Chondrichthyans from the Grippia bonebed (Early Triassic) of Marmierfjellet, Spitsbergen

The Grippia bonebed is located at Marmierfjellet in Flowerdalen, Isfjorden area of Spitsbergen. The bonebed occurs in the Vendomdalen Member, Vikinghøgda Formation, Early Triassic (Spathian). It is unique in Spitsbergen because of the richness of chondrichthyans, osteichthyes, amphibians and ichthyopterygians. This paper focuses on the chondrichthyans with more than 550 chondrichthyan teeth studied, together with three fin spines and one cephalic spine, assigned to 7 genera and 15 species, 8 of which are new from the Grippia niveau. The Hybodontiformes represent five of the identified genera, where most of the identified species belong to Hybodus and Acrodus. Hybodus sasseniensis, previously recorded from the older Dienarian substage, and is now shown to occur in the younger Spathian substage, and suggested to be the senior synonym of H. rapax. The possibility of Acrodus scaber and A. spitzbergensis being morphospecies, and the Acrodus genus in general, is further discussed. The occurrence of Hybodus microdus in the material is uncertain since it is shown that the mesio-distal length of the teeth is much larger than described by Stensiö in 1921. The Neoselachii (modern shark), previously only known from a fin spine of Nemacanthus from the Triassic of Spitsbergen, is now described from teeth belonging to possibly two new neoselachian species. They are referred to Synechodontiformes 1 and 2, awaiting more material to be processed.


Introduction
The Permian-Triassic Mass Extinction (PTME) Four lineages of fish and fish-resembling vertebrates survived the PTME; Cyclostomata, Conodonta, Chondrichthyes and Osteichthyes (Scheyer et al., 2014).One of the main groups of top predators who suffered most during the PTME were the chondrichthyans according to Benton et al. (2014) and Scheyer et al. (2014).Other studies (e.g., Schaeffer, 1973;Thomson, 1977;Patterson & Smith, 1987;Koot, 2013;Romano et al., 2016) state that the Chondrichthyes, and most fish-like creatures in general, were practically unaffected by the mass extinction since the more mobile chondrichthyans were able to survive anoxic conditions and possessed the ability to change food sources (Hallam & Wignall, 1997).Twitchett (2006) suggested that the chondrichthyans may have adapted to the environmental changes by reducing body size, often called the "lilliput effect".
The marine fauna shifted from a rich Chondrichthyeslife in the Carboniferous-Permian to one dominated by the Osteichthyes in the Mesozoic and Cenozoic.This appears to be a long-term process and not a direct result of mass extinction although the latter may have intensified the process (Romano et al., 2016).
The Chondrichthyes are divided in two subgroups, the Elasmobranchii (shark, skates and rays) and the Holocephali (chimaeras).Chondrichthyans have a cartilage skeleton, which is rarely preserved, whereas their teeth, consisting of enamel and dentine and regularly replaced, are commonly found as fossils (Cuny & Risnes, 2005).For these reasons our knowledge of ancient sharks is much enhanced thanks to their teeth.(Straube et al., 2008;Klug, 2010;Koot, 2013).
The majority of early sharks became extinct by the end of the Permian period, such as symmoriids, petalodontiforms, stethacanthids and bransonelliformes (Fig. 1).
unit was studied by the early expeditions of Nordenskiöld (1864-68) in the Isfjorden area.Here, field parties discovered several horizons with fish and tetrapod remains (Nordenskiöld, 1867).Hulke (1873) erected the hybodontiform shark species Acrodus spitzbergensis from material collected by Nordenskiöld in his expedition to Saurieberget in 1864 and 1868.Stensiö (1918) gave a short report about Hybodus sp., Acrodus spitzbergensis and Gyrolepis?sp. from the Hornsund area.He compared these specimens with his own material from the Isfjorden area and found them almost identical.The material from the Isfjorden area was later described in Triassic Fishes from Spitzbergen (Stensiö, 1921).This material was collected by small expeditions, initiated by Professor Carl Wiman, in 1912Wiman, in , 1913Wiman, in , 1915Wiman, in -1917 and several other small expedition during 1918 (Stensiö, 1921).Stensiö (1921) established nine new shark species after his findings.The only discovery of neoselachian at Spitsbergen is a Nemacanthus fin spine, identified by Stensiö (1921) (Koot, 2013).Cox & Smith (1973) described two Triassic chondrichthyans collected by CASP (formerly known as Cambridge Arctic Shelf Programme) in 1969.One was a cephalic spine belonging to a hybodont shark and the other was a hybodont Acrodus tooth, which was collected at Marmierfjellet, Spitsbergen in the upper part of Vikinghøgda Formation.Birkenmajer & Jerzmanska (1979) described material collected by the Polish Spitsbergen Expedition at Hornsund (Lower Triassic), south Spitsbergen, in 1960.The material consisted mainly of shark teeth from the genera Acrodus, Hybodus and Polyacrodus.During the Polish expedition of 1998 a considerable amount of fish remains were collected, amongst them were also shark teeth of the genera Acrodus, Hybodus and Lissodus (Blazejowski, 2004).Romano & Brinkmann (2010) described the newly discovered shark Palaeobates polaris from the Lusitaniadalen Member, Stensiö Fjellet.From the Vendomdalen Member they recognised eight chondrichthyan taxa, where seven were identified from teeth and one from a fin spine.
The primary collection method for shark fossils from Spitsbergen has during 150 years of research probably been surface picking (Stensiö, 1918(Stensiö, , 1921(Stensiö, , 1932;;Birkenmajer & Jerzmanska, 1979), but this method provides only a limited amount of material and understanding.
The material for this present study is the largest sampled from Spitsbergen and was collected from the bonebed in the Grippia niveau (see Hurum et al., 2018).This unique Early Triassic bonebed gives an age constraint for the described taxa and provides a new insight for the worldwide fossil record of sharks.
Neoselachii consists of all living sharks and rays and their fossil ancestors, together with some extinct Mesozoic and Cenozoic groups (Cappetta, 1987(Cappetta, , 2012;;Klug, 2010;Guinot et al., 2012).The first great radiation of neoselachians is assumed to have been in the Rhaetian period, where a transgression leads to shallow epicontinental seas (Cuny & Benton, 1999).Through most of the early history of the group in the Palaeozoic and Triassic, neoselachians are rare and very specialized (Kriwet et al. 2009).They developed different jaw systems whereby the jaw quickly opens and creates suction to draw in its prey.The mouth is placed underneath the head, unlike the hybodonts where the mouth was placed in front (Benton, 2015).Both the hybodonts and neoselachian have the characteristically replacement teeth, while the neoselachian additionally have a three-layered enameloid that is more resistant than the hybodonts single layered enameloid (Klug, 2010;Enault et al., 2015).The triple enameloid units enhance the preservation of neoselachians teeth (Klug, 2010).According to Klug (2008), the first large radiation of neoselachian occurred in Late Triassic-Early Jurassic, approximately 73 million years after their first known occurrence (Klug, 2008).

