Craspedochelys? sp., Rutimeyer, 1873

Madzia, Daniel, Szczygielski, Tomasz & Wolniewicz, Andrzej S., 2021, The giant pliosaurid that wasn’t-revising the marine reptiles from the Kimmeridgian, Upper Jurassic, of Krzyżanowice, Poland, Acta Palaeontologica Polonica 66 (1), pp. 99-129 : 105-113

Craspedochelys? sp.

Figs. 2–6 .

Material. — MZ VIII Vr-71, shell fragments from the canary-yellow Nerinea limestone, upper Kimmeridgian of Krzyżanowice, Poland ( Dąbrowska 1957; Borsuk-Białynicka and Młynarski 1968).

The specimen series collectively labelled MZ VIII Vr-71 was collected in 1962 from three pits located 15–20 m apart and consists of hundreds of shell fragments of varying sizes belonging to three or four individuals of different ontogenetic stages and supposedly representing a single species ( Borsuk-Białynicka and Młynarski 1968). The elements from each pit present little to no overlap, therefore their attribution to a single taxon is uncertain. They nonetheless originate from a single stratigraphic layer from the same locality and exhibit congruent characteristics of the shell surface, morphology of the sulci, shell thickness and degree of ossification. Due to the absence of significant differences, these elements are herein considered as representing a single taxonomic entity. Borsuk-Białynicka and Młynarski (1968) stated that the series included a central fragment of carapace, several carapace and plastron fragments likely belonging to another individual, two plastron fragments from a much larger individual, and numerous smaller pieces. The original description of this material ( Borsuk-Białynicka and Młynarski 1968) was very brief and no actual specimens were figured. Two of the numerous fragments were recently redescribed and figured ( Tyborowski and Błażejowski 2019a: fig. 4B, C, 2019b: fig. 4); however, the descriptions contain numerous errors, insufficient and selective comparisons and do not discuss nor reflect the relevant literature. Therefore, a thorough redescription was needed.

Description.—A central fragment of carapace ( Fig. 2 ): This is the only piece figured as an interpretive drawing in Borsuk-Białynicka and Młynarski (1968: fig. 2), although in a very stylized form, and as photographs in Tyborowski and Błażejowski (2019a: fig. 4C, D; 2019b: fig. 4C). The fragment is pieced together from numerous small fragments and a substantial part is reconstructed with plaster. Currently it is split into two parts, separated along the neurocostal connection.

Borsuk-Białynicka and Młynarski (1968) described this fragment as comprising the third and fourth neural and the costals: third to fifth on the right and third on the left. The layout of the sulci, however, indicates that the costals preserved on the right side of the fragment are either the second to fourth or fourth to sixth (the suture between the middle and posteriormost costal is fused and visible only viscerally), and thus the neurals represent either the second and the third or the fourth and the fifth ( Fig. 6A : 1). As mentioned by Tyborowski and Błażejowski (2019a), small fragments of another, preceding costal are preserved at the

