differs from all congeners in a higher spiral valve turn count (11-12 vs. 6-10) and in the morphology of branchial, trunk and lateral caudal-fin dermal denticles, which are loosely-spaced and not overlapping even in adult specimens of the new species, whereas they are closely-set and overlapping in all other Bythaelurus species. Compared to its morphologically and geographically closest congener, the new species further differs in a larger size, a ventral head with dark mottling, claspers that gradually narrow to the bluntly pointed tip without knob-like apex, and a surface of dermal denticles that is very strongly and fully structured by reticulations.
Description of the holotype. Values of the paratypes are presented in parentheses (n = 119 instead of n = 120 as the morphometrics of one paratype with deformed caudal fin were excluded). More complex differences between holotype and paratypes are described separately. Morphometric measurements and meristics are given in Table 1.
External morphology. Body firm and slender, subcircular in cross section at mid-trunk, laterally compressed and tapering posterior to cloaca; head region broad, long abdominal and caudal sections (Figs 1 and 2). No predorsal, interdorsal, or postdorsal ridges; no postanal ridge; no lateral ridges on caudal peduncle. Trunk shorter than tail, distance from tip of snout to anterior cloaca 45.1% TL (39.6-45.8% TL); pre-first dorsal-fin length 46.7% TL (41.8-47.6% TL), pre-second dorsal-fin length 64.4% TL (57.5-65.1% TL), ventral precaudal length 71.4% Table 1. Bythaelurus stewarti n. sp., morphometrics and meristics. Individual values for the adult male holotype (ZMH 26251) and one gravid female paratype (ZMH 26253), ranges for all other paratypes (n = 119 for the meristics except for n = 48 for the tooth row counts, but n = 118 for the morphometrics because the measurements of one paratype were excluded due to its deformed caudal fin), as well as means and standard deviations (SD) for all 121 type specimens concerning the meristics except for n = 50 for the tooth row counts, but for 120 type specimens concerning the morphometrics due to exclusion of the measurements of one paratype with deformed caudal fin. Proportional values are expressed as percentages of total length (TL) 70% ethanol preserved except for minimum, maximum, and mean of TL in mm. 3) times eye height; nictitating lower eyelids, anterior and posterior eye notches, and suborbital grooves present. Spiracles small and slit-like, close behind but well separated from eyes, dorsolaterally on head and somewhat lower than level of eye notches, spiracle length 4.2 (2.9-16.8) times in eye length and 7.9 (5.9-32.1) times in interorbital width. Gill slits moderately long, well separated, their upper ends clearly below level of lower edge of eye; gill area fully scaled, gill filaments not visible externally; gill openings increasing in size from first to fourth (first to third or fourth), the fifth smallest and above pectoral-fin origin. Nostrils oblique, expanding diagonally inwards from snout edge, clearly not reaching level of mouth, with large, triangular anterior nasal flaps and much smaller but still distinct posterior flaps; preouter nostril length 1.0 (0.6-1.5) times nostril width and 0.5 (0.4-0.6) times preoral snout length, nostril width 1.0 (0.7-1.4) times internarial width and 0.8 (0.7-1.3) times eye length. 2) times as long as upper ones. Tongue moderately long, flat and rounded, light-colored, densely set with small, knob-like papillae; entire roof of mouth densely set with moderately small, knob-like papillae ( Fig 5). Fleshy buccal curtain along inner margin of upper jaw extremely densely set with small, knob-like papillae, fleshy buccal curtain along inner margin of lower jaw loosely set with few very large, elongated, partially furcated papillae (Fig 6).
Upper jaw with approximately 77 (64-85) and lower jaw with about 74 (64-88) diagonal rows of small teeth (n = 50; Fig 6). Teeth in upper jaw tricuspidate with median cusp much longer than small lateral cusps (Fig 7A and 7B); teeth in lower jaw similar to those of upper jaw but with larger lateral cusps (Fig 7C and 7D). The median cusp decreases in size from symphysis to mouth corners, whereas the size of the lateral cusps increases, reducing the size difference of median and lateral cusps towards mouth corners in both jaws. Outer surface of crown furrowed by strong longitudinal ridges from base of cusps to tip and strongly structured by reticulations from basal areas to slightly into median cusp but to well into lateral cusps. Cutting edges of cusps without serrations.