Triassic chondrichthyans of Svalbard
The PTME in central Spitsbergen is identified by Wignall et al. (1998) and Bond & Wignall (2010) to be immediately above the base of the Vardebukta Formation.The latter provides evidence of anoxic conditions during the transition to the Triassic, as Erwin (1998) also suggested in his three phases hypothesis.This

Geological setting
The Spitsbergen Mesozoic Research Group's (SMRG) expedition of 2015, concentrated on the early Triassic marine deposits in the Isfjorden area.The chondrichthyans described in this study are all from the Grippia bonebed in the Vendomdalen Member, Vikinghøgda Formation, Sassendalen Group at Marmierfjellet, Spitsbergen, for details on the geology see Hurum et al. (2018) and Hansen et al. (2018).Wiman (1910) identified three vertebrate levels in the Sassendalen Group and labelled them the "Fish niveau", the "Lower Saurian niveau" and the "Upper Saurian niveau" based on their fossil contents.Wiman labelled a fourth layer, the Grippia niveau in 1928, which was first referred to by Stensiö (1921: XXV) as "a bonebed 33 m above the Fish horizon".The name is derived from the ichthyopterygian Grippia longirostris (Wiman, 1928).It also consists of ammonoids and is assumed to be an endemic fauna (Frebold, 1930).Maxwell & Kear (2013) connected the Grippia niveau to the Vendomdalen Mb.According to Romano & Brinkmann (2010), the Early Triassic chondrichthyan material was mainly collected from bonebeds under and above the Fish Niveau at Spitsbergen.
The Grippia bonebed excavated for this present study is located approximately 17 m above the Fish niveau and 25 m below the Lower Saurian niveau at Marmierfjellet.It is roughly 4-5 cm thick and consists mainly of bones, teeth and coprolites.For a stratigraphic log see Hurum et al. (2018), for detailed log and age discussion see Ekeheien et al. (2018) and Hansen et al. (2018).The material has a grey to black coloured shale matrix with yellow dolomite nodules, bones, teeth and coprolites scattered throughout.Due to the colour and moisture of the material, the different components (except the yellow nodules) are difficult to distinguish.The material is fragile, porous and breaks easily.The fragmentary fossils indicate active ocean currents that have reworked the material repeatedly, which according to Blazejowski (2004), results in a marine bonebed with disarticulated fossils.
The material analyzed is phosphatized, and X-Ray Diffraction (XRD) shows that it consists predominantly of apatite, however pyrite and dolomite/anchorite are also present.The pyrite content indicates anoxic water conditions.The presence of a tooth plate of a Ceratodus sp.(lungfish) among the material is intriguing.The genus is today limited to brackish and freshwater (Berra, 2001), and post-Palaeozoic lungfishes are mostly found in freshwater deposits (Schultze, 2004).However, he noted that Middle Triassic occurrences of lungfish are known from marine deposits.This could indicate a possible process where sediments from coastal areas have been transported to open ocean areas.This Grippia bonebed is, in many ways, comparable to the Late Triassic Rhaetic bonebed from Devon, UK, which is a result of the Rhaetic transgression (Allard et al., 2015, Korneisel et al., 2015).The age determination of the stratigraphic log is based on ammonoids, as they provide some of the best dating from marine deposits (Hounslow et al., 2008).The biostratigraphic data are correlated with the data from the Vikinghøgda ammonoid zones in Mørk et al. (1999) and the palynology zones based on Vigran et al. (2014).
Approximately 250-300 kg of material was obtained from the Grippia bonebed.From the material were 552 shark teeth, 3 fin spines and 1 cephalic spine identified.All identified specimens in this study are in the collection at the Natural History Museum, University of Oslo.

Methods
The size of the collection area was approximately 7 m 2 and 5-7 cm deep.Some of the larger samples were separated and individually wrapped in aluminium foil for protection.A few of the more fragile specimens were placed in small field jackets (e.g., Hybodus rapax PMO 230.098).
The bulk material was sieved with different fractions.In this study, only the material from the 2 mm (largest) fraction where examined, but the other sieving samples were preserved for later micropaleontological research.Caution should be taken considering the amount of the unidentified teeth in the smaller fraction not counted.
Sieving could not remove all of the clay and minerals from the teeth so the material in this study was further treated in an ultrasonic bath, Bandelin Sonorex RK 255 Transistor, for five minutes to remove most of the adhering clay.The material consists mostly of chondrichthyan teeth.Most of these teeth were fragmented, probably a combination of preservation, permafrost, excavation, transportation, and extraction by sieving.With the aid of a binocular microscope fragments were identified, picked using tweezers, and successfully assembled and glued together using PaleoBond (type Jurassic Gel 4540) and a hardener to activate the glue (Loctite 7452).
analysis.Two teeth of Hybodus rapax and one tooth of Acrodus spitzbergensis, Acrodus scaber, Polacrodus sp., Synechodontiformes sp.(1), Synechodontiformes sp.(2) were studied using the SEM.First the section was studied with a Hitachi S-3600 N Scanning Electron Microscope Energy Dispersive X-ray using an acceleration voltage between 20 -25 kV.Both low and high vacuum studies were performed.The sections were polished with fine sandpaper before being etched with diluted hydrochloric acid (HCl) at 10% for 10-120 seconds, and then coated in a gold alloy using 'Fine Coat-Ion Sputter JFC-1100' .The gold alloy was coated 2 to 4 times for 7 min on each 90 degree angle of the analysis disk.
Due to difficulties of assigning the teeth to a right or left transverse row in the jaw, most of the direction given in the description is based on the different labial, lingual or occlusal view.The terminology of morphological features in shark teeth follows that of Duffin (1985) and Koot (2013) and a morphological key is given in Fig. 1.Fig. 2 explains the enameloid histology of chondrichthyan teeth.

Institutional abbreviations:
PMO -Palaeontological Museum Oslo (Natural History Museum, University of Oslo).

Systematic palaeontology
For the measurements of the described teeth, see Table 1.Stensiö, 1918 Figure 3 Material: 262 identified teeth, 258 isolated and four partly in matrix.The teeth size ranges with a mesiodistal length from 3.1 mm to 18.4 mm and a linguo-labial width of 1.2 mm to 7.5 mm.Most of the teeth have broken cusps and only three of the 262 teeth have the central cusp intact.

