Resolution of the Acroteriobatus leucospilus species complex, with a redescription of A. leucospilus (Norman, 1926) and descriptions of two new western Indian Ocean species of Acroteriobatus (Rhinopristiformes, Rhinobatidae)

Recent sampling efforts and examinations of museum material provided evidence for a complex of species within Acroteriobatus leucospilus (Norman, 1926). The present manuscript contains a redescription of A. leucospilus involving the syntypes and additional material, as well as formal descriptions of two new species of Acroteriobatus Giltay, 1928. All specimens of both new species were found in the western Indian Ocean. Individuals of the first new species, hereafter referred to as Acroteriobatus andysabini sp. nov., were identified originating from Madagascar, and specimens of the second new species, hereafter referred to as Acroteriobatus stehmanni sp. nov., were only found off Socotra Islands at the junction between the Gulf of Aden and the Arabian Sea. Both new species appear to be endemic to the respective region and allopatric to A. leucospilus and occur in coastal waters to at least 80 m and 43 m depth, respectively. They differ from A. leucospilus in a number of characteristics including the maximum size and coloration in fresh. Acroteriobatus andysabini sp. nov. is a large species growing to more than 100 cm total length (TL) and with elongated bluish-gray spots on snout giving “stripe-nosed” appearance; numerous small bluish-gray spots on pectoral, pelvic, dorsal, and caudal fins; brown spots on trunk and fin bases; lateral tail folds striped orange and white; and ventral surface largely white but with a V-shape pattern of faint to dark speckled black spots on snout tip. Acroteriobatus stehmanni sp. nov. is a small species growing to ~62 cm TL and with sparse patterning with small bluish-gray circular spots confined to snout tip, posterior pectoral-fin margins, a pair on midbody, and few on posterior pelvic-fin margins, rather indistinct small to larger dark brown spots, and lateral tail folds and ventral surface white. Acroteriobatus leucospilus is a medium-sized species growing to ~96 cm TL and with patterning similar to A. andysabini sp. nov. but ventral surface uniformly white and lateral tail folds white or striped blue and brown. Taxonomical differences include nasal lamellae counts (42–48 in A. andysabini sp. nov. vs. 43–48 in A. stehmanni sp. nov. vs. 37–41 in A. leucospilus), snout angle (76–85° vs. 71–77° vs. 68–81°), and dorsal head length (24.2–33.5% vs. 17.2–22.8% TL vs. 24.0–29.2% TL). A key to the species of Acroteriobatus is given for the first time.


Introduction
The orders Pristiformes, Rhiniformes, and Rhinobatiformes (sensu Weigmann 2016) were recently revisited by Last et al. (2016a) and combined in the newly described order Rhinopristiformes based on molecular and morphological data, including the families Pristidae and Rhinidae and two new families: the Glaucostegidae and Trygonorrhinidae. Prior to this revision, the Rhinobatidae consisted of six genera, one of which was considered to be highly questionable, and 48 species, including five questionably valid species (Weigmann 2016). The revised family Rhinobatidae contains three genera: Acroteriobatus Giltay, 1928 with eight;Pseudobatos Last, Séret & Naylor, 2016a with nine;and Rhinobatos Linck, 1790 with 18 species (numbers updated from Last et al. 2016a, 2017. Species of Acroteriobatus are mostly confined to the western Indian Ocean with seven of the eight valid species occurring in this area and only two species also being recorded from the southeastern Atlantic. Species now assigned to Pseudobatos are restricted to the amphi-American region, while Rhinobatos is wider ranging, occurring mostly in the Indo-Western Pacific and eastern Atlantic (updated from Weigmann 2016). The three genera can be separated from each other externally by the morphology of their nasal flaps and molecular data (Last et al. 2016a). Species of Acroteriobatus attain maximum sizes of 601-1400 mm total length (TL) and are found in 0-366 m (usually shallower than 185 m) depths (updated from Weigmann 2016). The maximum size of 2050 mm TL, indicated for Acroteriobatus zanzibarensis (Norman, 1926) by Weigmann (2016) and Séret et al. (2016), was based on an erroneous size indication for the paratype in the original description of the species. The actual size of the paratype is~650 mm TL.
During the course of a broader study on western Indian Ocean guitarfishes, an undescribed species of Acroteriobatus was discovered from off Madagascar. Most of the paratypes of this new species were collected aboard the fishing vessel (FV) NOSY BE 11 during the Atimo Vatae expedition (see Bouchet 2014 for details). The Atimo Vatae (meaning "Deep South" in Antandroy language) expedition took place between April and June 2010. It involved a shore-based party of 40 scientists, volunteers, and trainees operating successively from Fort Dauphin (Tolagnaro) between 27 April and 17 May, and from Lavanono between 23 May and 13 June, collecting intertidally, with SCUBA and light dredging. Additionally, Institut de Recherche pour le Développement (IRD) deployed its research catamaran Antéa for sampling by diving, and a chartered shrimp trawler, FV NOSY BE 11, operated by Société des Pêcheries de Nossi Bé (Groupe Unima), was used to sample by dredging and trawling, conducting 119 hauls at 50-700 m depths. Another undescribed species of the genus was found while examining Acroteriobatus material from the northwestern Indian Ocean, collected by Matthias F.W. Stehmann during the 17 th cruise of the Russian RV Vityaz in 1988 and1989. In order to characterize the true Acroteriobatus leucospilus (Norman, 1926), the two syntypes of the species at BMNH were examined. Data of further, more recently collected material was added for a redescription of the species provided herein and to enable delimitation of two previously undescribed species, formally described in the present paper as well. A key to the species of Acroteriobatus is given for the first time.

