Re-description of two contemporaneous mesorostrine teleosauroids (Crocodylomorpha, Thalattosuchia) from the Bathonian of England, and insights into the early evolution of Machimosaurini

Teleosauroidea was a clade of successful, morphologically diverse ancient crocodylomorphs that were integral in coastal marine/lagoonal environments during the Jurassic. Within Teleosauroidea, the macrophagous/durophagous tribe Machimosaurini evolved specialized feeding strategies (e.g. hypertrophied jaw musculature; blunt, heavily ornamented dentition) and large body sizes (> 7 m), becoming an important component of Middle/Late Jurassic ecosystems. These ocean-dwelling giants are well known from the Callovian (Lemmysuchus) of Europe and the UK, and Kimmeridgian-Tithonian (Machimosaurus) of Europe and northern Africa. There are reports of fragmentary machimosaurin material from the Bathonian of Africa; however, the overall Bathonian teleosauroid material is poorly understood. While multiple specimens were described during the 19 and 20 centuries, little research has been done since. Here we re-describe two historically important Bathonian species from near Oxford, UK. We demonstrate that both ‘Steneosaurus’ larteti and ‘Steneosaurus’ boutilieri are indeed valid taxa and establish neotypes for both species and two new genera, Deslongchampsina and Yvridiosuchus. Our cladistic analysis finds Yvridiosuchus boutilieri as a basal member of Machimosaurini, and Deslongchampsina larteti to be closely related to Steneosaurus heberti. Interestingly, four distinct teleosauroid ecomorphotypes are present in the Bathonian of Europe, and teleosauroid ecomorphological diversity continued throughout the Callovian and Kimmeridgian/Tithonian in Europe and