Hybodus sasseniensis
The teeth are sorted into five morphotypes, which are given in Fig. 3  not been described for H. sasseniensis at Spitsbergen.Blazejowski (2004) described teeth with one to three lateral cusplets, which were mostly located on one side of the crown.Birkenmajer & Jerzmanska (1979) observed teeth with one to four lateral cusplets while Stensiö (1921) found teeth with up to six lateral cusplets.The tooth with six lateral cusplets (Stensiö, 1921,  The roots are generally apico-basally narrow and in addition usually fragmented with several disorderly distributed irregular foramina and nutrient canals as seen in PMO 230.103.The base of the root is bent in apical direction that results in a slightly apico-lingually tilt of the sulcus, except for PMO 230.102where the root base is straight. PMO 230.101 (Fig. 3 E1-4) represents one tooth of a group of 18 more massive teeth than the remaining 244 "regular" teeth.There is no gradual size distribution between the "regular" and massive group of teeth, which is unusual.They differ from other "regular" teeth (e.g., Stensiö, 1921;Birkenmajer & Jerzmanska, 1979;Blazejowski, 2004)   The vertical striations are coarse and issues from the linguo-labial groove in apical direction to the occlusal crest.By the central cusp, the striations continue apically to the cusp base.The labial striations do not encounter the occlusal crest.Rest of the crown is smooth with a shiny surface.PMO 230.098 has four lateral cusplets (Fig. 4 A1) oriented at the mesiodistal ends and are paired on each side, the first distal lateral cusplet is intact with an extended cusp similar to the central cusp, but blunter, and the second is smaller and more rounded.
The thick linguo-labial broadness of the root is well defined (Fig. 4 A, B, C-3) and the crown is positioned in the labial root area.The central part of the root is apico-basally highest right beneath the central cusp and the labial root is apically inclined.The apical root has a distinct edge that extends almost from the mesial to the distal end and apically the root ends in a concave orientation parallel to the labial groove.The labial groove is apico-basally high and concave, which gives a short interval between the root and crown (Fig. 4A).The lingual groove to PMO 230.100 is nearly horizontal due to the root plateau extended as a line from the mesial to the distal end (Fig. 4 C1).The foramina are distributed irregular in the root and a labial nutrient groove is positioned beneath the centreline of the inclined cusp, as in Stensiö teeth (1921:  Hybodus ?microdusStensiö, 1921 Figure 5 Material: Five isolated teeth identified.The teeth have a central cusp and several lateral cusplets.They range from a mesiodistal length of 8.9 to 11.3 mm and linguo-labial width of 3.2 to 4.3 mm.None of the apico-basal heights is intact.The teeth are much larger than for the H. microdus described before in e.g., Stensiö (1921), Birkenmajer & Jerzmanska (1979).
Description: PMO 230.106 (Fig. 5 A1-4) has a central cusp with two lateral cusplets on each mesiodistal side (A1).The occlusal crest continues from the mesiodistal end to the fracture surfaces of the cusp and cusplets, before it initiates on the opposite side of the fracture surface.On the linguo-labial crown shoulder there are oriented vertical striations creating small furrows.The linguo-labial grooves are apical-concave, the same as the apical root (Fig. 5 A1-2).Both linguo-labial roots have the same apico-basal height, similar to Acrodus.The labial root is filled with irregular foramina disorderly distributed (A1) while the lingual root has smaller and more rounded foramina (A2).
Polyacrodus sp.Jaekel, 1889 Figure 6 Material: Three identified teeth.Only one tooth has the mesiodistal length preserved, at 13.3 mm (PMO 230.130), while the apico-basal height range between 4.6 to 6.1 mm and the linguo-labial width range between 2.5 to 3.7 mm.P. sp. has a linguo-labial broad, relatively high and blunt central cusp with a lingual peg located at the basal cusp.
The teeth have apical high and massive roots, as Acrodus, and a crown with brittle-shape cusplets, more like the genus Hybodus.The lateral cusplets vary from three to four on each side of the central cusp.(Stensiö, 1921, plate 1, fig. 21-26) has the same lingual peg with approximately the same linguo-labial thickness compared to their mesiodistal length, but the crown has no lateral cusplets.The teeth described by Stensiö (1921) are significantly smaller (linguo-labial length of 1.5-2.0mm and apicobasal height of 0.5-0.75mm) compared to the teeth described here.While the Polyacrodus polycyphus Agassiz, 1837 described by Diedrich (2009) have similar size and contains lateral cusplets.
Towards the posterior position in the jaw the number of lateral cusplets increase, roots are flatter and become more linguo-labially wide (Diedrich, 2009).PMO 230.130 has an extra pair of lateral cusplets and is linguo-labially broader, which may indicate a more posterior position than PMO 230.131.The occlusal crest is continous from the mesial to the distal direction with skewness from the center of the tooth crown towards the labial side, and is intact through the crown.This results in a crown that is broader lingually than labially.The crown shoulders have ornamentation, formed as a mesh pattern, and located on the edge before the crown angles inwards into the linguo-labial grooves.Linguo-labial grooves are deep.The root are apico-basally high, massive, but narrower than the crown, and with a slight concavity as in the linguo-labial groove and basal crown.The lingual root is apico-basally higher than the labial and both have small irregular foramina located disorderly.
Acrodus scaber Stensiö, 1921 Figure 7 Material: 41 isolated teeth well preserved and many of them complete.The teeth identified are flat, elongated and ranging from a maximal mesiodistal length of 11.2 to 31.1 mm, apico-basal height of 4.8 to 9.7 mm and linguo-labial width of 3.6 mm to 8.2 mm.The root is significantly apico-basal higher than the crown in most of the identified specimens, similar to Stensiö's (1921) observations of A. scaber.A characteristic feature for this species compared to A. spitzbergensis is the radiating striations from the central part of tooth in mesiodistal and apico-basal direction to the linguo-labial grooves, in a fan-like shape.
The tooth is rather apico-basally flat, mesiodistally elongated and fluctuates apico-basally.The apically highest part of the crown is subrounded and slightly mesiodistally bent in the vertical plane, opposite to that of A. spitzbergensis, which has a more angular central crown.PMO 230.133 (Fig. 7 D1-4) has a small rounded elevation in the crown centre that indicates the middle part of the tooth.The crown differs from PMO 230.112,PMO 230.113 (Fig. 7 B1-4) and PMO 230.114 by being more massive, with rounded appearance and having linguo-labial fluctuations (Fig. 7 D3).PMO 230.114 (Fig. 7 C1-4) has a significant mesial offset crown peak to the left in labial view, which indicates a lateral to posterior position in a jaw.The distal end of the tooth is lingual bent, which results in a slight labial concavity of the tooth (Fig. 7   Acrodus oppenheimeri Stensiö, 1921 Figure 9 Material: 16 teeth, well preserved and complete (except one tooth missing the root).15 teeth are isolated and one is situated in a dolomitic matrix.The identified teeth are elongated with a large crushing surface.They range between a mesiodistal length of 3.1 to 11.5 mm, apicobasal height of 2.7 to 5.6 mm and linguo-labial width of 1.6 to 3.9 mm.The mesiodistal crown surface is approximately flat or slightly arching apically without any cusps/cusplets.The species resembles the A. gaillardoti in this study (PMO 230.120), but the distinct occlusal crest and the lack of linguo-labial crown bulging separate the species.
Description: A significant straight occlusal crest extends from the mesial to distal ends, but does not encounter the linguo-labial grooves.The vertical striations radiate apically in branches from the linguo-labial grooves to the occlusal crest.There are some short vertical striations located throughout the occlusal crest.The striations in the lingual crown centre, PMO 230.118 (Fig. 9 A1-4), (A2) are forming small circling structures as may indicate the central crushing part of the tooth.This area has slightly more elevated and coarser striations than the rest of the crown.The linguo-labial grooves have the same apical arching as the crown surface and they are linguo-labial deep.Both root and crown vary little in height throughout the mesiodistal length of the tooth.The root has randomly placed irregular foramina and the lingual sulcus is somewhat labial-concave, which results in a basal narrowing of the lingual root.
One of the smallest teeth identified of this species is PMO 230.119 (Fig. 9 B1-4).The labial crown is slightly broader than the lingual (Fig. 9    Acrodus sp.
Figure 12 Material: Five isolated teeth, two of them nearly intact.The crowns are flat, oval, elongated and with no ornamentation or striations on the surfaces.Similarities to the species A. lateralis, but due to the flat and unornamented crown we speculate whether or not they are the same species.The identified teeth are small, elongated and ranging from an apico-basal height at 1.9 mm to 3.9 and a linguo-labial width of 1.5 mm to 2.6 mm.Only two teeth have intact mesiodistal length, PMO 230.121 and PMO 230.122 with the lengths of 9.8 mm and 7.4 mm respectively.