Dermal denticles on dorsal and ventral ( Fig 8A and 8B) snout leaf-like to teardrop-shaped, densely set and overlapping, surface structured by reticulations in basal third, with two to four narrow ridges that do neither fuse nor reach the tip of the denticle. Snout denticles in juveniles of similar shape and also densely set and overlapping, but surface structured by reticulations in basal half and with only two to rarely three ridges that partially fuse and partially reach the tip of the denticle (Fig 9A and 9B). Denticles in branchial area, on lateral trunk (Fig 8C and 8D) and on lateral caudal fin tricuspidate, small, loosely set and not overlapping, with long and pointed median main cusp and shorter, pointed lateral cusps at lower level, surface very strongly and fully structured by reticulations, with about four narrow ridges that partially reach the tip of the median main cusp but are hardly detectable due to being camouflaged by the strong and dense reticulations. Branchial, trunk (Fig 9C and 9D) and lateral caudal-fin denticles in juveniles similar but even less densely set and partially unicuspidate without lateral cusps. Dermal denticles on anterior dorsal caudal-fin margin (Fig 8E and 8F) slightly enlarged, tricuspidate, densely set and overlapping, with long and pointed median main cusp and shorter, pointed lateral cusps at lower level, surface only weakly structured by reticulations only close to base, with a weak median ridge in basal half and two very strong and pronounced lateral ridges that do not fuse but reach the tip of the median main cusp, two additional, very weak lateral ridges in lateral cusps that reach or nearly reach the tip of the lateral cusps. Dermal denticles on anterior dorsal caudal-fin margin in juveniles (Fig 9E and 9F) strongly differ from those of adults, not enlarged, tricuspidate, loosely set and not overlapping, with long and pointed median main cusp and shorter, pointed lateral cusps at lower level (sometimes absent), surface very strongly and fully or mostly structured by reticulations, partially with a median ridge in basal half and two lateral ridges that fuse and reach the tip of the median main cusp, but ridges not present or hardly detectable due to being camouflaged by the strong and dense reticulations in about one third to half of the denticles. In embryos, the denticles are similar in shape on the snout but less densely set and hardly overlapping (Fig 10A), unicuspidate, as well as very loosely set and not overlapping in branchial area, on lateral trunk (Fig 10B and 10C), and on lateral caudal fin ( Fig 10D), but tricuspidate on anterior dorsal caudal-fin margin ( Fig  10D). In contrast to larger specimens, the latter denticles are somewhat smaller in size than those on the lateral caudal fin but still forming an obvious crest ( Fig 10D).
Second dorsal fin 0.7 (0.5-1.1) times as high as and about as long as first dorsal fin, anterior margin slightly convex, apex narrowly rounded, posterior and inner margins straight, free rear tip angular, base length 2.2 (1.6-4.0; no ontogenetic or sexual differences detectable) times fin height and 0.5 (0.4-1.1) times interdorsal space; second dorsal-fin origin over anal-fin midbase ( Fig 11B).
Caudal fin slender, relatively long (28.2-34.4% TL) and strongly asymmetrical, its length 5.0 (4.1-5.8) times fin height and 2.3 (2.3-3.7) times interdorsal space; dorsal caudal margin weakly convex, no lateral undulations; upper caudal lobe very low, lower caudal lobe much deeper, with straight pre-and postventral margins. Ventral caudal-fin origin anterior of dorsal caudal-fin origin due to long preventral margin, which is slightly shorter than the postventral margin and forms an elongated fleshy ridge in about anterior half of its length. Ventral corner bluntly angled; subterminal notch distinct; terminal lobe 5.6 (4.1-9.2) times in caudal-fin length; terminal caudal margin nearly straight without mesial indention ( Fig 11C).
