Evidence of macrophagous teleosaurid crocodylomorphs in the Corallian Group (Oxfordian, Late Jurassic) of the UK

Teleosaurids were a group of semi-aquatic crocodylomorphs with a fossil record that spanned the Jurassic Period. In the UK, abundant specimens are known from the Oxford Clay Formation (OCF, Callovian to lower Oxfordian), but are very rare in the Kimmeridge Clay Formation (KCF, Kimmeridgian to lower Tithonian), despite their abundance in some contemporaneous deposits in continental Europe. Unfortunately, due to the paucity of material from the intermediate ‘Corallian Gap’ (middle to upper Oxfordian), we lack an understanding of how and why teleosaurid taxic abundance and diversity declined from the OCF to the KCF. The recognition of an incomplete teleosaurid lower jaw from the Corallian of Weymouth (Dorset, UK) begins to rectify this. The vertically oriented dentition, blunt tooth apices, intense enamel ornamentation that shifts to an anastomosed pattern apically, and deep reception pits on the dentary unambiguously demonstrates the affinity of this specimen with an unnamed sub-clade of macrophagous/durophagous teleosaurids (‘Steneosaurus’ obtusidens + Machimosaurus). The high symphyseal tooth count allows us to exclude the specimen from M. hugii and M. mosae, but in absence of more diagnostic material we cannot unambiguously assign DORCM G.3939 to a more specific level. Nevertheless, this specimen represents the first mandibular material referable to Teleosauridae from the poorly sampled middle-upper Oxfordian time-span in the UK.