anterior right edge of the piece. The left preserved costal is attached incorrectly and more likely corresponds to the second preserved costal on the right side (the fourth according to Borsuk-Białynicka and Młynarski 1968, but actually likely either the third or the fifth). This is clearly evidenced by a sulcus on its external surface. Despite the damage, the sulcus interpreted by Borsuk-Białynicka and Młynarski (1968) as an interpleural sulcus is clearly sinuous and approaching the costoneural suture. These aspects make it markedly different from the unambiguous interpleural sulcus preserved on the contralateral side of the fragment, which is straighter and well-separated from the neurals, but congruent with the morphology of intervertebral sulci in the same piece. A supposed intervertebral sulcus was also marked in red close to the anterior end of the first preserved neural. Although plausible, this identification is ambiguous—such a position of the sulcus would require a departure from the typical turtle layout in which the intervertebral sulci cross only the odd-numbered neurals, and it cannot be ruled out that the observed groove might have originated as a result of damage. Despite the intervertebral sulci being sinuous, the areas of the vertebral scutes lack regular and well-defined radial striations, which are typical for many Jurassic turtles ( Meyer 1860; Wagner 1861; Maack 1869; Rütimeyer 1873; Bräm 1965; Joyce 2003; Jansen and Klein 2014; Sullivan and Joyce 2017), but particularly for Eurysternum wagleri Meyer, 1839a and Pelobatochelys blakii Seeley, 1875 (see Meyer 1839a; Seeley 1875; Anquetin and Joyce 2014; Pérez-García 2015c; Anquetin et al. 2017). Only several, short and irregularly placed groves are present along the anterior edge of the vertebral scute area ( Fig. 2A 1 ). The intervertebral sulcus is predominantly horizontal and the vertebral lacked a sagittal tip, unlike the vertebral scutes in, e.g., Achelonia formosa Meyer, 1860 , Eurysternum wagleri , Idiochelys fitzingeri Meyer, 1839b , or Palaeomedusa testa Meyer, 1860 (e.g., Meyer 1860; Maack 1869; Rütimeyer 1873; Joyce 2003; see Anquetin and Joyce 2014; Joyce and Mäuser 2020). The scute areas are completely flat, unlike in Platychelys oberndorferi Wagner, 1853 (e.g., Meyer 1860; Wagner 1861; Bräm 1965; Lapparent de Broin 2001; Sullivan and Joyce 2017). The vertebrals are moderately wide, about five times the width of the neurals. This proportion resembles Thalassemys hugii Rütimeyer, 1873 , Plesiochelys etalloni ( Pictet and Humbert, 1857) (with the exception of NMS 79 with exceptionally wide vertebrals) and Plesiochelys bigleri Püntener, Anquetin, and Billon-Bruyat, 2017a , as well as Craspedochelys spp. , but is smaller than that in Achelonia formosa , Chelonides wittei Maack, 1869 , Eurysternum wagleri , Idiochelys fitzingeri , Jurassichelon moseri ( Bräm, 1965) , Jurassichelon oleronensis Pérez-García, 2015b , Thalassemys bruntrutana Püntener, Anquetin, and Billon-Bruyat, 2015 , Thalassemys marina Fraas, 1903 , Solnhofia parsonsi Gaffney, 1975b , Solnhofia brachyrhyncha Anquetin and Püntener, 2020 , and likely Pelobatochelys blakii , and larger than in Tropidemys langii Rütimeyer, 1873 and Tropidemys seebachi Portis, 1878 (see Meyer 1839b, 1860; Maack 1869; Rütimeyer 1873; Seeley 1875; Portis 1878; Fraas 1903; Andrews 1921; Rieppel 1980; Joyce 2000; Anquetin et al. 2014b, 2017; Karl et al. 2012a; Püntener et al. 2014, 2015, 2017b; Anquetin and Joyce 2014; Pérez-García 2015a, c; Anquetin and Chapman 2016; Raselli and Anquetin 2019; Joyce and Mäuser 2020). MZ VIII Vr-71 differs from Tropidemys seebachi in the lack of infravertebral scutes ( Joyce and Mäuser 2020).