Material and methods
Institutional acronyms follow Sabaj (2019) except for HMNH NASU for the National Museum of Natural History, National Academy of Sciences of Ukraine (NASU), Kyiv (Kiev). External morphometric measurements were taken by a vernier caliper to one-tenth of a millimeter (mm) from the specimens preserved in 70% ethanol. Measurements were taken between perpendicular lines where relevant unless stated otherwise and largely follow Last et al. (2004Last et al. ( , 2016a and Ebert and Gon (2017). The following measurements were newly added in the present paper: snout length (pre-eyeball) measured directly and horizontally, skeletal interorbital width complemented by intereye width measured across mid-eye when it first becomes visible, widths of second and fourth gill openings, body depths at anterior orbit and caudal peduncle, lengths of inner and outer spiracular folds, and distance between bases of folds. Presocket snout length was measured as the direct length from the snout tip to the firm part of the nasal capsule adjacent the orbit (forward of eye socket), spiracle length was defined as the greatest length of the main cavity, preoral length was measured as the direct length from the snout to the posterior edge of upper jaw at its symphysis, mouth width was taken across the exposed width, and pelvic-fin insertion to dorsal-fin origin was defined as the horizontal distance from the pelvic insertion to the origin of the first dorsal fin. Skeletal morphometrics and meristics were taken and counted from radiographs. The description of the new species from off Socotra Islands considers brief field notes and color photographs of fresh specimens, kindly provided by Matthias F.W. Stehmann (ICHTHYS).
The map with catch locations of the examined specimens of all three species of the Acroteriobatus leucospilus species complex was generated based on the Global Relief Model ETOPO1 by NOAA, the National Oceanic and Atmospheric Administration (Amante and Eakins 2009). 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 RV Vityaz, see Weigmann et al. (2013) or Weigmann et al. (2015).

Diagnosis
A large Acroteriobatus species distinguished by the following combination of characters: dorsal surface smooth, without prominent thorns or tubercles, except for enlarged granular denticles partially around orbital rims and irregularly distributed along midline from nape to somewhat anterior to first dorsal-fin origin; absent between dorsal fins and upper caudal fin. Snout semi-translucent with elongated bluish-gray spots giving "stripe-nosed" appearance, numerous small bluishgray spots covering disc and pectoral, pelvic, dorsal, and caudal fins on a brown background with larger, slightly darker, indistinct brown spots covering the tail and fin bases; outer edges of pelvic-fin margins white, lateral tail folds striped

Description of the holotype
Values of the eight paratypes at SAIAB and CAS, as well as the two larger paratypes at MNHN in parentheses, more complex differences are described separately. Where relevant, ratios are based on horizontal measurements unless otherwise stated. Detailed morphometric measurements and meristics of the holotype, eight paratypes at SAIAB and CAS, and the two larger paratypes at MNHN are given in Table 1.
Dermal denticles mostly minute, close-set, covering entire body and fins; surfaces mostly smooth to the touch, except for a continuous row of~15 to 20 somewhat enlarged granular and smaller denticles around the anterior and inner rim of orbits, extending to the anterior edge of inner spiracle; enlarged, coarse denticles patchily distributed along midline from nape to~50 mm anterior to first dorsal-fin origin, absent along midline between dorsal fins and upper caudal origin; enlarged denticles absent on snout tip, anterior and lateral margins of disc, pelvic fins, and bases of dorsal fins; nasal flaps and lamellae naked. Crowns of denticles mostly oval, except largest denticles with slightly erect, bluntly acute cusp, base stellate. Thorns absent. Prebranchial sensory pore patch distinct, extending to just behind the level of third gill slit. Post-scapular sensory canal rather indistinct, weakly undulated anteriorly, terminating near pectoral-fin insertions; sensory pores minute, canal not forming a shallow groove (Fig. 3).

Coloration
Prior to preservation: dorsal surface light to medium brown with an orangish tinge, an organized pattern of smaller bluishgray circular spots, with a darker ring encircling a lighter bluish-gray center, and indistinct larger brownish spots (Figs. 1 and 4); smaller bluish-gray spots cover the snout, disc, pectoral, pelvic, dorsal, and caudal fins; larger brownish spots mostly cover the tail and fin bases, but not extending onto fins; snout semi-translucent with somewhat elongated bluish-gray spots giving it a stripe-nosed appearance (Figs. 1,4,and 12). Outer edges of pelvic-fin margins white, outer edges of posterior dorsal-fin margins translucent whitish, lateral tail folds striped orange and white. Ventral surface mostly white, except Individual values for the juvenile male holotype (SAIAB 97396) and one female paratype (SAIAB 189815), ranges for all other SAIAB and CAS as well as the two larger MNHN paratypes (n = 9), and means and standard deviations (SD) for all nine SAIAB and CAS as well as the two larger MNHN type specimens are presented. Proportional values are expressed as percentages of total length (TL) 70% ethanol preserved except for minimum, maximum, and mean of TL in mm for faint to dark speckled black spots forming a V-shape on rostral snout tip; spots becoming darker and more numerous with growth ( Fig. 7); in small (≤215 mm TL), newborn specimens short V-shaped black lines with faint dark spots at base extending from tip to about one-third snout length to midpoint of internostril space; larger individuals (>300 mm TL) with prominent dark to black spots, becoming more numerous and with spots extending to about the midpoint of rostrum; pectoral and pelvic fins mostly white, except fin edges brownish ( Fig. 7). After preservation: dorsal surface similar but bluish-gray spots less pronounced and large brownish spots mostly faded and not evident (Fig. 2). Ventral surface also similar, except edges of pectoral and pelvic-fin margins fade to a lighter brown (Fig. 3).