Introduction
Teleosauroids (Thalattosuchia) were a unique group of distant extinct relatives of modern crocodiles that inhabited marine and brackish ecosystems throughout the Jurassic (Buffetaut, et al., 1981;Buffetaut, 1982;Andrews, 1913;Hua, 1999;Foffa et al., 2015, in press;Johnson et al., 2015Johnson et al., , 2018Martin et al., 2016) and Early Cretaceous (Fanti et al., 2016). This near-global group of ancient crocodylomorphs have often been regarded as the Jurassic marine equivalents of extant gavials, due to many species having an elongate and tubular snout, high tooth count and dorsally directed orbits, which is suggestive of a primarily piscivorous diet (Andrews, 1909(Andrews, , 1913Buffetaut, 1982).
However, within Teleosauroidea, the tribe Machimosaurini is characterised by blunt tooth crowns with serrated carinae and extensive enamel ornamentation; proportionally shorter snouts and lower tooth count; and proportionally anteroposteriorly and mediolaterally enlarged supratemporal fenestrae, all of which suggest a macrophagous-durophagous lifestyle (Eudes-Deslongchamps, 1864, 1867, 1869Andrews, 1909Andrews, , 1913Buffetaut, 1982;Massare, 1987;Hua et al., 1994;Hua & Buffetaut, 1997;Vignaud, 1997;Martin & Vincent, 2013;Young et al., , 2014bYoung et al., , 2015aYoung et al., , 2015bFanti et al., 2016;Jouve et al., 2016;Foffa et al., 2018a). During the Middle Jurassic, machimosaurins were relatively rare compared to other teleosauroids, with isolated tooth crowns and indeterminate material known from the Bathonian of England, France and Morocco Jouve et al., 2016), and Lemmysuchus obtusidens  from the Callovian of England and France (Andrews, 1909(Andrews, , 1913Young et al., 2016;. The Oxfordian is a poorly sampled stage for teleosauroids, but L. cf. obtusidens and Machimosaurus sp. are known from England and France respectively Foffa et al., 2015, complete head with its lower jaw, which has been collected in the Cornbrash of the neighbourhood [of] Oxford (England). To this piece, bearing under the catalogue number vm-2357, the following indication, Crocodilus oxoniensis (Conybeare), was attached, a small note probably by the hand of M. de Blainville and bearing these words: "Found near Oxford in oolitic limestone known as Cornbrash, given by the institution of Bristol, probably one of Honfleur's species"]. In his monograph, E. Eudes-Deslongchamps (1867-69) disagreed with the naming of C. oxoniensis, claiming that he could not find any information pertaining to the description of this species, and that it was therefore invalid. Eudes-Deslongchamps (1867-69 p. 231) also stated the following about the C. oxoniensis 'holotype' (OUMNH J.1401): "c'est donc probablement un simple nom de catalogue donné dans une collection publique, et je pense que dans le cas où l'on viendrait à reconnaître que l'espèce des environs d'Oxford est la même que celle de Longues, on devrait les désigner l'une et l'autre sous le nom de Steneosaurus Boutilieri " ["… [so] it's probably a simple catalogue name given in a public collection, and I think that, in the event that we come to recognize that the species around Oxford is the same as that of Longues [the area where the French 'S.' boutilieri holotype was found], we should designate each under the name of Steneosaurus Boutilieri"].
Since de la Bêche and Conybeare (1821) never explicitly used the species oxoniensis in their work, it could be argued that it is indeed not a valid species name, but rather one only present on specimen labels. According to Article 12.2.7 of the Code of the International Commission on Zoological Nomenclature (ICZN), when publishing names before 1931, "the proposal of a new genus-group name or of a new species-group name in association with an illustration of the taxon being named, or with a bibliographic reference to such an illustration, even if the illustration is contained in a work published before 1758, or in one that is not consistently binominal, or in one that has been suppressed by the Commission (unless the Commission has ruled that the work is to be treated as not having been published [Art. 8.7])" is perceived as valid. De la Bêche and Conybeare (1821) did not designate a new species, nor did they provide an indicative illustration or drawing of the crocodile skull from near Oxford; the name itself was only given on a specimen label. Therefore, the name C.
Eudes-Deslongchamps (1868) incorporated material from both the French 'holotype' and OUMNH J.1401; E. Eudes-Deslongchamps (1867-69) also re-described 'S.' boutilieri using these specimens. In addition, Woodward (1885) explicitly reports that the cast (OUMNH J.1401) in Oxford has the names of Crocodilus oxoniensis Conybeare and Steneosaurus oxoniensis de la Beche. Woodward (1885: 501) also writes that " Mr. Edward Wilson, to whose kindness the writer is indebted for particulars of the Crocodilian fossils now in the Bristol Museum, is unable to discover any such label in the collection; and the present whereabouts of the original specimen seems to be unknown". Based on Woodward's (1885) comments, it appears as though the epithet 'oxoniensis' was never formally published, and thus is an invalid name (a nomen nudum).

The late 1800s
As mentioned previously, Phillips (1871) named and described a new species, Teleosaurus brevidens, using teleosauroid material housed in the Oxford Museum (however, he was unclear which exact specimen(s) he referred to). He mentioned these specimens as being "observed in Stonesfield and some other places in the Great oolite near Oxford, [as] Enslow Bridge and Kidlington" (Phillips 1871: 184 fig. 1 (Phillips 1871), he described the teeth of T. brevidens as "rather short [teeth]…a little curved, uniformly striated, the striae growing more prominent toward the point and finer toward the base… [a] slight trace of bicarination on these teeth, near the apex, which is usually blunt…" and regarded these features as among the defining characteristics of T. brevidens (Phillips appears to be referring to the anastomosing pattern that is characteristic of machimosaurin teeth). Presumably, Phillips (1871) was referring to OUMNH J.29850 and/or OUMNH J.1403, as he referred to not only the teeth but the skull and palatal material as well.
Hulke (1877)  Interestingly, Hulke (1877: 29) also briefly refers to Steneosaurus (Crocodilus) oxoniensis in his comparisons, and writes as if E. Eudes-Deslongchamps used the name as well (["…which M. Deslongchamps regards as the mature form of Steneosaurus Oxoniensis…"]). However, as mentioned previously, E. Eudes-Deslongchamps (1867-69) did not find oxoniensis as valid and referred to it as boutilieri. However, we agree with Hulke (1877) and consider S. stephani to be a separate and distinct species (see Discussion).