All the costals are incomplete distally and the rib necks are broken off at the level of the entrance into the costal plates. The neurals are coffin-shaped in outline, longer than wide, and wider anteriorly than posteriorly, unlike in platychelyids, and differ from the elongated but more rectangular and nearly parallel-sided neurals of Neusticemys neuquina ( Fernández and de la Fuente, 1988) (see Meyer 1860; Wagner 1861; Bräm 1965; Fernández and de la Fuente 1993; Gasparini et al. 1997; Lapparent de Broin 2001). Their anterolateral edges are much shorter than the posterolateral edges, measuring (as preserved) 12 mm and 34 mm for the first and 13 mm and 39 mm for the second preserved neural (contra 10 mm and 15 mm given by Tyborowski and Błażejowski 2019a), respectively. This, together with the anteroposterior elongation of the neurals and lack of the sagittal keel, distinguishes MZ VIII Vr-71 from Tropidemys spp. ( Rütimeyer 1873; Portis 1878; Karl et al. 2012a; Anquetin et al. 2014b, 2017; Püntener et al. 2014; Pérez-García 2015c; Anquetin and Chapman 2016; Joyce and Mäuser 2020). The lack of any keeling also distinguishes it from Pelobatochelys blakii (see Seeley 1875; Pérez-García 2015c; Anquetin et al. 2017). The first preserved neural is missing its anterior right part, has its corners less pronounced than the second preserved neural, and shows a convex rather than concave anterior edge, resulting in an overall rounder outline. However, this morphology seems to be taphonomic. The interneural suture is posteriorly convex. The vertebrae are not preserved, the neural processes are broken just below the neural plates. Unlike Idiochelys fitzingeri, MZ VIII Vr-71 does not show any signs of neural reduction ( Meyer 1839b, 1860; Rütimeyer 1873; Lortet 1892; Anquetin et al. 2017). Unlike in Thalassemys spp. , there are no striations perpendicular to the sutures ( Rütimeyer 1873; Anquetin et al. 2014b, 2017; Pérez-García 2015a; Püntener et al. 2015). The illustration presented for that specimen in the original paper ( Borsuk-Białynicka and Młynarski 1968: fig. 2) does not accurately reflect its shape and completeness, showing some elements that are not currently present and, judging from the original description, have likely never been. In the original paper, the illustration is described as “reproducing, according to the principles of symmetry, the lacking parts of plates and damaged edges, as well as […] aligning to each other the disposition of particular parts” ( Borsuk-Białynicka and Młynarski 1968: 219). Still, the specimen is recognizable and, therefore, it is puzzling why Tyborowski and Błażejowski (2019a: 745) describe it as a “new specimen”. Curiously, the dimensions of the specimen given in the text and shown in the figures in Tyborowski and Błażejowski (2019a, b) are clearly reduced. For example, the costal preserved on the left side of the fragment measures about 72 × 43 mm, not 51 × 29 mm, and the first preserved costal on the right side of the fragment (the most complete, but still broken distally) measures about 144 × 39 mm (length to width ratio equals 3.7), not 102 × 35 mm.

Two fragments of plastron and carapace with peripherals, likely belonging to a single individual ( Figs. 3 , 4A–C ): One of the pieces ( Fig. 3A ) was interpreted by Borsuk-Białynicka and Młynarski (1968) as consisting of the first neural, fragments of the first two costals and part of the hyoplastron with a pectoroabdominal sulcus. The identification of the carapace elements seems correct ( Fig. 6A : 2), as suggested by their shapes, interconnections, and by what appears to be a very gentle transverse sulcus crossing the supposed first neural and first costal—an appropriate location for the sulcus between the first two vertebral scutes. Posteriorly to the first neural, part of the second neural seems to be present as well. Neither of the elements are preserved completely. The first neural has its anterior part broken but seems to have a rounded rectangular shape in outline. The second neural has only its left part preserved but it appears to be coffin-shaped, with the anterior part wider than the posterior part and very short craniolateral edges. The interneural suture is posteriorly convex. All of these features are common for the first two neurals of thalassochelydians (e.g., Anquetin et al. 2014b). Both neurals lack a keel. The preserved proximal parts of the first and the second left costal are broken into irregular fragments and provide little morphological data. The fragment of the hyoplastron is trapezoidal and bears a single, nearly straight transverse sulcus. If the spatial correlation with the associated carapace fragments is any indication, this would indicate that the sulcus represents the humeropectoral, rather than the pectoroabdominal sulcus. There is, however, no clear morphological indication that would favour one interpretation over the other, because the lateral edge of the element is damaged.

The second fragment ( Fig. 3B ), likely belonging to the same individual, has been described as strongly damaged fragments probably pertaining to the seventh costal and neural together with a fragment of hypoplastron with the abdominofemoral sulcus ( Borsuk-Białynicka and Młynarski 1968). This interpretation seems to be incorrect—there appears to be no neural (although it cannot be ruled out that it was initially present and became separated over the years) and the external surface apparently bears two oblique sulci, one of which is faint and predominantly aligned anteroposteriorly, and the other sinuous and nearly transverse. This layout refutes the identification of the element as the seventh costal, but supports its recognition as the eight costal, which would agree with the relatively small size of the piece ( Fig. 6A : 3). Just like in the previous piece, the costal is incomplete and broken into several fragments. The hypoplastron is subtriangular, has damaged edges, and bears an oblique sulcus, which may be the femoroanal, rather than the abdominofemoral sulcus, as suggested by its straight, rather than curved trajectory and proximity of the (supposedly) eighth costal. Nonetheless, the interpretation of this fragment is ambiguous.