Size
The new species is described from 15 specimens (ten males and five females), ranging in size from 177 to 603 mm TL. All types are juvenile, the size at birth is~180 mm TL. Based on images of unretained specimens (Fig. 12), Acroteriobatus andysabini sp. nov. is a large species of Acroteriobatus growing to more than 1000 mm TL.

Distribution
Acroteriobatus andysabini sp. nov. is presently only known from the coastal waters surrounding Madagascar and may be endemic to this region (Fig. 13). Records are from off Tôlanaro, Anosy region, and Cape Andavaka (southeastern Madagascar), Toliara and Ankilibe (southwestern Madagascar), and Antongil Bay (northeastern Madagascar). Its depth range is from shallow coastal waters to at least 80 m. Fricke et al. (2018) reported A. leucospilus as a new species record for Madagascar but commented that the species had been misidentified previously as Rhinobatos (= Acroteriobatus) annuulatus by several earlier authors.

Etymology
The species name andysabini is in honor of Andy Sabin for his gracious support of the Lost Sharks project at the Pacific Shark Research Center.

Diagnosis
A small Acroteriobatus species distinguished by the following combination of characters: dorsal surface smooth, without prominent thorns or tubercles, except for slightly enlarged granular denticles partially around orbital rims and rather regularly distributed along midline from nape to or to somewhat anterior to first dorsal-fin origin; absent between dorsal fins Fig. 16 Acroteriobatus stehmanni sp. nov., ZMH 25557, adult female paratype, 602 mm TL fresh, in total dorsal view taken directly after catching. Scale bar: 5 cm. The photograph was taken and kindly provided by Matthias Stehmann Fig. 17 Acroteriobatus stehmanni sp. nov., ZMH 25558, adult female paratype, 594 mm TL fresh, in total dorsal view taken directly after catching. Scale bar: 5 cm. The photograph was taken and kindly provided by Matthias Stehmann Fig. 18 Acroteriobatus stehmanni sp. nov., ZMH 25560, juvenile male paratype, 201.4 mm TL, in total dorsal view. Scale bar: 5 cm and upper caudal fin. Snout semi-translucent with a few, somewhat elongated bluish-gray spots only giving it a very reduced stripe-nosed appearance, patterning with small bluish-gray circular spots generally sparse and confined to symmetrical patterns on snout tip, posterior pectoral-fin margins, a pair of tiny spots on midbody behind occipital joint, and few spots on posterior pelvic-fin margins; indistinct brown spots on the body and dorsal and caudal fins; ventral surface white except for a blackish blotch and two tiny black spots on ventral snout tip in smallest juvenile paratype. Nasal lamellae 43-48; upper jaw tooth row count~64-78; 181-186 post-synarcual centra; 194-200 total vertebral segments; 63-67 total pectoral skeleton radials.

Description of the holotype
Values of the seven ZMH paratypes in parentheses; more complex differences are described separately. Where relevant, ratios are based on horizontal measurements unless otherwise stated. Detailed morphometric measurements and meristics of the holotype and seven ZMH paratypes are given in Table 2.
Dermal denticles mostly minute, close-set, covering entire body and fins; surfaces mostly smooth to the touch, except for 4 (4-5) slightly enlarged granular denticles on each anterior orbital rim, 1 (1-2) on each inner posterior rim, and 2 (1-2) on the inner edge of each spiracle (Figs. 20 and 24a-d); 53 (41-60) enlarged, coarse denticles rather regularly distributed along midline from nape to first dorsal-fin origin (to origin also in most paratypes but to~60 mm before origin in adult male paratype ZMH 25554 and to~10 mm before origin in juvenile male paratype ZMH 25560; about 10 posteriormost denticles only weakly pronounced in both female ZMH paratypes) (Figs. 24e, i and 25); enlarged denticles absent along midline between dorsal fins and upper caudal origin, on snout tip, anterior and lateral margins of disc, pelvic fins, and bases of dorsal fins; nasal flaps and lamellae naked. Crowns of denticles mostly oval, except largest denticles with slightly erect, bluntly acute cusp, base stellate (Fig. 24a-e). Crowns of denticles in juvenile paratype ZMH 25560 with acutely pointed cusp and a minute cusplet on each side of the cusp (Fig. 24f-i). Furthermore, the denticles of this juvenile paratype are proportionally larger compared to larger specimens and those in midline are set much denser (Fig.  24i vs. Fig. 24e and Fig. 25b vs. Fig. 25a). With respect to the minute dermal denticles covering entire body and fins, those anterior to orbit are tricuspidate in this juvenile paratype with three well-developed, acutely pointed cusps (Fig. 24f). Thorns absent in holotype and paratypes. Prebranchial sensory pore patch distinct, extending to about the level of third gill slit. Post-scapular sensory canal distinct, weakly undulated anteriorly, terminating near pectoral-fin insertions; sensory pores minute, canal not forming a shallow groove (Fig. 15).