Description:
The crown is very brittle with several disorderly oriented fractures.PMO 230.121 (Fig. 12 A1-4) has a bulging labial crown with a slight skewness of the maximal linguo-labial width and apico-basal Palaeobates polaris Stensiö, 1921 Figure 13 Material: Three isolated teeth.The teeth are small with an angular central crown, and coarse striations fill the entire crown.Only two teeth have the mesiodistal length preserved, 6.3 and 7.3 mm, while the apico-basal heigth range from 2.4 mm to 3.7 mm and linguo-labial width range from 2.1 mm to 2.4 mm.Romano & Brinkmann (2010) presented a Palaeobates polaris model were they reconstructed the lower jaw and palatoquadrate dentition from material preserved in situ.The dentition files were divided in numbers from anterior to posterior position, respectively from M0 to M7 and P0-P6.The teeth described in this study belongs to the anterior file P1 (Romano & Brinkmann, 2010, fig. 3. C4).

Description:
The mesiodistal crown surface to PMO 230.134 (Fig. 13 A1-4) is apical angled towards the crown top (A1) and is narrower in the mesiodistal ends than in the central crown (A3), similar to Stensiö (1921 to the occlusal and transverse crests.Both the linguolabial grooves are distinct (Fig. 13 A1 & 2).The labial root is significantly apico-basally higher than the lingual root (Fig. 13 A1 & 2), with several irregular foramina, and the sulcus is labially concave, which results in a basal narrowing of the root.
Description: PMO 230.125 (Fig. 14 A1-4) is a small and nearly intact tooth with a labial tilting of the crown, but still larger than the teeth described by Stensiö (1921) that has a mesiodistal length of 1.5-2 mm and apicobasal height of 0.5-0.75mm.The teeth described by Blazejowski (2004) have similar size, with a mesiodistal length up to 7 mm.The crown has a large central cusp with distinct vertical and occlusal crests, which apical meets the tip in a sharp point (Fig. 14 A3).The occlusal crest extends nearly from the linguo-labial groove in mesiodistal direction through the apical peak of each cusp (Fig. 14 A1 & A4).The central cusp has a lingual swelling named the lingual peg with the transverse crest radiating through the center and down to the longitudinal crest.The longitudinal crest is distinct and extends mesiodistally.There are a few lingual striations on the crown while the rest of the crown is smooth.One lateral cusplet is located on each side of the central cusp.This differs from the teeth described by Blazejowski (2004) that have incipient cusplets rather than welldeveloped cusplets.The lateral cusplet may indicate an anterior position in the jaw (Blazejowski, 2004).The crown is linguo-labially thicker to the right in labial view (Fig. 14 A1) than to the left where there is less swelling of the cusplet (Fig. 14 A3).The mesiodistal crown shoulders are rounded and pinch out beyond the mesiodistal root.The height of the root is decreasing to the left in labial view (Fig. 14 A1) and the basal root is flat.Irregular foramina are disorderly distributed in the root.
Hybodontoidea basal plate of a cephalic spine Figure 15 Material: One fractured basal plate of a cephalic spine (PMO 230.142).
Description: PMO 230.142 (Fig. 15 A1-4) has both the central and lateral lobe preserved; the rest of the basal plate is missing along with the spine.The basal plate has a mesiodistal length of 24.6 mm, apico-basal height of 16.4 mm and a linguo-labial width of 9.5 mm.The spine attachment surface is identifiable as a distinct scar on the top of the basal plate (Fig. 15 A2).The scar is significantly larger than in the basal plate found by Stensiö  surface, there are deep grooves oriented in vertical direction.The lateral lobe end is rounded and slightly bent in apical direction.It has two large foramina near the lateral margin indentation.The rest of the basal plate is filled with small irregular to semi rounded pores.The plate is, with some exception, covered in black enamel (Fig. 15 A3).Because of the disarticulated preservation, the species of the basal plate is difficult to determine.
Hybodontoidea fin spine (1) Figure 16 Material: One almost complete fin spine and one apical fragment.
Description: PMO 230.139 (Fig. 16 A1-6) is elongated and gently curved posteriorly and right laterally with a mesiodistal length of 51.7 mm, apico-basal height of 12.9 mm and an anterior broadness of 6.6 mm.The spine is an distal part of a fin spine with a missing apical tip and basal part.In cross-section, the spine is rounded to eggshaped (Fig. 16 A6).The wideness narrows significant from the basal to apical side (Fig. 16 A1 & 2).Both lateral and anterior surfaces have coarse longitudinal striations (better referred to as ribs) distributed in regular intervals and covered with black enamel (Fig. 16 A1, 2 & 3).The ribs and the surface have approximately the same broadness and extend between them from the basal to apical part of the fin spine.The ribs meet in the apical top edge due to the narrowing in apically direction.Blunt, broken, denticles occur on the posterior edge with one series oriented in a footprint pattern (Fig. 16
Description: PMO 230.140 (Fig. 17 A1-6) is a massive fin spine, gently curved both laterally and posteriorly towards the apical tip.The spine has a mesiodistal length of 35.2 mm, a lateral width of 19.5 mm and an anterior broadness of 9.7 mm.This spine is probably an upper middle part, an observation supported by the missing apical tip (Fig. 17 A1) and the start of the posterior basal opening (Fig. 17 A4).The cross-section of the spine has a rounded to diamond-shape that is more longitudinal than latitudinal (Fig. 17  striations.The right-lateral side has almost entirely lost the enamel and is covered with matrix (Fig. 17 A2).Two thirds of the posterior wall consists of the same gentle vertical striations as the lateral sides while the last part starts with the basal opening (Fig. 17 A4), which is deep with thick lateral walls extending on each side.The anterior wall has eight blunt denticles, where six of them are oriented in a vertical line (Fig. 17 A3) and covered with black enamel.Stensiö (1921) by being smaller, with different orientation of the denticles and by having a posterior basal opening.The presence or absence of the basal opening could be an outcome of differences between first and second dorsal fin spine (Cione et al., 2002).
Synechodontiformes Duffin & Ward, 1993 Synechodontiformes are placed in the Neoselachii as they have three layered enameloid (SCE and BCE units), and teeth that are of the cutting-clutching or cutting sensu stricto type, whereas the hybodontiformes rarely have cutting dentition (Cappetta, 1987) and mostly only one layer (SCE) enameloid (Enault et al., 2015).