Claspers (Fig 12) of mature males rather long, terminating distinctly before anal-fin origin and very slender, lateral margins nearly straight, not undulated, extending to about one third of their inner margin length beyond pelvic-fin free rear tips; inner margin length 10.1% (10.4-11.3%) TL, base width 1.4% (1.4-1.5%) TL. Glans somewhat elongated, length about half clasper inner margin; only slightly tapering to tip in distal half and gradually narrowing to bluntly pointed tip without knob-like apex. Ventral and outer lateral surfaces of clasper covered with small tricuspidate clasper denticles (CD), similar to those on trunk; dorsal and inner lateral surfaces largely naked. The narrow slit-like apopyle opens the clasper groove proximally; the hypopyle ends the concealed clasper groove distally and is detectable as a small cavity next to the rhipidion, but both are concealed by the cover rhipidion and exorhipidion and thus not visible in Fig 12. The proximally concealed clasper groove (CG) opens widely in the distal glans. A large, fleshy flap, the envelope (EN), on outer lobe of glans, overlaps part of CG; outer lobe also with a large, subtriangular exorhipidion (ER), which consists of a proximal convex blade and a distal fleshy wall; several enlarged clasper denticles, the clasper hooks (CH) along inner edge of ER. Inner lobe with a fan-shaped flap, the rhipidion, that partially covers the concealed part of CG and itself is concealed by a movable blade, the large cover rhipidion (CR); inner lobe also with two blind cavities: the large and deep pseudopera (PP), which is partially concealed by EN and RH, and-on the inner margin-the large, longitudinally slit-like pseudosiphon (PS).
Size. The description of the size is based on total length measurements prior to preservation except for the two female near-term embryos measured in 70% ethanol. A small catshark and a medium-sized species of Bythaelurus reaching a maximum total length of about 437 mm for females and 435 mm for males. Males are mature at 389 mm and juvenile at 348 mm TL. The smallest known free-swimming specimens have total lengths of 156 (female) and 162 mm (male), respectively. The size at birth is estimated at around 150 mm based on the smallest   Distribution. The new species is known only from the Error Seamount (Mount Error Guyot) in 380-420 m depth (see map in the Discussion section). It is apparently a microendemic species restricted to this isolated Seamount.
Etymology. The new species is named after the late filmmaker and shark conservationist Rob Stewart, who inspired the second author and stimulated her interest in sharks.
Discriminant function analyses for Bythaelurus stewarti n. sp. and B. hispidus. The first DFA for Bythaelurus stewarti n. sp. and B. hispidus provided one significant function (Box-Test with χ2 = 430.430 and p < 0.0001; Wilks' lambda = 0.195 and p < 0.0001), which explains 100% of the total variation in the data. Group centroids for morphological characters of both species are projected in Fig 14A. Both species are clearly separated in the discriminant space defined by the first function. From the standardized coefficients (Table 2), the two characters that have the greatest influence on the discriminant function (characters most discriminatory) are the preoral length (POR) and caudal-fin postventral margin (CPoV).
In the second DFA, three significant functions were estimated based on morphological characters of males and females of Bythaelurus stewarti n. sp. and B. hispidus (Box-Test with χ2 = 944.979 and p < 0.0001; Wilks' lambda = 0.025 and p < 0.0001). Together, these functions explain 100% of the total variation in the data. The first two functions explain 92.875% of the total variation in the data (Table 3), which is sufficient for further detailed analysis. The third discriminant function explains 7.125% of total variation. Group centroids for morphological characters of males and females of both species are projected in Fig 14B. Fig 14D presents the individual specimens projected onto the first two discriminant functions. As all four groups were clearly separated in the discriminant space defined by the first two functions, the third function was not used. The first discriminant function explains 56.666% of total variation (Table 3). It mainly separates the males of Bythaelurus stewarti n. sp. and B. hispidus. The females of these species cannot be clearly separated by the first discriminant function. The second discriminant function accounts for 36.210% of total variation. The females of Bythaelurus stewarti n. sp. and B. hispidus are clearly discriminated by this function. Pelvic height (P2H) and tail height at pelvic base end (TAH) are the two characters that have the greatest weight on the first discriminant function and for the discrimination of the males of both species (Table 3). The contrasts between the pelvic fin height (P2H) and the pelvic anterior margin length (P2A) are mainly responsible for the discrimination of the second discriminant function and for the discrimination of the females of both species (Table 3).  (Table 4), the two characters that have the greatest influence on the discriminant function are the total length (TL) and clasper base width (CLB).