Coloration
Prior to preservation: dorsal surface light to medium brown with a greenish tinge; the patterning with small bluish-gray circular spots is sparse, and such spots are confined to symmetrical patterns on the snout tip, posterior pectoral-fin margins, a pair of tiny spots on midbody behind occipital joint, and few spots on posterior pelvic-fin margins. The spots have rather inconspicuous and very thin brownish margins. Indistinct small (on anterior disc) to larger (on posterior disc and tail) dark brown spots can be found in varying degrees of conspicuousness, sometimes very indistinct, but always  bluish-gray spots mostly not visible anymore, at most few detectable on posterior pectoral-fin margins; brownish spots more or less faded, some specimens, including one of the females with fresh coloration described above, show almost no dark spots anymore. Outer edges of pelvic-fin (and partially posterior pectoral-fin) margins, as well as lateral tail folds creamy instead of white as typical for preserved  (Fig. 14). Ventral surface uniformly creamy as well (Fig. 14); smallest juvenile paratype ZMH 25560 with prominent blackish blotch on underside of snout tip plus two tiny black spots on outer edges of anterior snout about half way from snout tip to nostrils (Fig. 27).

Size
The new species is described from 10 specimens (seven males and three females), ranging in size from 208 to 622 mm TL (fresh) and from 201 to 613 mm TL (preserved). Adult males range from 546 to 622 mm TL (fresh), a 378 mm TL male is a very early subadult specimen, and a 208 mm TL male is juvenile. Accordingly, males of the new species apparently start to mature at around 378 mm TL. The two female paratypes of 594 mm and 602 mm TL are presumed to be adult, considering that they are distinctly larger than the smallest adult male paratypes. Based on all known specimens, the new species is assumed to be a small Acroteriobatus species reaching only about 622 mm TL.

Distribution
Acroteriobatus stehmanni sp. nov. is presently only known from the coastal waters surrounding the Socotra Islands and may be endemic to this region (Fig. 13). It is known from 36 to 43 m depth.

Etymology
The species is named in honor of Dr. Matthias F.W. Stehmann for his invaluable contributions to chondrichthyan taxonomy, particularly with respect to skates. He taught SW and BS chondrichthyan taxonomy, collected all type specimens of the new species, and kindly provided fresh photographs of two of them.  Gilchrist & Thompson (1911: 55); Gilchrist & Thompson (1916: 284)

Diagnosis
A medium-sized Acroteriobatus species distinguished by the following combination of characters: dorsal surface smooth, without prominent thorns or tubercles, except for slightly enlarged granular denticles partially around orbital rims and rather regularly distributed along midline from nape to near first dorsal-fin origin; absent between dorsal fins and upper caudal fin. Snout semi-translucent with elongated bluish spots giving stripe-nosed appearance, numerous small bluish spots covering snout, pectoral, pelvic, dorsal, and caudal fins but not central disc on a sandy brown background with darker brown spots of varying sizes covering the disc, tail, fin bases, and fins, sometimes giving the caudal peduncle a striped appearance; outer edges of pectoral-and pelvic-fin margins blue, lateral tail folds white or striped blue and brown; ventral surface white. Nasal lamellae 37-41; upper jaw tooth row count 60-75; 187-192 post-synarcual centra; 200-205 total vertebral segments; 63-70 total pectoral skeleton radials.