The 1900s to present day
In the early 20th century, Watson (1911) briefly described the palatal view of S. stephani (NHMUK PV OR 49126), which was initially neglected by Hulke (1877) . Watson (1911, No. 18: 3)  It was not until Phizackerley's (1951) short monograph on teleosauroid specimens from Oxford that 'S.' boutilieri, T. brevidens or the de la Bêche & Conybeare specimen was one again examined in any detail. Phizackerely (1951detail. Phizackerely ( : 1184 initially referred to OUMNH J.29850 as being the type specimen of T. brevidens (note that, in 1951, the specimen was housed in the Department of Zoology at Oxford, so the specimen number was 1639/1; when it was moved to the Earth Sciences collection in the museum and given its current specimen number is unknown). Curiously, Phizackerley (1951Phizackerley ( : 1185 then stated that type material of T. brevidens was inaccessible, but then used1639/1 (OUMNH J.29850) as the type specimen of his new species 'Steneosaurus' meretrix. This very odd situation means that the type specimen of 'S.' meretrix was already the type specimen of T. brevidens, thus making them objective synonyms. Thus, both T. brevidens and 'S.' meretrix are considered as invalid species, and are junior synonyms of 'S.' boutilieri. Steel (1973) and Vignaud (1995Vignaud ( , 1997 referred to OUMNH J.1401 (the 'oxoniensis' skull) and the T. brevidens/ 'S.' meretrix holotype (OUMNH J.29850) as ' S.' boutilieri, although Vignaud (1997) referred to OUMNH J.1401 as C. oxoniensis. Godefroit (1996) also briefly mentions E. Eudes-Deslongchamps' (1867-69) holotype of 'S.' boutilieri (as well as Phizackerley's (1951) work on the aforementioned Oxford specimens and Hulke's (1877) S. stephani) in comparison with a new specimen of S. megistorhynchus. Vignaud (1998) also highlighted the characteristic robust shape of the teeth seen in 'S.' boutilieri.

Geology
The Great Oolite Group (GOG) is a Middle Jurassic (Bathonian) lithostratigraphic unit that is well known from extensive exposures in central England, UK. The GOG consists of three formations that are, from youngest to oldest, the Cornbrash Formation, the Forest Marble Formation, and the White Limestone Formation (Palmer & Jenkyns, 1975;Sellwood et al., 1985). Both the Forest Marble Formation and the White Limestone Formation were primarily deposited by laterally migrating tidal channels in an intertidal environment (Klein, 1963(Klein, , 1965Palmer & Jenkyns, 1975). Several important fossils are known from the Great Oolite Group, including pterosaurs (O'Sullivan and Martill, 2018), crocodylomorphs (de la Bêche and Conybeare, 1821;Eudes-Deslongchamps, 1867-69), dinosaur remains (Woodward 1910;Benson 2010) and reptilian eggs (Buckman, 1860).

The Cornbrash Formation (CF) is a Bathonian-Callovian (Middle Jurassic)
lithostratigraphic unit that underlies the Oxford Clay Formation (OCF) (Wright, 1977). This stratigraphic section consists of medium-to fine-grained limestone that is bioturbated, yields some reptilian (Hulke, 1877;Benton and Spencer, 1995) and many trace (e.g. Powell and Riding, 2016) fossils and stretches from the Weymouth area to the Scarborough area in the UK (Cox & Sumbler, 2002), similar to the OCF. The CF is comprised of two main units (the Cornbrash Limestone and the Cornbrash Shales) which form a transgressional marine cycle and marks the first marine invasion of the Jurassic delta subsequent to that represented by the Middle Bajocian Scarborough Beds (Wright, 1977). The CF in England is also correlative with 'Fuller's Earth inférieur' in France (see Vignaud, 1995). All currently known English specimens of ' S.' boutilieri (e.g. OUMNH J.1401, OUMNH J.29850) and 'S.' larteti (OUMNH J.29851) were found in the CF.

Etymology: named in honour of father and son French naturalists Jacques Amand and Eugène
Eudes-Deslongchamps, who thoroughly described the holotype specimen (in addition to numerous other teleosauroid taxa during the latter 1800s).