According to Borsuk-Białynicka and Młynarski (1968), the gathered material also includes a fragment of the carapacial rim bearing the sulci between the anteriormost four marginals ( Fig. 4A ) and two peripherals interpreted as the third right ( Fig. 4B ) and the fifth or sixth peripheral from an indeterminate side ( Fig. 4C ), all likely belonging to the same individual as the associated carapace and plastron fragments. The identification of the former fragment Fig. 4A ) seems to be accurate ( Fig. 6A : 4); this is supported by the presence of an acute anterior corner of the supposed first pleural and a gently downturned free edge. The element comes from the left side of the body. The areas of the first and the fourth marginal are incomplete anteriorly and posteriorly, respectively. In dorsal view, the first marginal was apparently narrower than the rest and had a straight, rather than protruding anterolateral edge. This differs from the anteriorly bulging first marginal area of Craspedochelys passmorei ( Andrews, 1921) , Jurassichelon moseri , Jurassichelon oleronensis , and Plesiochelys etalloni ( Fig. 6B, C ), but resembles most specimens of Craspedochelys jaccardi ( Pictet, 1860) ( Fig. 6D ) (with the possible exception of the holotype, which is slightly divergent in morphology from the specimens from Solothurn, see Pictet 1860; Anquetin et al. 2014b), Craspedochelys picteti ( Rütimeyer, 1873) ( Fig. 6E ), Plesiochelys bigleri (with the exception of MJSN BSY007- 257, see Püntener et al. 2017b), and Thalassemys hugii (see Rütimeyer 1873; Portis 1878; Andrews 1921; Bräm 1965; Rieppel 1980; Anquetin et al. 2014a, b, 2017; Pérez-García 2015a; Püntener et al. 2017b; Raselli and Anquetin 2019). The marginals are narrower than in pleurosternids such as Dorsetochelys typocardium (Seeley, 1869) , Pleurosternon bullocki ( Owen, 1842) , Riodevemys inumbragigas Pérez-García, Royo-Torres, and Cobos, 2015 , or Selenemys lusitanica Pérez-García and Ortega, 2011 (see Owen 1853; Milner 2004; Pérez-García and Ortega 2011; Pérez-García 2014; Pérez-García et al. 2015). Both the first and the second marginal formed an extensive contact with the first vertebral scute, unlike in Achelonia formosa , Palaeomedusa testa , Parachelys eichstaettensis Meyer, 1864 , Plesiochelys spp. , and (with the exception of Tropidemys seebachi NKMB Watt 09/162) most specimens of Tropidemys spp. ( Fig. 6B, C ; see also Meyer 1860; Rütimeyer 1873; Bräm 1965; Lapparent de Broin et al. 1996; Joyce 2003; Karl et al. 2012a; Anquetin et al. 2014a, b; Püntener et al. 2014, 2017a, b; Anquetin et al. 2017; Raselli and Anquetin 2019; Joyce and Mäuser 2020). In Craspedochelys spp. the contact between the second marginal and the first vertebral was apparently variable, but it was present at least