INTRODUCTION
Teleosauridae (Thalattosuchia: Crocodylomorpha) was a group of crocodylomorphs that inhabited lagoonal/coastal environments in the Jurassic (Andrews, 1909;Andrews, 1913;Buffetaut, 1982;Vignaud, 1995;Young et al., 2014a).During the late Middle and Late Jurassic a group of teleosaurids achieved large body sizes, heavily built skulls and blunt dentition indicative of a derived macrophagous feeding habit (Young et al., 2014a;Young et al., 2014b;Young et al., 2015a).This group includes 'Steneosaurus' obtusidens from the Oxford Clay Formation of England (Callovian) and the genus Machimosaurus from the late Oxfordian to Tithonian of Europe and Africa (Sauvage, 1897-98;Hua et al., 1994;Lepage et al., 2008;Martin & Vincent, 2013;Young et al., 2014a;Young et al., 2014b;Young et al., 2015a;Martin, Vincent & Falconnet, 2015).It is worth noting that the presence of Machimosaurus in the upper Oxfordian is based on isolated tooth crowns from France and Portugal, and a partial lower jaw (Musée de la Princerie (Verdun, France), 2007.0.14) from France (Sauvage, 1897-98;Hua, 1996;Lepage et al., 2008;Young et al., 2014a;Young et al., 2014b).None of these specimens was found to be diagnostic at the specific level and they were referred to Machimosaurus sp.based on low tooth count (estimated as lower than 'S.' obtusidens-see 'Discussion') and characters of the posterior mandible, which unfortunately cannot be assessed based upon the available 'S.' obtusidens specimens.
The content of the genus Machimosaurus (sensu Young et al., 2014a) has been questioned by Martin and colleagues (2015).They consider Ma. hugii the only valid species within Machimosaurus (Martin & Vincent, 2013;Martin, Vincent & Falconnet, 2015).However, it should be noted that Martin and colleagues did not actually address the monospecifity within Machimosaurus as they focused on the validity of Ma. buffetauti.They suggest that this taxon is the same as Ma.mosae, and both should be referred to Ma. hugii as originally proposed by Martin & Vincent (2013).Their argument for the synonymy between Ma. buffetauti and Ma.mosae is based on the fact that the diagnoses produced by Young et al. (2014a) and Young et al. (2015b) to distinguish these two taxa would be either accountable for intra-specific variation or by post-mortem deformation on the specimens.We recognise that the specimens have undergone some deformation (as is the norm for fossil specimens); however, we disagree with Martin and colleagues conclusions and consider these differences true morphological traits-perhaps due to differing perception of morphological species.However, Machimosaurus species (sensu Young et al., 2014a) still differ from each other in stratigraphic occurrence, basioccipital apophysis cross-section, relative size and shape of the basioccipital tuberosities, relative size of the paroccipital processes and the expansion at their lateral ends, tooth morphology, tooth enamel surface morphology, and tooth count (exceeding modern crocodylian intra-specific variation Larson, 2013;Brown et al., 2015)).Perhaps more importantly, it should also be noted that the supposed synonymy of Ma. buffetauti and Ma.mosae does not prove the monospecificity of the genus Machimosaurus.Whilst Martin and colleagues appeal to the arguments proposed in Martin & Vincent (2013), the updated diagnosis that Young et al. (2014a) and Young et al. (2015b) proposed for Ma.hugii type species, and the unique tooth enamel morphology of the species (Young et al., 2014b), was not taken into account in Martin, Vincent & Falconnet (2015).Consequently, we do not find Martin, Vincent & Falconnet (2015) arguments to be compelling; as such we assert the taxonomic content of Machimosaurus as described in Young et al. (2014a) and Young et al. (2015b) to be valid and adopted it as phylogenetic framework in this paper.
Nevertheless, after a long debate, recent studies agree in considering Machimosaurus and 'Steneosaurus' obtusidens morphologically distinct taxa (Martin & Vincent, 2013;Young et al., 2014a;Young et al., 2014b;Young et al., 2015a), which we agree with.Working under this assumption means that there is a significant gap separating 'Steneosaurus' obtusidens and the first unambiguous Machimosaurus species.This timespan roughly corresponds with the Oxfordian stage, which in the UK is represented by a series of formations stacked in between the fossil-rich Oxford Clay Formation (OCF) and the Kimmeridge Clay Formation (KCF) (Cope, 2006).The middle-late part of the British Oxfordian has been referred to as 'Corallian Gap' due to the rarity of fossil vertebrates known from this period (Young, 2014).While little fossil material is known from this interval, the 'Corallian Gap' is significant because it marks a period of drastic changes in the sub-Boreal marine faunas.Marine reptiles were particularly affected, and plesiosaur, thalattosuchian and ichthyosaur fossils are rare in the Oxfordian and Corallian beds (Benton & Spencer, 1995) after being much more common in the Callovian beds.Additionally, the OCF marine reptile faunas are very different from those of the KCF in terms of both composition and taxonomy (Young, 2014).Unfortunately, the absence of diagnostic Corallian material hinders our understanding of how and why marine reptile faunas changed so dramatically during this time.
Teleosaurids are no exception to this pattern, as only rare, fragmentary, and mostly non-diagnostic material has been reported from the Oxfordian of the UK (Benton & Spencer, 1995).In contrast to other marine reptile groups, teleosaurids are, in fact, very poorly represented (only two genera) even in the fossil-rich Kimmeridge Clay Formation (Young & Steel, 2014;Young et al., 2014a).The 'Steneosaurus' obtusidens + Machimosaurus sub-clade (Young et al., 2012;Martin & Vincent, 2013;Young, 2014)common in continental Europe formations-is exclusively represented by a handful of teeth in the KCF (Krebs, 1967;Hua et al., 1994;Ruiz-Omeñaca, Piñuela & Garcia-Ramos, 2011;Martin & Vincent, 2013;Young & Steel, 2014;Young et al., 2014a).The paucity of teleosaurid material in the UK is puzzling.It may be a consequence of environmental changes (i.e., the UK becoming a deeper water environment), but this is still a matter of debate that can only be clarified by new discoveries and environmental-diversity studies.
Here we describe a fragmentary but informative teleosaurid specimen (DORCM G.3939) from the Corallian of Weymouth Bay (Dorset) (Figs. 1 and 2).We demonstrate that it belongs to the sub-clade 'Steneosaurus' obtusidens + Machimosaurus and that it shows close affinities to 'Steneosaurus' obtusidens.However, as the genus 'Steneosaurus' is still in need of revision, we are unable to recognize any unambiguous diagnostic features of a particular species, so we refer the specimen to 'Steneosaurus' cf.obtusidens.