Feeding ecology
The analysis of radiographs of all 121 type specimens of the new species revealed that 34 of 121 stomachs (28.1%) were empty (score 0), 78 (64.5%) were partially filled (score 1), and 9 (7.4%) were full or nearly full (score 2). Examples of the different scores can be found in Fig 15. The present findings are in line with those of other studies on different Bythaelurus spp. Nevertheless, the actual number of empty stomachs might be lower in the present study as soft food remains generally do not depict well in radiographs. In a study on B. canescens (Günther; for holotype data see [17]) by Acuña & Villarroel [18], the number of stomachs with food was 312 of a total of 513 stomachs examined (61%), accordingly the number of empty stomachs was 201 (39%). In a study by Lopez et al. [19], 37 of 50 stomachs of B. canescens sampled (74%) contained food, correspondingly 13 stomachs (26%) were empty. For B. hispidus Nair & Appukuttan [20] reported that 57 of 241 examined stomachs (24%) were empty. Nevertheless, in the same paper the authors indicated 52 of 241 empty stomachs (22%) when differentiating between adults and juveniles. Here, 36 stomachs showed traces of food, 60 stomachs were quarter full, 43 stomachs were half full, 17 three fourths full, 23 were full, and 10 gorged with food. Of 121 stomachs of B. hispidus examined by Akhilesh et al. [21], 24% were empty, 18% with trace contents only, 15% half full, 21% three quarters full, and 22% full.
In the present study, the diet mostly consisted of teleosts, cephalopods, and decapods. This is evidenced by the findings from radiographs, as well as the results of the dissection of female paratype 295 mm TL, ZMH 26252. The components most commonly found in radiographs, as well the dissected stomachs were partially digested teleost fish eyes (Fig 16, A). However, in radiographs teleost eyes and cephalopod eyes (Fig 16B) might be mixed up due to similar appearance. Based on radiographs, the frequency of occurrence of roundish structures, likely Table 2. Standardized coefficients of the first discriminant function (DF1) of the first discriminant function analysis (DFA) separating the two species Bythaelurus stewarti n. sp. and B. hispidus based on morphological characters. In bold, characters with the greatest weight in DF1. For explanation of morphological characters, see Table 1. teleost eyes but possibly-to a lesser extent-cephalopod eyes, was 53.6%. Further teleost remains detected were parts of vertebral columns (Fig 15C, 15E and 15G), which were detected in radiographs only, other skeletal parts (Fig 16C), which were detected in both radiographs and the dissected stomach, as well as scales (Fig 16D), which were detected in the dissected stomach only. Based on radiographs, the frequency of occurrence of skeletal parts of teleosts was 20.0%. In addition to eyes, beaks ( Fig 16E) and arms (Fig 16F) of cephalopods were also found in the dissected stomach, soft parts like arms probably being hardly detectable in radiographs. The beaks and arms were identified as Ancistrocheirus cf. lesueurii (D'Orbigny). With respect to decapods, different body parts were found in the dissected stomach (Fig 16G-16I), but in radiographs mainly claws or parts of claws were detected. The other crustacean parts are probably hard to detect in radiographs. The antennae and antennulae were likely from swimming decapods of the paraphyletic suborder "Natantia", the claws were probably from brachyurans. In the radiographs, the frequency of occurrence of decapod remains was 14.5%. Another diet possibly ingested by the new species but hardly detectable in radiographs and not found in the examined specimens are polychaetes. Nevertheless, polychaetes were not found to be important food items in other Bythaelurus species, with only one ( [18]) of two ( [18,19]) studies on B. canescens reporting very low numbers of polychaetes and none of three ( [20][21][22]) studies on B. hispidus detecting any polychaetes. Therefore, it is conceivable that polychaetes are not or hardly ingested by the new species. In the current study, based on radiographs the frequency of occurrence of unidentified stomach contents was 11.8%, one large example of unidentified food remains found in the dissected stomach is shown in Fig 16J, possibly representing a partially digested shrimp stomach. The composition of prey items found in the current study is in line with results of earlier studies on species of Bythealurus: based on the percent index of relative importance (%IRI) Acuña & Villarroel [18] found crustaceans to be the most important prey item in B. canescens (~64%), followed by teleosts (~23%), unidentified digested remains (13.2%), cephalopods (<0.2%) and polychaetes (<0.1%). A different diet was reported by Lopez et al. [19] for B. canescens, who used a generalized index of food (%GI) and found siphonophores to be the most important prey item (68%), followed by cephalopods (20%) and teleosts (12%).