Our statement of the characters that we regard as differentiating Deslongchampsina larteti
from other taxa is given by the species diagnosis below. 3. The neotype can be recognised through both the description below and Fig. 1. 4. The holotype is presumed destroyed in 1944 during the bombing of Caen. 5. The holotype had a partial skull; the description and figure given by J. A. Eudes-Deslongchamps (1866) show it was a gracile, mesorostrine skull with rounded (little constricted) premaxillae, oval orbits, large antorbital fenestrae, robust and pointed teeth and slight but noticeable ornamentation. As such, the neotype is consistent with what is known of the former name-bearing type. 6. Unfortunately, the neotype is not from the same locality or country as the holotype.
However, both types are from the same age (Bathonian) and relative formation (the Cornbrash Formation in England is correlative with "Fuller's Earth inférieur" in France), and have both been referred to as the same species (see Historical Background). 7. The neotype is the property of a recognized scientific institution, OUMNH, which maintains a research collection with proper facilities for preserving name-bearing types, and is accessible for study. mediolaterally thin posterior processes of the nasals*; gradual and well-developed anteroventral sloping of the nasals*; presence of large, elongated antorbital fenestrae, and internal antorbital fenestra between 25-50% of the length of the orbit (shared with Steneosaurus gracilirostris Westphal, 1961, and 'S.' brevior); frontal width subequal with orbital width (shared with the Chinese teleosauroid previously referred to as Peipehsuchus (see Li, 1994), Mycterosuchus nasutus Andrews, 1913 [Foffa et al., in press]); palatine anterior margin terminates level to 21st maxillary alveoli, or more distal alveoli (shared with Steneosaurus leedsi Andrews, 1909, Myc. nasutus and B. megarhinus); four premaxillary alveoli; large, robust, non-compressed teeth with a pointed apex and high relief enamel ridges (similar to Steneosaurus edwardsi).

Description
A partial skull, broken into two pieces (rostrum and occipital), is the neotype of and differing from the small, subcircular antorbital fenestrae seen in other teleosauroids (e.g.  obtusidens LPP.M.21). It is slightly anteroposteriorly thickened and has small sparse pits for ornamentation in the medial part. The postorbital-squamosal contact is not preserved. While the specimen is broken into two pieces (see above) near the middle of the supratemporal fenestrae, they appear to be anteroposteriorly elongated and rectangular in shape ( Fig. 1A-B).

Myc. nasutus
Parietal: The single parietal is a relatively large, mediolaterally thickened bone ( Fig. 1A- (Andrews, 1909(Andrews, , 1913Romer, 1956;Nesbitt, 2011). It is difficult to determine where exactly the mandibular symphysis begins, as the dorsal surface of the dentary is poorly preserved, but it starts approximately at the 16 th or 17 th alveolus. There are at least 23 dentary alveoli preserved on the left side. The interalveolar spacing is variable throughout the dentary and the alveoli are subcircular in shape ( Fig. 2A).
In lateral view ( Fig

Our statement of the characters that we regard as differentiating Yvridiosuchus boutilieri
from other taxa is given by the species diagnosis below. 3. The neotype can be recognised through both the description below and Fig. 3. 4. The holotype is presumed destroyed in 1944 during the bombing of Caen. 5. The holotype had a partial skull; the description and figure given by E. Eudes-Deslongchamps (1868) showed it to be large, rugose and robust with broad, rounded palatines and robust teeth. As such, the neotype is consistent with what is known of the former name-bearing type. 6. Unfortunately, the neotype is not from the same locality or country as the holotype.
However, E. Eudes-Deslongchamps compared the holotype with our proposed neotype, and considered that they were from the same species and included both in his future description of the species. In addition, both the holotype and our proposed neotype come from the equivalent to the Great Oolite Group in England. 7. The neotype is the property of a recognized scientific institution, OUMNH, which maintains a research collection with proper facilities for preserving name-bearing types, and is accessible for study.