BACKGROUND INFORMATION
The specimen, DORCM G.3939, was found at the Nothe Grit (SY30 686 787)corresponding to the south flank of the Nothe Gardens by ''S Winch Esq.''Paul Ensom (assistant Curator at DORCM at the time) physically collected DORCM G.3939, which was then presented to the Dorset County Museum (DORCM) in 1980, where it is currently on display in the Jurassic Gallery.
The location where the specimen was found is in close to the type section of the Nothe Grit, which is the base bed of the Corallian Group in the area (House, 1993).The Nothe Grit Member is a localised geological succession dated to the Oxfordian in South Dorset.It overlies the Oxford Clay Formation and constitutes a succession of thick beds of greyyellow-brown sandstones, often including clay and extensive calcareous concretion (Coe, 1995).It is thought to represent a near-shore subtidal environment (House, 1993).The specimen has never previously been formally described, and is not the Teleosaurus specimen from the same locality that is mentioned in Benton & Spencer (1995).In their review of the British Late Jurassic reptile sites, the authors briefly mentioned a crocodilian from ''Nothe, Weymouth (688788; Teleosaurus from Lower Calcareous Grit)'' (sic in Benton & Spencer, 1995).This specimen is not DORCM G.3939, as the authors were referring to a lower jaw erroneously referred to Teleosaurus (the specimen being pliosaurid) by Newton (1878), and later mentioned by Delair (1958, p. 60).CROCODYLOMORPHA Hay, (1930) (sensu Nesbitt, 2011) THALATTOSUCHIA (Fraas, 1901)

DESCRIPTION AND COMPARISONS
DORCM G.3939 is an almost complete and undistorted lower jaw that preserves the symphyseal region, and is approximately ∼30 cm long (Fig. 2).The anterior tips of both dentaries are incomplete so that the dentary alveoli 1 and 2 (D1, D2) are partially or entirely missing.The rest of the tooth rows continue up to the D19 alveolus on the right side and the D20 alveolus on the left side.Exceptionally, most of the teeth are intact and in situ, and two additional crowns were found associated with the specimen.
The surface details (sutures and ornamentation) are largely obscured by the poor superficial preservation of the specimen (Figs.2A-2D).The bone texture is not smooth but appears sculptured by deep anastomosed grooves similar to other large bodied macrophagous teleosaurid dentaries (Martin & Vincent, 2013;Young et al., 2014a).
The sutures of DORCM G.3939 are difficult to discern, although close examination of the posterior breakage surface reveals the contacts between the splenials and dentary laterally, and between the splenials along the midline (Figs.2E and 2F).The splenialdentary sutures are very close to the midline contacts, indicating that the mandible broke across the splenial anterior process.This can be assessed by looking at the lateral extent of the splenials, which in the posterior-most section of teleosaurid mandibular symphyses make up a considerable part of the entire mandible width.This contribution gradually decreases as the splenials taper anteriorly.When these observation are applied to DORCM G.3939, we can confidently conclude that in this specimen the splenial must have reached at least a few (estimated 3 or 4) alveoli anterior to the D19 alveolus.Unfortunately we cannot be sure of the exact extent of the splenial due to the poor visibility of the sutures in dorsal view.This estimate differs from all Machimosaurus specimens, where the splenial reaches up to the D13 alveolus in Ma. buffetauti, and the D11 alveolus in Ma. mosae (Hua, 1999;Young et al., 2014a) (Table 1).Unfortunately, comparisons cannot be extended to Steneosaurus edwardsi (PETMG R178) and 'S.' obtusidens (NHMUK PV R 3168) because the dorsal surface of the lower jaws is inaccessible in both these specimens (Andrews, 1913;Table 2 in Young et al., 2014a) (Table 1).
The coronoids are not visible on either side, even in cross-section, demonstrating once again that the anterior break of the specimen occurred in front of the anterior-most extent of the coronoids.The importance of this observation is further elaborated upon in the Discussion.(Andrews, 1913;Young et al., 2014a) and direct observations on the specimens.Machimosaurus mosae (Hua, 1999;Young et al., 2014a) 19 16 11 5
edwardsi (PETMG R175, PETMG R178), and S. herberti (OUMNH J1420) are comparatively more robust than those of S. leedsi and My.nasutus, they lack the distinctive surface enamel ornamentation that DORCM G.3939 shares with 'S.' obtusidens and Machimosaurus spp.(Fig. 5).Every DORCM G.3939 tooth is consistently ornamented by parallel, densely packed, high-relief apicobasal ridges that are variable in length (Figs.4A and 5A).This pattern is maintained up to 3/4 of the crown height, where the texture becomes anastomosed (Figs.4A and 5A) (Young et al., 2014a;Young et al., 2014b;Young et al., 2015a).All teeth are distinctively carinated and have denticles that are non-contiguous and poorly detectable without visual aids (Young et al., 2015a); both features are clearly visible only close to the apex (Fig. 4B).Conversely, the presence of carinae is variable in Machimosaurus spp.The crowns of DORCM G.3939 display false and true zyphodonty as shown in 'Steneosaurus' obtusidens (Young et al., 2015a).
The unusually well preserved dentition and lack of distortion in the specimen shows that teeth proceeded in an alternating pattern (Figs. 2A,2C and 4C).Fully grown teeth alternate with smaller crowns, which are likely representative of earlier developmental stages.This alternating pattern has been observed in modern reptiles before, and may be related to tooth replacement rhythm (Edmund, 1969).It would, in fact, guarantee that there is a fully grown crown every two tooth positions, minimising the disadvantage of continuous tooth replacement through life (Edmund, 1969).All the largest tooth apices are worn and polished (Fig. 4C).However, the apices of the smaller crowns are rounded as in Machimosaurus and 'Steneosaurus' obtusidens and not pointed as in all other teleosaurids (Figs.4B and 5) (Andrews, 1913;Lepage et al., 2008;Martin & Vincent, 2013).