Morphological characters DF1
In the geographically and morphologically closest species, B. hispidus, Nair & Appukuttan [20] in a study on specimens caught from the Gulf of Mannar, India, detected that 61% of diet by volume (%V) was composed of teleosts, followed by cephalopods (18%), crustaceans (16%), mud (6%), gastropods (0.2%), and algae (0.1%). In their study, in adult specimens fish dominated in the food, followed by cephalopods and crustaceans, whereas in juveniles crustaceans ranked first, followed by fish, while cephalopods were not found. In a more recent study on the diet of B. hispidus by Akhilesh et al. [21], the analysis of stomach contents (%IRI) revealed teleosts as the primary diet (52%), followed by crustaceans (37%) and gastropods (1%).

Parasites
Numerous nematodes were found in the stomach walls of two partially dissected paratypes ( Fig  17). Based on their morphology they have been identified as members of the family Ascarididae. Ascaridid nematodes have previously been reported from other catshark genera, e.g. [23,24].

Remarks
The reproductive mode was determined to be yolk-sac viviparous based on embryos found in adult female paratype ZMH 26253. Both uteri were filled with one near-term embryo each.  A) paratype, ZMH 26254, adult male, 391 mm TL, score 0, (B) paratype, ZMH 26254, juvenile female, 296.4 mm TL, score 0, (C) paratype, ZMH 26252, juvenile female, 299 mm TL, score 1 containing a teleost and few eye lenses, (D) paratype, ZMH 26254, adult male, 417 mm TL, score 1 containing a decapod (probably brachyuran) claw, (E) paratype, ZMH 26252, juvenile male, 340 mm TL, score 2 containing, among other remains, two decapod (probably brachyuran) claws and one teleost, (F) paratype, ZMH 26252, juvenile female, 299 mm TL, score 2 containing two decapod (probably brachyuran) claws and few eye lenses, and (G) paratype, ZMH 26254, adult male, 375 mm TL, score 2 containing a decapod (probably brachyuran) claw and a teleost. https://doi.org/10.1371/journal.pone.0207887.g015 The embryos both were females with 137.3 and 145.3 mm TL, respectively. The embryos were each enclosed in a thin, membranous egg case in utero. Yolk-sac viviparity is the rarest and most advanced reproductive mode within the Scyliorhinidae sensu lato, in which three reproductive modes occur [21]. In this reproductive mode females give birth to live young from a thin egg case in utero without direct link between embryo and mother via a uterine connection [21].
With regard to the apparent difference in maturity and maximum sizes between B. stewarti n. sp. and B. hispidus, intraspecific variation and phenotypic plasticity, caused by different environmental characteristics should be taken into consideration. According to Pitcher & Hart [25], individuals have evolved the capacity to overcompensate reproductive output in the face of variability in annual mortality. Compensation may result from changes in individual growth rates, age and size of maturation, reproductive output, and survival. One example is the spiny dogfish, Squalus acanthias Linnaeus, that has experienced reductions in neonate size, fecundity, age and size at maturity after decades of exploitation in the northwestern Atlantic [26,27]. However, the various further differences between B. stewarti n. sp. and B. hispidus warrant the description as a new species.
As catsharks of the genus Bythaelurus are generally inactive, demersal sharks [28,29], it is assumed that the new species from the Error Seamount is geographically isolated. Although other seamounts exist in the wider radius of the Error Seamount, they are usually located in much deeper waters, where B. stewarti n. sp., B. hispidus and similar species apparently do not occur. Furthermore, a deep channel of almost 4000 m depth separates the Error Seamount from the Socotra Islands. Therefore, the new species is supposed to be endemic to the Error Seamount, arisen from a speciation process, potentially driven by special environmental conditions and prey availability on the seamount. The assumption of the presence of an isolated population of Bythaelurus catsharks on the Error Seamount is supported by the fact that high levels of endemism are generally assumed for seamounts [30][31][32][33][34][35]. However, sampling biases might be caused by the occasional sampling of rare but widespread species [36,37]. Furthermore, the complex effects that seamounts have on ocean circulation and, accordingly, deposition of sediment and organic matter are poorly understood due to the large diversity in seamount size, shape, and distribution [30,35]. Nevertheless, as a result of their sharp relief across the floor of the ocean basins, seamounts generally have strong influences on water movements [33]. Processes like trapping of plankton (trophic focusing) and increased primary productivity resulting from localized upwelling generally cause seamounts to harbor richer benthic communities than surrounding deep-sea habitats [30,38]. The available rich food resources might possibly favor the growth of the specimens on the Error Seamount to larger sizes than all known populations of B. hispidus. As the Error Seamount is a flattopped guyot, it probably has a rather quiescent summit, covered with soft sediment, and steep slopes of mostly bare rock following Clark et al. [35], possibly supporting the development of rich benthic communities. Nevertheless, values of bottom temperature and salinity, taken during cruise 17 of RV "Vityaz" and kindly provided by Sergei A. Evseenko and Aleksej V. Mishin (IORAS, P.P.Shirshov Institute of Oceanology of the Russian Academy of Sciences, Moscow), as well as M.F.W. Stehmann are very similar at the catch locations of the new species as compared to the catch locations of B. hispidus.