Synechodontiformes sp. (1)
Figure 18 Material: 36 isolated teeth.Neoselachian teeth have not been described from Spitsbergen before.Most of the teeth are fragmented with only the central part preserved, the teeth have a dominant and high central cusp with no or few lateral cusplets.The crown is high while the root is less dominant.They range from a mesiodistal length of 7.2 to 11.3 mm, apico-basal height of 3.2 to 7.3 mm and linguo-labial width of 1.6 to 3.9 mm.

Description:
The central cusp of PMO 230.128 is located in the mesiodistal centre, while the central cusp is bent to the right in mesiodistal view (Fig. 18 B-D1) for PMO 230.110-1 and PMO 230.129 (Fig. 18 D1-3) and slightly lingually.The crown surface is smooth and unornamented except for the occlusal crest and a few transverse crests radiating apically from the linguolabial grooves.The occlusal crest extends from mesial and distal ends to the apex of the central cusp and forms a sharp cutting edge, which is apically broader at the base of the central cusp base.PMO 230.128 (Fig. 18 A1-4) has an apico-basal high central cusp with a conical shape and two broken lateral cusplets on each side, and thus resembles the teeth of Mucrovenator minimus (Cuny et al., 2001).The labial crown shoulder is distinct with a longitudinal sharp edge, which is rounded in the mesiodistal ends (Fig. 18 A1).It differs from Mucrovenator minimus that has no groove separating the crown and root.an interval between crown and root while the lingual is more concave and appears as a line.

Synechodontiformes sp. (2)
Figure 19 Material: 13 teeth, 12 isolated and one in matrix.This tooth morphology has not previously been described from Spitsbergen.The teeth range between a mesiodistal length of 4.7 to 11.2 mm, apico-basal height of 1.5 to 4.6 mm and linguo-labial width of 1.1 to 3.  cusp, more elongated cutting edge and higher apicobasal root.

Description:
The central mesiodistal crown meets in a cutting edge with both the cusp and cusplets.The crown has a triangulated shape in cross-section (Fig. 19 A-C4).PMO 230.127 (Fig. 19 A1-4) has an elevated central cusp with five small lateral cusplets on each side.PMO 230.126 (Fig. 19 B1-4) is similar to PMO 230.127, but has a crown with a less distinct central cusp, smaller cusplets and more mesiodistal elongation.To the left of the central cusp there are two intact lateral cusplets (smaller version) of the cusplets in PMO 230.127, while to the right there is 10 to 11 lateral cusplets (Fig. 19 B1).The right cusplets look more like small elevations of the crown instead of separate cusplets, better defined as sub-cusplets.PMO 230.117 (Fig. 19 C1-4) differs from the others by having a pointier and more distinct cusp, resembling slightly the A. spitzbergensis central cusp.The occlusal crest emerges from the mesiodistal crown shoulders and continues throughout the cutting edge as a sharp line.Striations radiates apically from the linguo-labial grooves on the crown shoulder, creating furrows.Fewer vertical striations continue apically to the apexes of the cusp and cusplets.The crown is linguo-labial narrower than the root, and is oriented slightly labially.The linguolabial grooves are distinct, shallow, forming an apicobasal interval between the basal crown and apical root, and appear as a smooth black area.In the root beneath the central cusp, there are large irregular foramina that extend from the basal root to the labial groove.The lingual root is short with a distinct basal edge that forms the start of the sulcus, which are angled approximately 45 degrees from the cutting edge (A-C 4).

Description:
The specimen PMO 230.141 (Fig. 20 A1-5) is a small spine fragment assumed to be the middle section of a fin spine based on the Nemacanthus splendens described by Böttcher (2015).
The fin spine has a mesiodistal length of 13.5 mm, a lateral width of 7.8 mm and an anterior broadness of 3.3 mm.In cross section, the fin spine has an oval shape and a possible nutrient canal is oriented in the centre (Fig. 20 A4 & 5).The anterior ridge is defined by distinct, dark and smooth enamel with some surface fractures (Fig. 20  A3).The lateral side has up to five longitudinal rows of small tubercles oriented vertically parallel to each other and is continuous basally from apical fracturing point to the angled start of the basal fin spine.The longest rows have nine to ten tubercles.The tubercles are oval to sub rounded and consist of black enamel which slightly extends from the surface (Fig. 20 A1 & 2).Some of the tubercles are missing the black surface enamel and are then either grey or a mix of grey and black.The basal part of the fragments are characterized by an angular grey surface tilting towards the anterior side (Fig. 20 A1) with vertical striations oriented parallel to the black anterior ridge, which is similar to the Nemacanthus splendens in Böttcher (2015, fig. 8.12 a-d).The angular grey surface is not present in the remodelling of Nemacanthus monilifer in Storrs (1994, fig.3J), which may exclude PMO 230.141 affinities to that species.The fin spine resembles the "Generically indeterminable selachian; fin spine" in Stensiö (1921,