DISCUSSION
The rarity of Oxfordian vertebrate fossils means that DORCM G.3939 is a pivotal specimen in understanding the evolution of teleosaurids.This specimen is particularly important because it fills a stratigraphic gap between 'Steneosaurus' obtusidens and the species of Machimosaurus, and-excluding teeth-it is the youngest macrophagous teleosaurid yet known in the UK.The dentition of DORCM G.3939 unambiguously demonstrates that it belongs to the 'Steneosaurus' obtusidens + Machimosaurus sub-clade of macrophagous teleosaurids.This confirms that the macrophagous teleosaurid lineage was still present in the UK at the beginning of the Corallian.This has immediate consequences because neither 'Steneosaurus' obtusidens nor Machimosaurus has been reported from the early Oxfordian.This means that if DORCM G.3939 belongs to Machimosaurus, then the first appearance of this genus should be stretched to the very base of Oxfordian.Conversely, if DORCM G.3939 belongs to 'Steneosaurus' obtusidens, then this specimen represents the latest occurrence of this taxon, and the first in the Late Jurassic.
Unfortunately, with the exception of the distinctive Machimosaurus hugii, 'Steneosaurus' obtusidens and Machimosaurus teeth are virtually indistinguishable (Young et al., 2014a;Young et al., 2014b;Young et al., 2015a), so tooth morphology is insufficient to diagnose DORCM G.3939 at the specific level.However, although DORCM G.3939 is an incomplete jaw and has no unambiguous autapomorphies, what is preserved provides useful information that can help pin down its systematic affinities.The mandibular morphology is enough to demonstrate that DORCM G.3939 is more similar to 'Steneosaurus' obtusidens than to Machimosaurus buffetauti.This is based on mandibular symphysis tooth count and morphological features of the mandible.
First, DORCM G.3939 has a high tooth count.The DORCM G.3939 mandibular symphysis is incomplete, and thus the tooth count of 19-20 is a conservative estimate that is likely to be higher.We estimated ∼28 dentary teeth for this specimen based on the splenial and post symphyseal tooth count in M. buffetauti (SMNS 91315) and Machimosaurus sp.(Musée de la Princerie 2007.0.14) (Table 1).Importantly, macrophagous teleosaurids show a trend of decreasing snout length and tooth count.This trend starts in the Callovian 'Steneosaurus' obtusidens (∼29 dentary teeth; Andrews, 1913) and continues through the Kimmeridgian-Tithonian from the oldest Machimosaurus species (M.buffetauti-lower Kimmeridgian) to the youngest (M.hugii-Kimmeridgian to lower Tithonian) (Table 1) (Martin & Vincent, 2013;Young et al., 2014a;Young et al., 2015a) 1).However, we can estimate ∼24 dentary teeth for Musée de la Princerie 2007.0.14, based on the symphyseal tooth count relative to the splenial in M. buffetauti (SMNS 91315), the Machimosaurus specimens with the highest tooth count in the genus.This estimate is comparatively lower than our estimate for DORCM G.3939.
Finally, additional evidence supporting close affinity between DORCM G.3939 and 'Steneosaurus' obtusidens comes from the structure of the lower jaw in dorsal view.The posterior end of the teleosaurid mandibular symphysis is marked by a noticeable lateral enlargement of the mandible, due to the divergence of the mandibular rami.This structure can be seen in all teleosaurid specimens (Andrews, 1913;Lepage et al., 2008;Martin & Vincent, 2013;Young et al., 2014a).Critically, such enlargement is not visible in DORCM G.3939.There are two possible explanations for the absence of this feature: (a) such enlargement is absent in DORCM G.3939, or (b) DORCM G.3939 broke anteriorly to the enlargement and therefore the preserved fossil does not show it.We consider explanation (a) unlikely as a lateral enlargement of the rami is consistently shown in all teleosaurids, including well preserved, three-dimensional specimens.The absence of the coronoid and the position of the break along the splenials are compatible with explanation (b).If correct, this corroborates that the symphyseal tooth count in DORCM G.3939 was higher than 19-20 (∼28 estimated dentary teeth), and thus compatible with a 'long-snouted' blunt-toothed teleosaurid: 'Steneosaurus' obtusidens.
Multiple lines of evidence demonstrate that DORCM G.3939 is more similar to 'Steneosaurus' obtusidens than it is to M. buffetauti.However, we cannot currently recognize any unequivocal synapomorphies that link DORCM G.3939 to 'S.' obtusidens, mostly because the genus Steneosaurus is in need of revision.In the absence of unambiguous diagnostic features we conclude that DORCM G.3939 should be provisionally referred to 'Steneosaurus' cf.obtusidens.Further study of the 'Steneosaurus' obtusidens holotype and new discoveries from the Corallian Gap are needed to confirm this assignment.
Interestingly, the Corallian Nothe Grit Member was deposited under near-shore tidal conditions.The discovery of a macrophagous teleosaurid within such an environmental setting is not surprising, but the absence of piscivorous teleosaurids is.Among teleosaurids, only the macrophagous Machimosaurus is hypothesised to have been able to venture into open seas (Young et al., 2014a).Conversely, excluding rare exceptions, piscivorous teleosaurids have mainly been found in lagoonal-coastal environments.The paucity of teleosaurids in the KCF of the UK may be environmental, as this unit was deposited in deeper water than the OCF (Gallois, 2004).It is, however, surprising that the only teleosaurid material within the 'Corallian Gap' of the UK is currently represented by macrophagous teleosaurid teeth.A possible explanation may be that, despite similar water depth conditions, the Oxfordian environment would have been very different from the nutrient rich, shallow and warm Callovian seas.However, there is another possibility.The Corallian is less exposed and sampled compared to the industrially-exploited OCF and KCF.The rarity of marine reptile material may simply be a consequence of sampling bias.Support for this statement comes from the report of abundant crocodilian teeth from Smallmouth Sands (lowermost Kimmeridgian) (Young & Steel, 2014), the shallowest deposition depth of the KCF Dorset Succession.It is then possible that teleosaurids became rare in the UK only during the Kimmeridgian, when the British Jurassic seaway transitioned from a coastal to a deep outer-shelf environment (Gallois, 2004).We predict that further studies on the poorly known material from the Corallian Gap will elucidate the time and mode of decline of Teleosauridae in the British Late Jurassic formations.