in Craspedochelys picteti NMS 9149 (former NMS 608; Fig. 6D, E ; see also Bräm 1965; Anquetin et al. 2014b). Borsuk-Białynicka and Młynarski (1968) mentioned that the areas of the first three pleurals were preserved in MZ VIII Vr-71, but this was incorrect; in addition to most of the lateral margin of the first pleural, only the craniolateral corner of the first vertebral is preserved. If at the time of the original description there was an additional part preserved posterior to the existing series that bore the sulcus between the first and the second pleural, it has been detached since. The sulci are indistinguishable ventrally in a manner similar to, e.g., Craspedochelys jaccardi (see Bräm 1965; Anquetin et al. 2014b, 2017), either as a result of very limited ventral exposition of the anterior marginals (body wall being unusually close to the edge of the carapace) or as a taphonomic effect. In marginal aspect, the fragment is gently sinuous, reaching the dorsal peak at the level of the second marginal scute, and the third and fourth marginals increase in dorsoventral diameter due do the downturn of their free edges becoming more pronounced. As mentioned by Borsuk-Białynicka and Młynarski (1968), the sutures are not visible in that specimen and there are no costoperipheral fontanelles. The latter may differentiate MZ VIII Vr-71 from, e.g., Achelonia formosa , Chelonides wittei , Eurysternum wagleri , Hydropelta meyeri ( Thiollière, 1850) , Idiochelys fitzingeri , Jurassichelon spp. , Neusticemys neuquina , Pelobatochelys blakii , probably Solnhofia spp. , Thalassemys spp. , and Tropidemys seebachi ( Meyer 1839b, 1860; Maack 1869; Seeley 1869, 1875; Rütimeyer 1873; Portis 1878; Lortet 1892; Fraas 1903; Rieppel 1980; Fernández and de la Fuente 1993; Gasparini et al. 1997; Joyce 2000; Anquetin and Joyce 2014; Anquetin et al. 2014b, 2017; Pérez-García 2015a; Püntener et al. 2015; Anquetin and Püntener 2020; Joyce and Mäuser 2020), although the preserved section comes from the anterior part of the carapace, so the presence of more posterior fontanelles cannot be excluded. The isolated bone described as the third right peripheral ( Figs. 4B , 6A : 5) indeed resembles its counterpart from the left side both in the layout of the sulci and in the gradual increase of the downturn of its edge. The identification of the piece described as the fifth or sixth peripheral ( Fig. 4C ) is, however, dubious: the downturn of the edge and the dorsoventral diameter do not match the fourth peripheral of the preserved series, and the fragment does not show any intermarginal sulcus nor trace of the bridge. Both isolated peripherals are broken off rather than separated along the suture lines. Neither of the fragments shows whether the bridge was ligamentous or osseous.