Description
The neotype of Y. boutilieri (OUMNH J.1401) (Fig. 3) is a partial rostrum, preserved from the anterior-most part of the snout until the anterior margin of the supratemporal fenestrae.
OUMNH J.1401 is broken into two sections (Fig. 3): the first piece is complete from the anterior margin to the right M15 and left M18 alveoli (a posteriorly-directed horizontal break), and the second piece includes the right M15 alveolus back to the supratemporal fenestra anterior margin. One area of the rostrum (including the right fourth premaxillary alveolus and the first four maxillary alveoli on both sides) is reconstructed with plaster (see  (Fig. 5G-H), a partial mandible, has been referred to as 'S'. boutilieri so we describe it here as well (however, it is unknown who referred this specimen to the species boutilieri, or when they did so). In OUMNH J.1403, the basisphenoid is exposed along the palatal surface anterior to the quadrates and bifurcates the posterior part of the single pterygoid ( Fig. 5C-D).
Pterygoid: In ventral view, the single fused pterygoid is well preserved in OUMNH J.1403 ( Fig. 5C-D), as well as on the right side in OUMNH J.29850 (Fig. 4C-D). The pterygoid is anteroposteriorly elongated and mediolaterally expanded; it is also slightly dorsoventrally thick in comparison with other teleosauroids (e.g. S. leedsi NHMUK PV R 3806; the Chinese teleosauroid IVPP V 10098). The anterior pterygoid process articulates with the palatines.
The pterygoid is concave and posterodorsally curved, most notably in the posterior area ( Other elements: In all specimens, the proötics are not visible and the laterosphenoids are poorly preserved; we were therefore unable to describe these bones properly.

Mandible:
The nearly complete mandible of OUMNH J.29580 (Fig. 4A-H) is cemented to the cranium so the dorsal surface is not visible, as mentioned above. It measures approximately 74 cm in length; however, the posterior-most part of the mandible is not preserved. In OUMNH J.1404, the dorsal surface of the partially complete mandible is exposed ( Fig. 5G-H), with the posterior portion not preserved.  Andrews, 1909Andrews, , 1913Romer, 1956;Nesbitt, 2011  Articular: In OUMNH J.29850, the left articular is not preserved; however, the right articular is a separated piece from the mandible (Fig. 4K). In lateral view, the posterior mandibular rami is sharply curved dorsoposteriorly, similar to S. heberti (MNHN.F 1890-13) and Machimosaurini. The retroarticular process is anteroposteriorly elongate, mediolaterally thin and triangular-shaped in dorsal view (Fig. 4K). The anteroposterior keel is small and thin but visible, and the posterior end of the retroarticular process is slightly rounded. The middle area of the retroarticular process is substantially narrower than the glenoid fossa (39%), as opposed to other teleosauroids (e.g. 55% in S. edwardsi PETMG R178; 53% in S. leedsi NHMUK PV R 3320; 65% in S. heberti MNHN.F 1890-13).
Dentition: Throughout the dentition, the teeth (Fig. 4L) of Yvridiosuchus boutilieri are large and robust with a blunt apex, and are more similar to Lemmysuchus than Machimosaurus, being slightly less conical and weakly curved in the anterior dentary. In OUMNH J.29850, the largest tooth (M14) measure approximately 1.5 cm in apicobasal length. The enamel ridges are small yet well-developed, parallel to one another and reach the top of the apex. The enamel is thinner towards the base of the crown and becomes progressively thicker towards the apex (Fig. 4L), similar to that seen in Lemmysuchus and Machimosaurus (Young & Steel, 2014;. There are numerous protruding apicobasal enamel ridges on the teeth, giving them a 'wrinkled' texture; these apicobasal ridges are close to one another and run parallel from the base of the crown to approximately three quarters of the entire tooth. At the apex, the ridges are considerably shorter and are organized in the typical anastomosed pattern that has been described for other members of Machimosaurini (L. obtusidens NHMUK PV R 3168; Mac. buffetauti SMNS 91415;Mac. hugii MG-8730-1;Mac. rex ONM 1-25;Young et al., , 2015aJouve et al., 2016;. The teeth have true denticles and false denticles (Young et al., 2015a), although the latter cannot be seen clearly with the naked eye.