CONCLUSION
In this paper we described the first mandibular material of Teleosaurid from the Corallian Gap of the UK.The specimen can be referred to a sub-clade of macrophagous teleosaurids including 'Steneosaurus' obtusidens and Machimosaurus based on its possession of unique craniodental characters.However, there are no clear synapomorphies linking it to another member of this sub-clade, or autapomorphies showing that it is a new taxon.Nevertheless, we showed that the mandible geometry and the tooth count (absolute and relative to the splenial) match those of 'Steneosaurus' obtusidens more closely than any Machimosaurus specimens.In absence of unambiguous autapomorphies of 'Steneosaurus' obtusidens or Machimosaurus buffetauti we refer DORCM G.3939 to 'Steneosaurus' cf.obtusidens.Future discoveries of more complete Corallian specimens and the revision of the 'Steneosaurus' obtusidens holotype will clarify if the tooth counts of DORCM G.3939 relative to splenial are indeed compatible with the Callovian teleosaurid.In a broader context, DORCM G.3939 demonstrates that the 'Steneosaurus' obtusidens + Machimosaurus lineage was still present in the British formations at the beginning of the Oxfordian.

Figure 1
Figure 1 Stratigraphic chart and map.Middle-Late Jurassic stratigraphic chart and map of the UK.The red circles show the stratigraphic level and locality where DORCM G.3939 was found.

Table 1 Comparative mandibular tooth count of Callovian-Kimmeridgian teleosaurids from UK formations.
Data derive from available literature

.
Based on the figures reported in Table 1, DORCM G.3939 is more compatible with 'Steneosaurus' obtusidens than other macrophagous teleosaurids.The late Oxfordian Machimosaurus sp.(2007.0.14) does not preserve the anterior part of the mandibular symphysis, and it is then difficult to compare with DORCM G.3939 (Table Sedgwick Museum, Cambridge, United Kingdom