There is at least one other fragment with well-preserved sutures and a broken ridge on the visceral surface ( Figs. 4D , 6A : 6), likely belonging to a different individual and representing a part of the intermediate element present in thalassochelydians (which is thought to at least partially articulate with the vertebral column; Anquetin et al. 2014b), the posterior left part of the preceding neural, and the proximal portion of the last left costal. The incompleteness of that fragment, however, makes this identification tentative. The preserved part of the neural suggests it was coffin-shaped, as the more anterior neurals. Its dorsal surface bears several oblique grooves, but it is not certain whether they were present in life or are of taphonomic origin (and, in the former case, whether they are pathological or not). The intermediate element receives the posterior edge of the costal via a posteriorly convex suture and expands posterolaterally. Neither of the elements is crossed in the preserved part by a sulcus, but the last intervertebral sulcus was located either more anteriorly, in the anterior part of the last neural, or more posteriorly, in the posterior part or behind the intermediate element (as in, e.g., Craspedochelys jaccardi NMS 102a; see Anquetin et al. 2014b: fig. 4M, N). All of the fragments are in agreement with the central part of the carapace ( Fig. 2 ) in suture and sulci morphology, whenever the sutures and/or sulci are preserved.

Six large fragments of costals and indeterminate parts of the carapace ( Fig. 5A–F ): The costals were interpreted by Borsuk-Białynicka and Młynarski (1968) as the first, the third or fourth, and the seventh or eighth, but no justification for these identifications was provided. For certain, MZ VIII Vr-71 includes two fragments of a second, fourth, or sixth costal (based on the presence of the pleurovertebral and interpleural sulci in one [ Fig. 5A ] and no sulci in the proximal part of the second [ Fig. 5F ]), a fragment of a probable third or fifth costal (based on the presence of a sinuous, likely intervertebral sulcus [ Fig. 5B ]), and at least seven large (over 30 mm) parts, tentatively identified as costals but lacking sulci on the external surfaces ( Fig. 5D–E ). The latter may represent areas of either vertebral or pleural scutes, but their state of preservation makes it impossible to establish whether the preserved parts are proximal or distal, and therefore their sequential identity is unknown. There is no isolated element readily identifiable as the first costal (contra Borsuk-Białynicka and Młynarski 1968; Tyborowski and Błażejowski 2019a, b). Since Borsuk-Białynicka and Młynarski (1968) did not figure nor describe in detail the fragment mentioned as an isolated first costal, it is impossible to verify whether they singled out the same piece that was described and figured by Tyborowski and Błażejowski (2019a: fig. 4B, 2019b: fig. 4A, B). In any case, although the general outline of the latter element superficially resembles the shape of the first costal in “plesiochelyids” (e.g., Anquetin et al. 2014b), such an identification is incorrect and refuted by the lack of any sulci on its external surface and rib-related structures viscerally, as well as by the presence of a single, natural, non-sutural edge ( Fig. 5G ; the fragment was broken into several pieces during DM, TS, and ASWs personal examination for this study, but the figures show its appearance before it became damaged). This edge is unlike the borders of the intercostal fontanelles, as it is nearly straight, spans along the entire preserved length of the element, and is not associated with a longitudinally striated free distal part of a rib (compare with, e.g., Meyer 1839b, 1860; Maack 1869; Rütimeyer 1873; Seeley 1875; Portis 1878; Lortet 1892; Fernández and de la Fuente 1988; Joyce 2000; Karl et al. 2012b; Anquetin and Joyce 2014; Anquetin and Püntener 2020; Joyce and Mäuser 2020; Püntener et al. 2020). Furthermore, the edge itself is recessed, i.e., it is deep to the external surface and even slightly deeper than the visceral surface, forming a low lip viscerally. The area between the edge of the element and its main body is set at a conspicuous, obtuse angle to the external surface and the outline of the latter is gently concave along their border between them, yet both surfaces show similar microsculpture suggesting that they were overlain by epidermal scutes. Given all of these characteristics, the interpretation of the element as a hyo- or hypoplastron fragment close to the inguinal or axillary notch is more likely. Just like for the central part of the carapace, the element is significantly larger than indicated by Tyborowski and Błażejowski (2019a, b), measuring about 75 × 40 mm. In all the isolated costals, the morphology of the sulci and sutures, whenever preserved, is congruent with the remaining fragments.

Two thick plastron fragments ( Fig. 5H ): The pieces are undiagnostic beyond identification as plastral elements. We agree with Borsuk-Białynicka and Młynarski (1968) in their conclusion that these belong to a large individual, as indicated by their thickness, and that their position within the shell is indeterminate.

Small, unidentifiable shell fragments: Hundreds of small, unidentifiable shell fragments are present as part of the MZ VIII Vr-71 series. Some of these bear sulci, but due to their small size and damage it is impossible to reliably identify their position within the shell.

Discussion.— Borsuk-Białynicka and Młynarski (1968) attributed all the gathered material (not part of it, contra Tyborowski and Błażejowski 2019b) to Tretosternon aff. punctatum , a taxon with a complicated taxonomic history (see Lapparent de Broin et al. 1996; Lapparent de Broin and Murelaga 1999; Milner 2004; Joyce 2017), which is now considered a nomen dubium ( Milner 2004; Joyce 2017). Borsuk-Białynicka and Młynarski (1968: 219) mistakenly cited the geological age (apparently unaware that the British material was dated to the Early Cretaceous rathen than the Kimmeridgian, see Joyce 2017 and references therein), flatness of the carapace and “its typical sculpture” as “very characteristic of the genus”.