Enameloid microstructure
The chondrichthyan tooth enamel is characterized by two units.The outer homogenous Single Crystallite Enameloid (SCE), where the crystallites have no preferred orientation, and the inner Bundled Crystallite Enameloid (BCE), where the crystallites forms bundles (Enault et al., 2015).The BCE unit is further divided in three different components: Parallel Bundled Enameloid (PBE), Tangled Bundled Enameloid (TBE) and Radial Bundled Enameloid (RBE), where their terminology is based on the crystallites orientation (Enault et al., 2015;and Fig. 2 for nomenclature explanation).
The teeth are less acid resistant near the center part of tooth, which results in the inner surface to be more affected by the acid than the outer surface.
Acrodus spitzbergensis (PMO 230.334,Fig. 21 A1-3): The enamel average range in apico-basal height from 130 µm at the linguo-labial sides to 280 µm at the occlusal crest and is recognizable by two layers of SCE.The first layer consist of homogeneous crystallites with none preferred orientation and the second is defined by canals mostly elongated perpendicular to the dentine boundary, most likely due to foramina (dental tubules) (Fig. 21 A2 & A3).These canals are mainly located from the linguo-labial crown shoulder apical to the crown top.The contact between the dentine and enameloid is unclear due to the intermingled tissues (Fig. 21 A2).
Acrodus scaber (PMO 230.109,Fig. 21 B1-3): The enamel comprises two layers of SCE and range in apico-basal height from 210 to 350 µm, separated by different orientation of the crystallites (Fig. 21 B2 & 3).The layers are oriented as those in A. spitzbergensis, however, the boundary between the dentine and enameloid is more distinct (Fig. 21 B2).It has thick enamel, compared to A. scaber, throughout the whole cross section with a slight narrowing towards the linguo-labial grooves.
Polyacrodus sp.(PMO 230.131,Fig. 21 C1-3): The enamel is comprised of two layer of SCE, as in A. scaber and A. spitzbergensis, with an apico-basal range of 54 to 97 µm and oriented in the same way as the Acrodus teeth (Fig. 21 B1-3).The outer homogenous layer is thicker than the inner layer with perpendicular oriented crystallites at the linguo-labial crown sides while at the occlusal crest the inner layer expand to almost the same apico-basal height, similar to what was observed by Cuny et al. (2001).The cross section of Polyacrodus sp.differs from the other investigated species by having a distinct pulp cavity (Fig. 21 C1).The pulp cavity has been filled by calcite crystallites based on quantitative chemical analyse.parallel oriented crystallites layer towards the dentine surface as in Acrodus, Hybodus and Polyacrodus (Fig. 22 B2 & 3).The SLE layer range in apico-basal height from 98 to 166 µm.In addition, the enamel has RBE extending from the outer SLE approximately perpendicular to the dentine surface (Fig. 22 B2 & 3).This feature is best presented in the untreated cross-section (Fig. 22 B2), while the acid treated surface indicates the transition between the possible PBE layer (~98 µm) and dentine (Fig. 22 B4).
Hybodus rapax (1) (PMO 230.143,Fig. 22 C1 & 2): A linguo-labial cross-section picture was taken before the surface treatment of the labial central cusp were performed (Fig. 22 C1).The cross-section reveals seemingly five different layers, same as in H. rapax (2).The SCE enamel range from approximately 70 to 80 µm and is divided in two, one outer homogenous layer and an inner layer oriented perpendicular to the dentine surface, as in Acrodus and Polyacrodus (Fig. 22 C1 & 2).Remaining layers may be a result of a mixture of orthodentine and osteodentine, as for Lissodus angulatus in Blazejowski (2004).
Hybodus rapax (2) (PMO 230.336,Fig. 22 D1-3): The enameloid consist of a SLE layer, which range from 63 µm by the crown shoulder to 93 µm at the occlusal crest.This cross-section is similar to the H. rapax (1) with the same division of the layers, except an additional not coherent thin pattern between the SLE layer and the dentine (Fig. 22 D2).The cross-section reveals seemingly five different layers, same as in H. rapax (2).The outer layer is divided in to two SLE layers, the first external layer is homogenous and the second internal layer has crystallites oriented perpendicular to the dentine surface, as in Acrodus and Polyacrodus (Fig. 22 D2 & 3).The third to fifth layers may be a result of mixed osteodentin and dentin, or plausibly the BCE unit is present.2 for SEM preparations and parameters.SCE -single crystallite enameloid.