Description
The description is based on the two syntypes (BMNH 1905.6.8.12 andBMNH 1920.7.23.1), as well as nine nontype specimens. Where relevant, ratios are based on horizontal measurements unless otherwise stated. Detailed morphometric measurements and meristics are given in Table 3.
Nostril moderately large, oblique, all nasal flaps well developed; anterior aperture subcircular; nostril length 1.7-3.0 times anterior aperture width, 0.75-1.13 times anterior nasalflap base length, 0.5-0.94 times distance from nostril to edge of disc, 0.81-1.31 times internarial distance. Anterior nasalflap base greatly extended, penetrating horizontally into internarial space, mesial edges barely separated (by about width of posterior nasal flap); innermost extension of flap not narrow; flap base length 1.88-2.86 times its width at process, 1.88-3.33 times anterior aperture width; minimum distance between mesial insertions of flaps 6.75-26.5 in greatest  distance across nostrils anteriorly, 2.63-9.5 in minimum internarial distance; process of flap about twice as long as wide, narrowing distally to blunt pointed tip, and adjoining with posterolateral nasal flap and anterior aperture posterior margin. Posterolateral nasal flap well developed, lobe-like, broadest medially, length 2.0-5.5 times its width; originating just behind lateral extremity of anterior nasal aperture, extending posteromesially as a free fold almost to the level of insertion of posterior nasal flap. Posterior nasal flap well developed, lobe-like, base length 1.6-2.5 times its width, its inner edge nearly reaching innermost margin of nostril; width 0.5-1.0 times anterior aperture width, 0.92-2.0 times posterolateral nasal-flap width (Figs. 34 and 35). Mouth width 1.59-2.24 times nostril length, 5.94-7.25 in precloacal length; positioned beneath posterior margin of orbit. Upper jaw slightly convex, upper lip broadly arched; lower lip not pronounced, not separated from oral groove by ridges of strongly corrugated skin; corners of mouth with several strong, short lateral grooves. Teeth arranged in quincunx, small, close-set, rhombic, broad-based with short, bluntly rounded cusp; upper and lower teeth similar in shape and size; upper and lower jaws each with~60-75 tooth rows. Gill openings strongly wavy; length of third gill slit 1.63-3.4 in nostril length, 4.33-7.0 in distance between fifth gill slits; distance between first gill slits 1.31-1.46 times distance between fifth gill slits; distance between fifth gill slits 2.54-3.24 times internarial distance, 1.3-1.62 times mouth width, 0.32-0.38 of ventral head length (Fig. 35).
Dorsal fins erect, relatively tall, dissimilar in size and shape; first dorsal-fin anterior margin weakly convex, curving rearwards towards rounded apex; posterior margin weakly convex near tip, then becoming weakly concave to nearly straight; second dorsal-fin anterior margin weakly convex, curving towards pointed or slightly rounded apex; posterior margin nearly straight; free rear tips rounded of both dorsal fins, almost forming right angle, not produced; first dorsal fin Individual values for the subadult male syntype (BMNH 1905.6.8.12) and the juvenile male syntype (BMNH 1920.7.23.1), ranges for all examined nontype specimens (n = 9), as well as means and standard deviations (SD) for all 11 examined specimens are presented. Proportional values are expressed as percentages of total length (TL) 70% ethanol preserved except for minimum, maximum, and mean of TL in mm height subequal to second, length of first 0.79-1.15 times height, its base length 1.67-2.8 times inner margin length; second dorsal-fin length 0.84-1.21 times its height, base length 1.5-3.3 times inner margin length (Fig. 36a, b). First dorsal-fin origin well posterior to pelvic-fin free rear tips, interspace 1.04-1.99 times interdorsal distance; interdorsal space relatively short, 1.32-2.19 times second dorsal-fin height, 2.13-3.17 times base of first dorsal fin, 0.98-1.33 times tail width at origin of first dorsal fin, 1.29-1.9 times interspace between second dorsal-fin insertion and upper origin of caudal fin. Caudal fin relatively small; dorsal caudal margin weakly convex, length 1.03-1.99 times preventral margin length; ventral lobe broadly rounded; posterior margin weakly concave; fin tip acutely pointed (Fig. 36c). Dermal denticles mostly minute, close-set, covering entire body and fins; surfaces mostly smooth to the touch, except for 7 slightly enlarged granular denticles on each orbital rim and 3-4 on the inner edge of spiracle; up to~50 enlarged, coarse denticles rather regularly distributed along midline from nape to near first dorsal-fin origin, absent along midline between dorsal fins and upper caudal origin; enlarged denticles absent on snout tip, anterior and lateral margins of disc, pelvic fins, and bases of dorsal fins; nasal flaps and lamellae naked. Crowns of denticles mostly oval, except largest denticles with slightly erect, bluntly acute cusp, base stellate. Juveniles have proportionally larger denticles than larger specimens and the crowns of the denticles have acutely pointed cusps. The minute dermal denticles covering entire body and fins are also proportionally in juveniles, giving the skin a velvet touch. Thorns always absent. Prebranchial sensory pore patch distinct, extending to about the level of third gill slit. Postscapular sensory canal rather indistinct, strongly undulated anteriorly, terminating near pectoral-fin insertions; sensory pores minute, canal not forming a shallow groove (Fig. 29).
Rostral cartilage broad, its shaft just slightly increasing in width posteriorly from rostral node; rostral node rounded at apex, not angular, broadly expanded, relatively short, axis at maximum node width 34.4-41.2% of length of rostral cartilage from tip; anterior fontanelle relatively broad posteriorly, tapering gradually towards rostral node, dorsolateral edges of cartilage surrounding fontanelle (rostral ridges on surface of snout) well separated posteriorly, not constricting medially; rostral cartilage 54.2-58.1% of neurocranium length, ventral edges of rostral cartilage united; nasal capsules large, with their transverse axes directed anterolaterally; width of cranium across nasal capsules 1.15-1.27 times nasobasal length (base of rostrum to occipital condyles); width of nasal capsule 1.3-1.45 times its length; basal plate relatively broad, its minimum width 3.35-3.61 times in nasobasal length; cranial roof with small, oval-shaped fenestra, with its anterior edge located behind precerebral cavity by a distance~1.5 times its length; anterior cartilage triangular, narrow, posteriorly wedgeshaped, without an anterior lobe extending past nasal capsules; preorbital processes well developed; postorbital processes moderately large, not bifurcate; greatest width across processes 1.78-1.97 times in nasobasal length (Fig. 37).

Coloration
Prior to preservation: dorsal surface sandy brown with a pronounced and colorful, organized pattern of numerous small and large bluish circular spots, with a darker ring encircling a lighter bluish center, and small and large brownish spots (Fig. 31); the bluish spots cover the snout, pectoral, pelvic, dorsal, and caudal fins but not the central disc; the brownish spots cover the disc, tail, fin bases, and fins, sometimes giving the caudal peduncle a striped appearance (Fig. 31); snout semi-translucent with elongated bluish spots giving it a stripe-nosed appearance (Fig. 31). Outer edges of pectoraland pelvic-fin margins blue, lateral tail folds white or striped blue and brown (Fig. 31). Ventral surface white. After preservation: dorsal surface similar but bluish spots less pronounced and brownish spots mostly faded and not evident (Figs. 28 and 30). Ventral surface also similar but creamy rather than white as typical for preserved specimens (Fig. 29).

Size
The 11 examined specimens ranged in size from 223 to 960 mm TL. Acroteriobatus leucospilus has been reported to reach 1200 mm TL (Compagno et al. 1989;Séret et al. 2016;Weigmann 2016), but no voucher specimens of such size are known and the maximum size might possibly be based on the larger A. andysabini sp. nov. instead. The largest verified specimen is the adult female of 960 mm TL examined in the present study. The 416 mm TL male syntype BMNH 1905.6.8.12 and the 430 mm TL non-type male SAIAB 40000 are both subadult, and the 690 mm TL non-type female SAIAB 34588 is gravid. Following Wallace (1967), a male of 402 mm TL was juvenile, one of 415 mm TL was subadult, and specimens of more than 565 mm TL were mature, with gravid females ranging from 564 to 921 mm TL. Considering the rather small maturing size of about 415-565 mm TL, a maximum size above 1000 mm TL appears unlikely. Wallace (1967) indicates a size of 93 mm TL and a weight of 1/4 lb for the smallest embryo seen and 250 mm TL for the smallest juvenile on record (allegedly also weighing 1/4 lb). Based on the maximum verified size of 960 mm TL, A. leucospilus is a medium-sized species of the genus.