The sculpture, which for over a century was considered diagnostic for Tretosternon spp. , was problematic in itself. Based on one of the two syntype specimens, which have never been figured and are now lost ( Lydekker 1889; Delair 1958; Lapparent de Broin et al. 1996; Lapparent de Broin and Murelaga 1999; Milner 2004; Joyce 2017), Owen (1842: 165) stated that in his newly described Tretosternon punctatum “the external surface of […] the carapace is closely pitted with minute irregular impressions, smaller than a pin’s head”. However, he also referred the specimens figured earlier by Mantell (1827: pl. 6: 1, 3, 5) and (at least some of them) already described as “ Trionyx bakewelli ” Mantell, 1833 (currently “ Helochelydra ” bakewelli , see Joyce 2017) at that time, which clearly show a finely tuberculated surface, to T. punctatum . It is unknown whether this inconsistency was the result of a misinterpretation of Mantell’s (1827, 1833) figures by Owen (1842). This led to a confusion causing helochelydrid-like tubercles rather than pits to be cited as characteristic for the taxon (e.g., Lydekker 1889). The description of the pitted, lost syntype was more recently interpreted as being consistent with the morphology present in Pleurosternon bullocki , suggesting synonymy ( Lapparent de Broin et al. 1996; Lapparent de Broin and Murelaga 1999; Milner 2004), but Owen (1842) never noted any particular similarity in the sculpture of the two, despite describing them side by side in the same paper. The surficial characteristics of the specimens figured by Mantell (1827, 1833) are typical for helochelydrid turtles, to which many of the specimens historically attributed to the genus “ Tretosternon ” are now referred ( Lapparent de Broin et al. 1996; Lapparent de Broin and Murelaga 1999; Milner 2004; Joyce 2017). The shells/carapace of these turtles are commonly ornamented in little tubercles and granulations ( Lapparent de Broin and Murelaga 1999; Milner 2004; Joyce et al. 2011; Scheyer et al. 2014; Joyce 2017; Herzog 2019), which in some species are prone to breakage, leaving minute pitting ( Joyce et al. 2011; Scheyer et al. 2014; Joyce 2017; Herzog 2019). Arguably, these are more similar to the trionychid sculpturing cited by Owen (1842) than is the shell surface of Pleurosternon bullocki ( NHMUK 21351, TS personal observation). It may be thus speculated that the “impressions” observed by Owen (1842) were in fact such damaged tubercles, but this cannot be proven until the missing specimen is rediscovered in the future. The tubercles of “ Helochelydra ” bakewelli are, admittedly, not easily broken off ( Joyce et al. 2011; Joyce 2017), but their sculpture is otherwise extremely similar to that in some other helochelydrids in which the tubercles are prone to breakage ( Joyce et al. 2011), so such a mistake seems plausible. The Pleurosternon bullocki hypothesis was favoured over the helochelydrid one by Lapparent de Broin et al. (1996) and Lapparent de Broin and Murelaga (1999) due to the presence of striations perpendicular to the sutures in the specimen described by Owen (1842) which, as the authors argued, are known only in pleurosternids ( Owen 1853; Lapparent de Broin et al. 1996; Lapparent de Broin and Murelaga 1999; Milner 2004). One of the helochelydrid specimens presented by Joyce et al. (2011: fig. 2B) and several described and pictured by Herzog (2019), however, have tubercles coalescing near the sutures, producing perpendicular ridges/striations, and prone to breakage tubercles on the main body of the bones. These linear structures do span “for the extent of two or three lines” (slightly over four to six millimetres), potentially matching Owen’s (1842: 165) description better than the longer, over 15 mm striations in Pleurosternon bullocki (TS personal observation). In any case, the sculpture of the carapace of MZ VIII Vr-71, which is inconsistent with that of “ Tretosternon ” was noted multiple times ( Lapparent de Broin and Murelaga 1999; Lapparent de Broin 2001; Joyce et al. 2011; Anquetin et al. 2017; Joyce 2017). As a result, modern authors considered the specimen either an indeterminate turtle ( Lapparent de Broin and Murelaga 1999; Lapparent de Broin 2001; Joyce et al. 2011) or an indeterminate “plesiochelyid” ( Anquetin et al. 2017; Joyce 2017; Tyborowski and Błażejowski 2019a, b). Indeed, the external surface of MZ VIII Vr-71 is mostly smooth, even despite it being covered by numerous larger, irregular pits of varying depths ( Figs. 2A 1 , 3 , 4A 1, B 1, C 1 , 5A–E, F 1, G 1 ). Similar pits occur commonly in European Jurassic aquatic turtles, regardless of their taxonomic affinities, and their distribution and morphology is asymmetric and inconsistent between individuals (e.g., Rütimeyer 1873; Bräm 1965; Lapparent de Broin et al. 1996; Meyer 2011; Anquetin et al. 2014b; Püntener et al. 2014; Anquetin and Püntener 2020; TS personal observation). Thus the pits likely have extrinsic origins (pathologic, taphonomic, a result of weathering or preparatory work) and lack taxonomic value. Very few fragments exhibit finer, more concentrated pitting on the surface ( Fig. 5A, D, F 1 ), but its extent is very limited, its origin appears to be taphonomic and the dorsal surface of the fragments is in some cases coarse ( Fig. 5F 1 ). This matter will be tackled in a separate paper (Tomasz Szczygielski, Justyna Słowiak, Irena Raselli, and Dawid Surmik, unpublished material).