Discussion
In this study, more than 550 chondrichthyan teeth, three fin spines and a cephalic spine were identified.The chondrichthyans identified in the material from the Grippia bonebed collected in 2015 belongs to 7 genera and 15 species.A likely origin for the Grippia bonebed material is a coastal margin, the findings of a fragmented tooth plate of the lungfish Ceratodus and a fragmented Mytilis strengthens this idea.The pyrite present in the Grippia bonebed indicates anoxic to dysoxic water  2 for SEM preparations and parameters.RBE -radial bundled enameloid, SCE -single crystallite enameloid, TBE -tangled bundled enameloid.
conditions (Bond & Wignall, 2010), which could have been developed in a coastal dysoxic area, or, most likely, at the cold deep-water shelf.
H. sasseniensis is previously only known from the Myalina bed(s) (Brevassfjellet) from the older Dienarian substage (Stensiö, 1921), and after this study also from the Spathian substage.The H. sasseniensis teeth described in this study have up to 7 lateral cusplets next to the central cusp (e.g., PMO 230.107), which previously has not been observed in this species from Spitsbergen (e.g., Stensiö, 1921;Birkenmajer & Jerzmanska, 1979;Blazejowski, 2004).According to Stensiö (1921) and Birkenmajer & Jerzmanska (1979), the number of lateral cusplets on the crown depends on the mesiodistal length of the tooth, which also applies for the material in this study.Another observation for H. sasseniensis is that along with the increase of lateral cusplets and mesiodistal length, the linguo-labial grooves becomes more apical-concave.The increase of apical-concavity is perhaps due to a lateral to posterior position in the jaw or as an occasion of mesiodistal elongation.The identified H. sasseniensis material is divided in two groups, based on the size differences.Although the "regular" group has specimens with same mesiodistal length or linguo-labial width, the majority is noteworthy smaller, especially in the latter direction.The 14 teeth described by Birkenmajer & Jerzmanska (1979) range from a mesiodistal length of 1 to 7 mm, which are generally smaller than the "regular" group of H. sasseniensis described in this study.Stensiö (1921) described 6 teeth with mesiodistal length ranging from 6 to 12 mm, which are the same size as the range in the "regular" group, while they are mesiodistally shorter than the massive group of H. sasseniensis.The size distribution in a species is considered to have a gradual distribution, e.g., from anterior to posterior position in a jaw, between upper or lower jaw or from young to adult specimens (Cappetta, 2012).Therefore the size difference between the "regular" and the massive group in this study is highly interesting.One possible conclusion may be that the massive group could belong to another species than H. sasseniensis.However, the major difference is the size while the morphology is the same as the "regular" group.Another interpretation may be that the species rarely grows to large sizes, but when they first manage, they grow to noteworthy sizes.A third explonation is that both the "regular" and massive group of teeth may be found in the same jaw of a H. sasseniensis.To verify any of these interpretations articulated material would have to be obtained.
The presence of Hybodus microdus in this material is questionable.Stensiö (1921) established the species with two main characteristic features; being small and with slight ornamentation.The teeth described herein have ornamentation on the crown and a massive cusp base, similar to Stensiö (1921), but they are not small.The maximal mesiodistal lengths of Stensiö (1921)'s teeth are approximately 1.5 mm, while the teeth in this material range to a mesiodistal length of 11.3 mm.The large difference could be a product of the surface picking conducted by Stensiö (1921), where he collected a small matrix block with imbedded fossils, compared to the larger amount of material collected for this study.The large amount of material may have provided a greater range of teeth; however, smaller teeth could have been preserved in the 1 mm fraction from the sieving of the material that has not yet been studied.The H. microdus specimens described in Birkenmajer & Jerzmanska (1979) has a mesiodistal length up to 3.5 mm.One feature that slightly contradicts the identification of the Hybodus genus is the apico-basal tall root, which resembles more an Acrodus.This morphological feature may indicate a different species or in fact different genus and could be a result of heterodont dentition.
The majority of species in the Hybodus genus is repeatedly thought to be synonyms or morphospecies (Stensiö, 1921;Brinkmann & Jerzmanska, 1979;Rieppel, 1981).Stensiö (1921) regarded the Agassiz splitting of the genus Hybodus into several species as unnecessary, since most of the species should be considered as morphospecies.However, due to the small amount of specimens the differences were still significant (Stensiö, 1921).At the same time, Stensiö (1921)  The striations are coarser than those of A. spitzbergensis, which can also be observed in this material, nevertheless Stensiö (1921) stated that the species resembled each other greatly.These two species are also found in this study and support Stensiö's (1921) observation.However, there may still be morphological differences dividing the species.A. scaber is mesiodistally longer, apico-basally higher and linguo-labially broader than the remaining Acrodus species in this study.However, the tooth size is not enough to fully distinguish the species.The size may vary due to a monognathic or diagnathic heterodonty, sexual dimorphism or even differences between young and adult specimens (Cappetta, 2012).
The Acrodus oppenheimeri specimens collected by Stensiö (1921) was not given a specific location, but are linked to the areas: "Mt.Tschermak, Mt.Saurie and south of Sassen Bay on the mountain between De Geer Valley and Flower Valley" (Stensiö, 1921, p.25).
The Acrodus oppenheimeri in this study differs from the teeth collected by Stensiö (1921, plate 3, fig. 5-11, P.105 and pers.obs.J.B.) by being mesiodistally straighter, smaller and without any lateral cusplets.This is perhaps due to its different position in the jaw rather than to different species, hence the reconstruction of the jaw done by Stensiö (1921), where the posterior teeth are significantly different to the anterior.Birkenmajer & Jermanzka (1979) have described a tooth with the same morphology, but they assign it to A. spitzbergensis rather than an A. oppenheimeri.This does not correspond with the results in this study, proving how difficult it is to determine species according to the striations and shape.
This study is the first to describe Acrodus gaillardoti from Spitsbergen.Diedrich (2009) previously observed the species in the Middle Triassic (Anisian-Ladinian boundary) from the Upper Muschelkalk bonebed, Bissendorf, Germany.Earlier findings of the species also refer to the mainland Europe Muschelkalk unit, e.g France (Woodward, 1889).The species resembles the A. oppenheimeri, but without the distinct occlusal crest and with a linguo-labial crown bulging, which are the characteristic features that separate the species.Delsate & Duffin (1999) stated that A. gaillardoti is significantly larger than A. lateralis, while in this study the mesiodistal length and linguo-labial width between those species are similar.The A. gaillardoti may be a morphospecies of A. spitzbergensis and A. scaber; however, to strengthen this hypothesis, articulated specimens would have to be obtained.
The Acrodus sp.differs from the Acrodus lateralis by a mesiodistal flatter crown without any striations or crests (smooth and shiny surface).In general the A. lateralis has striations extending throughout the crown (Diedrich, 2009;Henz & Hertel, 2012;Böttcher, 2015) and occlusal crest (e.g., Delsate & Duffin, 1999).The crown in Acrodus sp. is more linguo-labial slender than the crown in A. lateralis.However, the Acrodus sp. is not excluded from belonging to A. lateralis in this study.
The presence of only one tooth of Lissodus angulatus does not necessarily indicate that the species is infrequent in the Grippia bonebed; it may have been sieved through the 2 mm fraction and is preserved in the 1 mm fraction material not yet studied.
The hybodontoidea fin and cephalic spines are rarely possible to sign to a specific hybodont genera or species due to only small differences.Unless the spines are found with one type of teeth, the attribution to different genera or species is almost impossible (Cappetta, 2012).
The material from this study is not articulated, which enhances the difficulties in determining the genera or species of the spines.Therefore, the two fin spines and one cephalic spine have been attributed to the superfamily Hybodontoidea.
The neoselachian are viewed as the sister group to the hybodonts (Maisey, 1982(Maisey, , 1984;;Klug, 2010).Due to the hybodonts originating in the Carboniferous, the early origin of the neoselachian is highly possible (Klug, 2010;Andreev & Cuny, 2012).The only previously evidence of neoselachian from Spitsbergen was the discovery of a Nemacanthus fin spine by Stensiö (1921).Parts of similar fin spines are also identified from the material used in this study.(Andreev & Cuny, 2012).However, the Duffinselache holwellensis (Duffin, 1998) differs from PMO 230.117 by not having any vertical striations or lateral sub-cusplets.The Duffinselache holwellensis was previously interpreted as a hybodont genus by Duffin (1998) named "Polyacrodus" holwellensis.Andreev & Cuny (2012) signed the species to genus Duffinselache and stated that the root was rarely preserved among hybodonts.This statement is contradictory to the material from the Grippia bonebed.Almost all of the hybodont teeth have the root preserved.
The presence of RBE and PBE in the enameloid of shark teeth reinforces the present of neoselachian in this study.The BCE unit is not recognized in the enameloid of Acrodus, Hybodus or Polyacrodus in the untreated or acid treated cross-sections.While the RBE are clearly recognized in Synechodontiformes sp.
(1) in the untreated section, the PBE and TBE is more unnoticeable, but this may be due to the fact that in general the TBE layer is poorly developed in primitive Neoselachian (Cuny & Risnes, 2005).Synechodontiformes sp.(2) has clear indications of RBE in the untreated surface and possible a PBE and TBE, as in Synechodontiformes sp.(2), although Cuny & Risnes (2005) states that RBE is poorer developed in the Triassic Synechodontiformes than the more modern neoselachians.The presence of RBE and PBE supports together with the morphological studies the occurrence of Synechodontiformes at Spitsbergen.The weak presence of the delicate BCE units might be due to the preparation of the specimens, which requires being accurate and careful.As stated by Cuny et al. (2001, p 296) -"No tangled-bundled enameloid (TBE) has been found, but this might be due to the difficulty of correctly preparing such minute teeth" The presence of neoselachian in our sample changes the perspective of the shallow epicontinental ecosystem in the Spathian substage.The three different trophic adaptions; grinding-tearing-and cuttingtype found in the assemblage, are indications of the different ways the shark obtained their food.Both sharks with tearing-and cutting-type dentition indicate more active hunters than the sharks with grinding-type dentition.This implies a competitive environment.
The neoselachian teeth are used for scratching and butchering, while the teeth of some hybodonts (e.g., Hybodus) are used for capturing and holding on to their prey.The first great radiation of the neoselachians is usually identified to the Rhaetian, as a result of rapid transgression creating epicontinental seas (Cuny & Benton, 1999;Klug, 2010).The appearance of two different synechodontiforms in the Grippia bonebed could be a hint of an earlier radiation of the neoselachians.
Hybodont teeth from the Early Triassic (Schytian (old terminology)) Induan-Olenekian are found on the Japanese islands, such as Polyacrodus minimus (Agassiz, 1839) and an unidentified Hybodontidae indet.(Goto, 1994).In China, in the Middle to Late Triassic, several hybodont teeth have been identified, such as Polyacrodus contrarius (Chen et al., 2007).These findings have a similar lingual peg as Polyacrodus sp. in this study; however the teeth show several distinct differences such that they can only be related to the same genus.Neoselachian teeth are identified from the Ladinian-Carnian (Chen et al., 2007), but they do not resemble the neoselachian teeth from Spitsbergen with their apical high and sharp lateral cusplets next to a massive central cusp.
The Early Triassic from Oman consists of Hybodontiformes and neoselachian shark teeth, which rarely extends more than 2 mm mesiodistally (Koot, 2013).These teeth are significantly smaller than the observed teeth in the Grippia bonebed, but are similar to the older teeth described by Stensiö (1921) and Birkenmajer & Jerzmanska (1979).The observations may indicate an increase of the chondrichthyan body size during the Early Triassic.The invertebrates were restricted in body size during the earliest of Induan (Twitchett, 2006), which may have influenced the chondrichthyan body size.The Neotethys (Oman, India, Timor and Iran) are, according to Koot (2013), similar in composition of hybodonts and neoselachian sharks.