Phylogenetic analysis
Methods. We conducted a phylogenetic analysis to test the evolutionary relationships of Deslongchampsina larteti gen nov. and Yvridiosuchus boutilieri gen. nov. within Thalattosuchia, using a modified version of the dataset provided by Foffa et al. (in press), which is based off of Ősi et al (2018). This dataset is continuously being updated, as it forms the foundation of the ongoing Crocodylomorph SuperMatrix Project. The dataset was first presented in Ristevski et al. (2018); however, it has been extensively updated subsequently (see Ősi et al. (2018) and Foffa et al. (in press) for full details). All data are summarised in Supplementary data files.  Chatterjee, 1985 was used as the outgroup taxon. The differences between our analyses and those presented by Foffa et al. (in press) . 17-18, 176, 355, 362, 367, 374, 464 The cladistic maximum parsimony analysis of the dataset was conducted using TNT 1.5 Willi Hennig Society Edition (Goloboff et al., 2008;Goloboff and Catalano, 2016), following the methodology used in Young et al. (2016 (2) ran the analysis once more using implied weighing (k = 12).  (Fig. 6A). The overall strict consensus topology recovered from this analysis is extremely similar to that presented by Ristevski et al. (2018), Ősi et al. (2018) and Foffa et al. (in press).

Results
In the strict consensus tree (Fig. 6A) The majority rule consensus tree shows increased resolution (Fig. 6B) Lastly, when the analysis was run once again using implied weighting ( Fig. 6C): (1)  Overall, the crocodylomorph interrelationships found in our analysis are similar to those recovered in previous iterations of this constantly growing dataset (Ristevski et al., 2018;Ősi et al., 2018;Foffa et al., in press;Sachs et al., in review). All phylogenetic analyses are similar concerning the following aspects:

1.
The monophyly and positioning of Thalattosuchia within Crocodyliformes.

2.
The separation of Thalattosuchia into two distinct clades: Teleosauroidea and Metriorhynchoidea.

4.
Within Teleosauroidea, Steneosaurus gracilirostris is the likely basal-most species, with two subclades being recovered: (1)  Furthermore, neither the rostrum nor associated teeth are preserved in NHMUK PV OR 49126, so we cannot assess whether it has critical machimosaurin characters (e.g. large and numerous neurovascular foramina or blunt tooth crowns with anastomosed apical enamel ornamentation). Therefore, we currently agree with Hulke's (1877) diagnosis and provisionally retain NHMUK PV OR 49126 as a distinct taxon, S. stephani.

Bathonian teleosauroids of northern Africa
As mentioned previously, Jouve et al. (2016)  An interesting observation to note is that while Sub-Boreal teleosauroid ecosystems of England change in diversity through time, the ecological structuring (which has briefly been explored; see Hua, 1997;Hua & Buffetaut, 1997;Foffa et al., 2018a), in particular during the  (Foffa et al., in press) and Steneosaurus jugleri Meyer, 1845. In addition, the heavily armoured, more terrestrial and longirostrine generalist ectomorphs vanished.
The Bathonian-aged Yvridiosuchus boutilieri also represents the current oldest known member of Machimosaurini from England, and the current oldest substantial machimosaurin material from anywhere in the world. This shows that, while rare, machimosaurins were already an important component of Bathonian ecosystems and had already evolved key characteristic machimosaurin features earlier than originally thought (i.e. blunt anastomosing teeth, parallelogram-shaped supratemporal fenestrae, deep reception pits).

Conclusions
While 'Steneosaurus' larteti and 'Steneosaurus' boutilieri are both morphologically and historically important teleosauroid taxa, little work has been done on them since the mid-20 th century. Here we re-describe one complete specimen of 'S.' larteti, OUMNH J.29851, consisting of a nearly complete skull, and use it to establish a new genus, Deslongchampsina.
We then re-describe four specimens of 'S.' boutilieri (the designated neotype consisting of a partial skull; one complete skull and nearly complete mandible; one nearly complete skull;                (Lamouroux, 1820), NHMUK PV R 119a.
Refer to the main text for the abbreviations list. Scale bars: 5 cm.