Comparative molecular analyses of B. stewarti n. sp. and B. hispidus are desirable for further clarification of the phylogenetic relationships and degree of differentiation. For this purpose, Thomas Knebelsberger (DZMB) kindly tried to perform molecular analyses from tissue samples taken internally from 10 paratypes of the new species and 14 specimens (13 specimens collected during cruise 17 of RV "Vityaz" plus specimen BMNH 1898.7.13.21) of B. hispidus. Unfortunately, these attempts were not successful due to formaldehyde-fixation of the "Vityaz" material and possibly due to long-time storage of the BMNH specimen. Although non-formaldehyde-fixed material of B. hispidus from other regions has become available (see [4]), nonformaldehyde-fixed specimens of the new species are not available, unfortunately, and may well not become available for a long time due to the remote type location.

Discussion
Bythaelurus stewarti n. sp. differs from all congeners in the distribution, which is apparently restricted to the Error Seamount. It further differs from all congeners in a higher spiral valve turn count (11-12 vs. 6-10) and in the morphology of the branchial, trunk and lateral caudalfin dermal denticles, which are loosely-spaced and not overlapping even in adult specimens of the new species, whereas they are closely-set and overlapping in adults of all other Bythaelurus species.
The new species is morphologically and geographically closest to B. hispidus. The most obvious difference of the new species compared to B. hispidus is the size: in B. stewarti n. sp., adult males range between 389 and 435 mm TL and juvenile males between 162 and 348 mm TL (subadult males and male embryos unknown). In contrast, in B. hispidus specimens from the geographically closest and in the comparative material by far best represented location Socotra Islands (n = 100), adult males have total lengths between 274 and 338 mm, subadult males between 254 and 285 mm TL, and juvenile males between 166 and 266 mm TL (one male embryo has 126.7 mm TL 70% ethanol preserved). From off the Andaman Islands, even smaller adult males (BMNH 1898.7.13.21 and USNM 221384) of B. hispidus with only 258.5 and 260 mm TL, respectively, and from the Gulf of Mannar, an even smaller subadult male specimen (SMF 11465) with only 212.9 mm TL were examined. These values are in line with specimens from southern India studied by Akhilesh et al. [21], who found that most males between 200 and 230 mm TL possessed partially calcified claspers, the smallest mature male had 228 mm TL and the largest immature male 265 mm TL. In this study, the maturity size (95% C.I.) of males was 235 (223-245) mm TL. In addition, the fresh ground color is darker in the new species and the ventral head usually bears dark grayish-brown mottling (vs. mostly uniformly yellowish or whitish in B. hispidus). Distinct differences can be found in the morphology of the dermal denticles: in adult specimens of the new species, the snout denticles are broader and have less pronounced ridges, the denticles in branchial area, on lateral trunk and on lateral caudal fin are smaller, less elongate and set much less densely and their surface is very strongly and fully structured by reticulations (vs. structured by reticulations only in basal fourth) and the denticles on anterior dorsal caudal-fin are less elongate and set slightly more loosely. When comparing the denticles of juveniles of the new species with those of B. hispidus specimens with similar total lengths, even more pronounced differences become apparent: the snout denticles of juvenile B. stewarti n. sp. are strongly structured by reticulations, stouter in shape and have less pronounced ridges, the denticles in branchial area, on lateral trunk and on lateral caudal fin are even less densely set and have even smaller lateral cusps than in adults, and the denticles on the anterior dorsal caudal-fin margin are similar to the trunk denticles of adults, leading to strong differences compared to B. hispidus. Another distinct difference between B. stewarti n. sp. and B. hispidus can be found in the clasper morphology: the claspers of the new species are more slender and gradually narrowing to the bluntly pointed tip without knob-like apex (vs. claspers broader and with distinct knob-like apex). Furthermore, the new species has more spiral valve turns (11-12 vs. 8-10) and numerous differences in morphometrics, particularly in preoral length, caudal-fin postventral margin, pelvic height, tail height at pelvic base end, pelvic anterior margin of both sexes, and total length and clasper base width of adult males, are evident statistically.