Conclusions
• This study provides a well preserved chondrichthyan substage from the Early Triassic of Spitsbergen.The Grippia bonebed is the richest discovery of Early Triassic chondrichthyan teeth from Spitsbergen.Of the 15 identified species, wherein eight are new for the Grippia niveau, contributes greatly to the understanding of chondrichthyans after the PTME.The teeth are well preserved and often with the root firmly attached to the crown, which is a rare preservation for hybodont teeth that are usually missing the root.
• The Hybodus sasseniensis have up to 7 lateral cusplets next to the central cusp, which previously has not been found in this species from Spitsbergen.This species is only known from the older Dienarian substage, but now also from the younger Spathian substage.The H. sasseniensis is represented by two groups, "regular" and massive, depending on the size.The massive group consists of the largest identified H. sasseniensis from Spitsbergen.
• In the determination of shark species, the heterodont dentition is frequently overlooked and the occurrence of morphospecies is therefore common.This study raises a question about the close resemblance of Hybodus sasseniensis and H. rapax, Acrodus scaber and A. spitzbergensis and between Acrodus species in general.This is also true for the variations in teeth size of a monognathic or diagnathic heterodont dentition, sexual dimorphism and even between young and adults.
• The presence of Hybodus microdus Stensiö (1921) in the Grippia bonebed is questionable due to the more than 7 times larger mesiodistal length and an apicobasal higher root than the teeth previously described by Stensiö (1921).However, the close resemblance of the morphology of the teeth indicates that they belong to the same species, and that the "lilliput effect" may have influenced the marine life after the PTME.
• This study is the first to describe Acrodus gaillardoti from Spitsbergen, whereas it is previously known from the younger Anisian-Ladinian Upper Muschelkalk bonebed.The species differs from the Acrodus oppenheimeri by the central linguo-labial bulging of the crown and lack the mesiodistally elongated occlusal crest.However, the A. gaillardoti is not excluded from being morphospecies of A. scaber and A. spitzbergensis.
• The neoselachian teeth are a new discovery from Spitsbergen and the material in this study consists of two new species of the Synechodontiformes.With their cutting edges and multiple enameloid layers they differ considerably from the identified Hybodontiformes in the Grippia bonebed.
• The shallow epicontinental to deep water shelf-areas in the Spathian Boreal sea contained a majority of active chondrichthyan hunters with cutting and tearing dentition, and a smaller amount of durophagous chondrichthyans with grinding dentition.The chondrichthyans indicate an endemic and latitudinal restriction in the Smithian -Spathian, also supported by the ammonoids known from the Vikinghøgda Formation.
• To better understand the chondrichthyans survival and diversity after the PTME, the discovery of articulated material is needed.This would give the possibility to differentiate the morphospecies from the true species.

Table 1 .
Measurements of the described and illustrated specimens.* -measured to the fracture point, Na -not available.
as A-E.The crowns are characterized by having a large and dominant central cusp with two to seven lateral cusplets on each side.H. sasseniensis has coarse vertical striations from the labial groove (see Fig.2for explanation) to approximately the middle of the cusp; the rest of the apical cusp is smooth.The striations continue in general to the apex of the cusplets.PMO 230.104 (Fig.3 B1-4) has coarser striation than PMO 230.103 (Fig.3 A1-4), whereas the striation on PMO

Table 2 .
SEM preparation parameters.x -not performed.
Diedrich (2009)t the species H. rapax undeniably resembled the H. longiconus, which Agassiz defined in 1843.Rieppel (1981)described that the H. longiconus resembles the teeth of H. plicatilis, while the H. sasseniensis in this study resembles the teeth of the latter species.The H. plicatilis is characterized by the straight and pointed central cusp, vertical striations apical radiating towards the apex and the root with approximately vertical linguo-labial grooves(Rieppel, 1981).This description for H. plicatilis fits well with the description of H. sasseniensis herein.In 1921, Stensiö established both H. rapax and H. sasseniensis, but with a precaution that with better material the studies may provide a similarity in which case they should be put in the same species.The same point of issue is being taken into account in this study and it is likely that H. rapax and H. sasseniensis belong to the same species.As a result, the difference of H. rapax and H. sasseniensis teeth is unclear and the two species should be considered as one species for which H. sasseniensis Stensiö 1918 has nomenclatural priority.Agassiz, 1837 is described byDiedrich (2009)to have similar size and lateral cusplets.The Polyacrodus genus differs from the Hybodus genus by having the lateral cusplets oriented closer together and with no high or extremely dominant central cusp.Comparing the Polyacrodus to Acrodus genus shows that they have similar apico-basal high root while the Acrodus has no central cusp or lateral cusplets such as the Polyacrodus genus.Based on these observations and the enamel study (which will be discussed later), the specimens are assigned to the Polyacrodus genus and referred to as Polyacrodus sp.
(Cuny et al., 2001)eeth described from Grippia bonebed range up to a mesiodistal length of 11.3 and an apico-basal height of 7.3 mm.The teeth described here are slightly older, (late Olenekian, Spathian) than the M. minimus from the Anisian of Nevada(Cuny et al., 2001).
Cuny et al. (2001) resemble the Mucrovenator minimusCuny et al. (2001)(Anisian, Nevada) by the cutting edge, conical shape of central cusp and apico-basal narrow root.The major difference is, amongst others, the size.The M. minimus teeth described byCuny et al. (2001)range from a mesiodistal length and apico-basal height never exceeding