Distribution
Acroteriobatus leucospilus so far has been recorded from off the central Transkei coast (Eastern Cape Province) to the northern KwaZulu-Natal Province (South Africa) and Mozambique. Ebert et al. (2021), based on the examined material, extended its range to Zanzibar and Dar es Salaam, Tanzania (Fig. 13). It is known from 0 to 100 m depth (Compagno et al. 1989;Séret et al. 2016;Weigmann 2016) but seems to be most abundant in depths of around 20 fathoms or 37 m (Wallace 1967).

Discussion
The genus Acroteriobatus can be subdivided into two subgroups: those with elongated bluish-gray spots on the dorsal snout surface giving it a stripe-nosed color pattern and those without. The six species with stripe-nosed pattern include A. andysabini sp. nov., A. stehmanni sp. nov., A. leucospilus, Acroteriobatus salalah (Randall & Compagno, 1995), Acroteriobatus variegatus (Nair & Lal Mohan, 1973), and A. zanzibarensis. Among the stripenosed subgroup, only A. andysabini sp. nov. and A. variegatus have black blotches or spots on the ventral snout surface at all sizes. Acroteriobatus salalah and the poorly known A. zanzibarensis sometimes have dark blotches or spots on the underside of the snout, but not always. The other two species (A. stehmanni sp. nov. and A. leucospilus) have a plain white ventral snout color (except for the smallest juvenile paratype ZMH 25560 of A. stehmanni sp. nov.). The four species of the other subgroup (Acroteriobatus annulatus (Smith in Müller & Henle, 1841), Acroteriobatus blochii (Müller & Henle, 1841), Acroteriobatus ocellatus (Norman, 1926), and Acroteriobatus omanensis Last, Henderson & Naylor, 2016b) all lack a stripe-nosed pattern on the dorsal snout surface and have a plain light to white ventral snout surface without a dark blotch.
Acroteriobatus andysabini sp. nov. can be separated from all other Acroteriobatus species by a combination of characters including black spots on the ventral surface of its snout at all sizes and in both sexes. All other known members of the genus, except for A. variegatus, are either plain white or variably may have a dark blotch on the snout tip depending on the size. Its closest geographic congener, A. leucospilus, is plain white on the ventral snout surface at all sizes from neonates to adults, while in A. andysabini sp. nov., the black spots become more dense with size. Furthermore, A. andysabini sp. nov. has a higher nasal lamellae count than A. leucospilus (42-48 vs. 37-41), more obtuse snout angle (76-85°vs. 68-81°), bluishgray but no dark brown spots dorsally on central disc, outer edges of pelvic-fin margins white, lateral tail folds striped orange and white (vs. no bluish-gray but dark brown spots dorsally on central disc, outer edges of pectoral-and pelvicfin margins blue, lateral tail folds white or striped blue and brown), and larger maturity (above 603 mm vs. 415-565 mm TL) and maximum (more than 1000 mm vs. 960 mm TL) sizes. In addition, the anterior disc margin is weakly concave anteriorly, becoming convex posterior to constriction at the level of first propterygials, turning to weakly concave posterior to the level of posterior margin of spiracles and again becoming weakly convex towards broadly rounded apex (vs. straight); there is a continuous row of~15-20 somewhat enlarged granular denticles around the anterior and inner rim of orbits, extending to the anterior edge of inner spiracle (vs.~7 denticles on each orbital rim and~3-4 on the inner edge of spiracle), and the nasal capsules are proportionally wider (width 1.56-1.87 vs. 1.3-1.45 times length). Geographically, A. andysabini sp. nov. appears to be the only species of the genus occurring off Madagascar, where it is apparently endemic. Acroteriobatus leucospilus is distributed along the eastern African coast from KwaZulu-Natal Province (South Africa) to Zanzibar, Tanzania. Acroteriobatus andysabini sp. nov. differs from A. salalah particularly in the shape of the snout and disc (snout moderately long and narrowly angular anteriorly, snout angle before eyes moderately acute, 76-85°, disc wedge-shaped with undulate anterior margin vs. snout short and obtusely angled at~88°, disc heart-shaped with convex anterior margins) and size (maximum size >1000 mm vs. 7 8 0 m m T L ) . C o m p a r e d t o A . v a r i e g a t u s a n d A. zanzibarensis, this new species grows larger (maximum size >1000 mm vs. 750-790 mm TL), has a less acutely angled (76-85°vs. 59-64°) snout without (vs. with) orange stripes, and shows a pronounced pattern of bluish-gray spots on disc, tail, and fins (vs. only few bluish-gray spots on disc, tail, and fins).