A detailed comparison of the new species with B. alcockii (Garman) is not possible due to the lack of specimens or detailed descriptions. That species was described from the Arabian Sea without exact locality data and is known only from the holotype, which has been lost [28] (K.V. Akhilesh, pers. comm., 2014). Following Alcock's [39] description of the lost holotype of B. alcockii, B. stewarti n. sp. has a strongly different coloration (blotched dorsally and mottled ventrally vs. blackish with hoary gray surface and some fins whitetipped posteriorly in B. alcockii) and tooth morphology (teeth in upper jaw tricuspidate with median cusp much longer than small lateral cusps, teeth in lower jaw similar to those of upper jaw but with larger lateral cusps vs. teeth with median and lateral cusps of subequal length). Furthermore, the new species was caught in much shallower water (380-420 m vs. 1134-1262 m depth). B. alcockii was originally described as Halaelurus alcockii and preliminarily placed in the subgenus Bythaelurus by Compagno [3]. However, the dark coloration and long snout combined with great catch depth, reported for B. alcockii by Alcock [39], indicate that this species possibly belongs to Apristurus rather than to Bythaelurus, like assumed by Compagno [28]. However, the tooth morphology described for B. alcockii by Alcock [39] does not fit the characters of neither Bythaelurus nor Apristurus species and therefore seems questionable. Overall, the validity of the species is questionable [1,[3][4][5]9,28,29]. As the only known specimen has been lost, it is currently impossible to resolve this issue and B. alcockii remains a species of uncertain validity and generic assignment.  B. alcockii, B. bachi, B. clevai, B.  lutarius, B. naylori, B. stewarti n. sp., B. tenuicephalus and B. vivaldii are found exclusively in the western Indian Ocean, whereas the nineth species, B. hispidus, is also known from the eastern Indian Ocean [1]. Nevertheless, intrageneric differences in general morphology and body shape, the presence or absence of oral papillae, the presence or absence of a crest of enlarged dermal denticles on the anterior dorsal caudal-fin margin, and, particularly, genetics and reproductive modes indicate that Bythaelurus possibly comprises species from at least two different genera. Another possible grouping arises from the presence or absence of oral papillae: species with numerous oral papillae are B. bachi, B. canescens, B. clevai, B. dawsoni, B. giddingsi, B. hispidus,  possibly B. immaculatus, B. incanus, B. stewarti n. sp., B. tenuicephalus, and B. vivaldii, species without (or with rudimentary) oral papillae are B. lutarius and B. naylori. The presence of oral papillae in B. immaculatus is unknown but it is assumed that the species has oral papillae based on its apparently close morphological relationship to B. canescens and B. incanus. White & Last [41] did not examine the holotype of B. immaculatus for the presence of oral papillae (W.T. White, pers. comm., 2015). Furthermore, none of the type specimens could be found upon recent requests (H. A third possible grouping arises from the reproductive modes of Bythaelurus species that was reviewed by Francis [42]. He noted that some species, e.g. B. canescens and B. dawsoni, are oviparous, whereas others are viviparous.

Review of Bythaelurus species
A detailed comparison of morphological and morphometric characteristics, as well as maturity and maximum sizes of all valid species of Bythaelurus can be found in Table 5, an overview of geographic and depth distributions, reproductive modes and meristic characteristics of the 14 species of Bythaelurus is shown in Table 6.
Generally, the taxonomy and biology of Bythaelurus species are poorly known. So far, studies on the biology and distribution are restricted to few species, i.e. B. canescens, B. dawsoni, B. hispidus and B. lutarius, and partially based only on a small number of specimens examined [18][19][20][21]42,44,48,53,54,60]. Therefore, more specimens are needed of several species, especially of those from the Indian Ocean and B. immaculatus. In order to further improve the knowledge of Bythaelurus species in this area, a comprehensive redescription of B. hispidus is currently in preparation.