Acroteriobatus stehmanni sp. nov. can be separated from all other Acroteriobatus species by a combination of characters including the patterns of bluish-gray and brown spots on the dorsal surface. In contrast to all other members of the stripe-nosed subgroup of Acroteriobatus, A. stehmanni sp. nov. only has a few, somewhat elongated bluish-gray spots, giving it a very reduced stripe-nosed appearance, a generally sparse patterning with small bluish-gray circular spots, confined to symmetrical patterns on snout tip, posterior pectoralfin margins, a pair of tiny spots on midbody behind occipital joint, and few spots on posterior pelvic-fin margins. The brown spots on the body and dorsal and caudal fins are indistinct, and the ventral surface is white except for a blackish blotch and two tiny black spots on the ventral snout tip in the smallest juvenile paratype. Compared to its morphologically closest congener, A. leucospilus, A. stehmanni sp. nov. also has a higher nasal lamellae count (43-48 vs. 37-41), smaller maturity (starting to mature at~378 vs. maturing between 415 and 565 mm TL), and maximum (622 mm  . Furthermore, A. leucospilus seems to be allopatric, distributed along the eastern African coast from KwaZulu-Natal Province (South Africa) to Zanzibar, Tanzania. Geographically, A. stehmanni sp. nov. appears to be sympatric only with A. salalah, which has very recently been confirmed for the Socotra Islands (Bogorodsky et al. 2021). Acroteriobatus stehmanni sp. nov. can easily be distinguished from A. salalah by the shape of the snout and disc (snout moderately long and narrowly angular anteriorly, snout angle before eyes moderately acute, 71-77°, disc wedgeshaped with straight anterior margin vs. snout short and obtusely angled at~88°, disc heart-shaped with convex anterior margins) and dorsal color pattern (snout with only a few, somewhat elongated bluish-gray spots, giving it a very reduced stripe-nosed appearance vs. snout usually with pronounced stripe-nosed pattern). Compared to A. variegatus and A. zanzibarensis, this new species is smaller (maximum size 622 mm vs. 750-790 mm TL, maturing starts at 378 mm TL vs. maturity size 580-640 mm TL) and has a less acutely angled (71-77°vs. 59-64°) snout without (vs. with) orange stripes and without (vs. with) pronounced stripe-nosed pattern.
Differences between both new species (A. andysabini sp. nov. and A. stehmanni sp. nov.) include the dorsal color pattern (bluish-gray but no dark brown spots dorsally on central disc, outer edges of pelvic-fin margins white, lateral tail folds striped orange and white in A. andysabini sp. nov. vs. very reduced stripe-nosed pattern and generally sparse patterning with small bluish-gray circular spots in A. stehmanni sp. nov.), snout angle (76-85°vs. 71-77°), shape of anterior disc margin (weakly concave anteriorly, becoming convex posterior to constriction at the level of first propterygials, turning to weakly concave posterior to the level of posterior margin of spiracles and again becoming weakly convex towards broadly rounded apex vs. straight), dorsal head length (24.2-33.5% vs. 17.2-22.8% TL), ventral coloration (blacks spots on the ventral snout at all sizes and in both sexes vs. ventral snout white except for a large blotch and two tiny spots in smallest juvenile paratype), as well as maturity (above 603 vs. starting to mature at~378 mm TL) and maximum (more than 1000 mm vs. 622 mm TL) sizes. Additionally, A. andysabini sp. nov. has a continuous row of~15-20 somewhat enlarged granular denticles around the anterior and inner rim of orbits, extending to the anterior edge of inner spiracle vs. 4-5 slightly enlarged granular denticles on each anterior orbital rim, 1-2 on each inner posterior rim, and 2 on the inner edge of each spiracle in A. stehmanni sp. nov. and the post-scapular sensory canal is rather indistinct vs. distinct. Furthermore, the preoral length is 2.0-2.62 vs. 2.63-3.02 times mouth width and 3.75-4.89 vs. 5.46-6.9 times internarial distance, the direct preorbital snout length is 1.92-2.39 vs. 2.59-2.78 times interspiracular distance and 3.09-3.63 vs. 3.83-4.32 times skeletal interorbital space, the distance between bases of spiracular folds is 0.3-0.7 vs. 0.85-1.83 times length of shortest fold, and the second dorsal-fin base length is 1.34-2.21 vs. 2.01-3.5 times inner margin length.
Biogeographically, Acroteriobatus andysabini sp. nov. appears to be endemic to Madagascar, while its closest geographic congener, A. leucospilus, appears to occur from off the central Transkei coast (Eastern Cape Province), South Africa, to at least Zanzibar, Tanzania. Literature accounts only list A. leucospilus as occurring off South Africa and Mozambique (Compagno et al. 1989;Séret et al. 2016), but here, we can confirm its distribution further north. Recently, A. leucospilus was reported from around Madagascar for the first time (Fricke et al. 2018), but that record, as well as subsequent records by Ghilardi et al. (2019), are now referable to A. andysabini sp. nov. Acroteriobatus zanzibarensis is geographically close but presently only known from off Zanzibar and Kenya. The only other geographically close species are the common A. annulatus, which occurs off Namibia to South Africa, and the poorly known A. ocellatus, which is only known from a few specimens off the Eastern Cape Province near Algoa Bay, South Africa (Compagno et al. 1989;Séret et al. 2016;Ebert et al. 2021). Four other species (A. omanensis, A. salalah, A. stehmanni sp. nov., and A. variegatus) are only known from the northern Indian Ocean, and all have very limited distributions (Séret et al. 2016;present study). The validity of A. variegatus and its separation from A. zanzibarensis are currently being reassessed, as both species are possibly conspecific (unpublished data; see also Séret et al. 2016). Acroteriobatus blochii occurs in the southeastern Atlantic off the west coast of South Africa to Namibia (Compagno et al. 1989;Ebert et al. 2021).

Conservation implications
Shark-like rays, specifically those of the order Rhinopristiformes, are among the most threatened species of cartilaginous fish globally (Dulvy et al. 2014;Kyne et al. 2020). This is due to life history characteristics such as slow growth, late age at maturity, and low fecundity, combined with a mostly nearshore coastal (<100 m depth) habitat that makes them vulnerable to overfishing and habitat degradation. The sawfishes (Pristidae) are perhaps the best-known family but also are one the most Critically Endangered groups, with all species having been extirpated from much of their ranges (Dulvy et al. 2016;Yan et al. 2021). More recently, the w e d g e f i s h e s ( R h i n i d a e ) a n d g i a n t g u i t a r f i s h e s (Glaucostegidae) have come into the public conscious as 15 of 16 species were assessed as being Endangered or Critically Endangered based on the International Union for Conservation of Nature (IUCN) Red List of Threatened Species categories (Kyne et al. 2020). While steps have been taken to address concerns over these better known rhinopristiform batoids, the small-bodied guitarfishes (Rhinobatidae) have mostly been overlooked even though they share the same life history and habitat characteristics, and are subject to intense fishing pressure and habitat degradation (Moore 2017;Jabado 2018;Jabado et al. 2018). Further compounding conservation and fisheries management efforts have been lack of taxonomic clarity and speciesspecific identification.
The implications for conservation and management concerns surrounding the two newly described Acroteriobatus species are considerable. First, the small-scale fishery for guitarfish in Madagascar (Fig. 12) has increased, coinciding with an increased demand for shark fins, including guitarfishes, in Asia (Humber et al. 2017). In fact, guitarfishes were one of three most important species and accounted for 75% of the overall elasmobranch landings in Madagascar (Humber et al. 2017). Two species of guitarfish are now known to occur off Madagascar: the new species A. andysabini sp. nov. and the recently described Rhinobatos austini Ebert & Gon, 2017. However, these two species are lumped into a single category, i.e., guitarfishes, without species-specific information, which is fundamental to effective conservation and management policy (Moore 2017). Presently, there are no management or conservation measures in place for sharks and rays in Madagascar (Humber et al. 2017). This is of significant concern since the distribution of A. andysabini sp. nov. seems to be restricted to Madagascar. Second, the distribution of A. leucospilus is much more restrictive than previously thought; it now appears to only occur on the east coast of South Africa to Tanzania (present study; Ebert et al. 2021). Acroteriobatus leucospilus has been assessed as Endangered by the IUCN Red List of Threatened Species (Pollom et al. 2019), but with its distribution now being more restrictive than previously thought, management and conservation policies may need to be reviewed. Third, A. stehmanni sp. nov. presently appears to be restricted to the waters surrounding Socotra Islands and may be subject to intense traditional fisheries (Moore 2017;Jabado 2018;Jabado et al. 2018). Improvement for the situation of the two guitarfish species occurring off Madagascar might be achieved through the national plan of action for sharks and rays in Madagascar (Plan national de gestion et de conservation des requins et des raies à Madagascar, PNGCRR) (Anonymous 2019a), associated to an implementation plan for the period 2019-2023 (Anonymous 2019b).
Most Acroteriobatus species, including the two new species, appear to be endemic or have very restrictive distributions. Nine of 10 Acroteriobatus species occur in the western Indian Ocean (WIO), and all are subject to intense coastal fisheries, which are reflected in the IUCN Red List of Threatened Species (IUCN 2020); the southeast Atlantic A. blochii only occurs outside this region. The conservation status of WIO Acroteriobatus guitarfishes reveals that three species are threatened, including A. variegatus (Critically Endangered), A. leucospilus (Endangered), and A. annulatus (Vulnerable), and A. salalah and A. zanzibarensis are Near Threatened. Two species (A. ocellatus and A. omanensis) are so poorly known they were assessed as Data Deficient, yet they occur in coastal areas subject to intense fisheries.
The recognition of two new species (A. andysabini sp. nov. and A. stehmanni sp. nov.), with the redescription of A. leucospilus, clarifies the taxonomic status and distribution of these three small-bodied guitarfishes. This is essential for improved data collection and research and for more effective conservation and management policy decisions.
for Aquatic Biodiversity (NRF-SAIAB). General assistance and support was kindly provided by Angus Paterson, Paul Cowley, Elaine Heemstra, Phil Heemstra, Roger Bills, Mzwandile Dwani, Nkoshinathi Mazungula, Vuyani Hanisi, and the fish collection staff (SAIAB), Rhett Bennett, Mike Markovina, Abdalla Said Abdulla, Naseeba Sidat, Jorge Sitoe, and Eleuterio Duarte (Wildlife Conservation Society); Stela Fernando (Instituto de Investigação Pesqueira, Mozambique); Baraka Kuguru (Tanzania Fisheries Research Institute); Marsha Englebrecht; Andy Sabin; Moss Landing Marine Laboratories; and the Save Our Seas Foundation, South African Shark and Ray Protection Project, implemented by the WILDTRUST, funded by the Shark Conservation Fund and the South African Institute for Aquatic Biodiversity. The authors are grateful to journal editors and reviewers of the manuscript for their helpful comments.
Funding Open Access funding enabled and organized by Projekt DEAL. This project was supported in part by the Save Our Seas Foundation, South African Shark and Ray Protection Project, implemented by the WILDTRUST, funded by the Shark Conservation Fund and the South African Institute for Aquatic Biodiversity.

Declarations
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All applicable international, national, and/or institutional guidelines for animal testing, animal care, and use of animals were followed by the authors.
Sampling and field studies All necessary permits for sampling and observational field studies have been obtained by the authors from the competent authorities and are mentioned in the acknowledgements, if applicable. The study is compliant with CBD and Nagoya protocols.
Data availability All data generated or analyzed during this study are included in this published article.
Author contribution SW and DAE conceived and designed the research. SW, DAE, and BS conducted the examinations. SW and DAE wrote the initial version of the manuscript. All authors read and approved the final manuscript.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.