The smoothness of the shell and the lack of consistent, small-scale ornamentation distinguishes MZ VIII Vr-71 from helochelydrids and Jurassic and Berriasian European pleurosternids, Dorsetochelys typocardium , Pleurosternon bullocki , “ Pleurosternon ” portlandicum Lydekker, 1889, Riodevemys inumbragigas , and Selenemys lusitanica (see Owen 1842, 1853; Lapparent de Broin and Murelaga 1999; Milner 2004; Joyce et al. 2011; Pérez-García and Ortega 2011; Scheyer et al. 2014; Pérez-García 2014; Pérez-García et al. 2015; Herzog 2019; Joyce and Anquetin 2019). The shell is relatively thick, differentiating it from the “eurysternids” (e.g., Anquetin et al. 2017; Püntener et al. 2020). There is no trace of costoperipheral fontanelles, further refuting attribution to numerous “eurysternid” taxa, Achelonia formosa , Thalassemys spp. , Tropidemys seebachi , Jurassichelon spp. , Neusticemys neuquina , and Pelobatochelys blakii (see above). The shape of the neurals is more regular than, e.g., in platychelyids ( Meyer 1860; Wagner 1861; Bräm 1965; Lapparent de Broin 2001; Sullivan and Joyce 2017). MZ VIII Vr-71 is therefore most similar to “plesiochelyids”, namely Craspedochelys spp. and Plesiochelys spp. The only identifiable characters pointing towards a Craspedochelys spp. affinity are the anteroposterior dimension of the first marginal, which in MZ VIII Vr-71 was smaller compared to the second marginal, unlike the more bulbous and protruding first marginal of Plesiochelys etalloni ( Fig. 6B, C ), a well-defined contact between the second marginal and the first vertebral (absent in Plesiochelys spp. , but present at least in Craspedochelys picteti NMS 9149, Fig. 6B– D ) and the proportions of the costals. The shape of the first marginal may be of little taxonomic value, since some variation of that nature is present in thalassochelydians, in general, and “plesiochelyids”, in particular (e.g., Anquetin et al. 2014b; Püntener et al. 2017b). The longest, most complete (although still broken distally) costal (either the second or the fourth), however, attains the length/width ratio of 3.7. This already exceeds the ratios given for the fourth costal by Anquetin et al. (2014b) for Plesiochelys etalloni (3.1–3.6), encompassing also Plesiochelys bigleri (see Püntener et al. 2017a; Raselli and Anquetin 2019), and it seems realistic that, when complete, it could have reached the range within Craspedochelys spp. (4.3–4.8). Unfortunately, the midsection of the plastron and the bridge are not preserved, so there are no data on whether the bridge was ligamentous or osseous and if any plastral fontanelles were present. The material does not allow any comparisons with Owadowia borsukbialynickae , Plesiochelys planiceps ( Owen, 1842) , Portlandemys mcdowelli Gaffney, 1975a , and Portlandemys gracilis Anquetin, Püntener, and Billon-Bruyat, 2015 . Due to its incompleteness, in accordance with tentative inferences of some previous authors ( Anquetin et al. 2017; Joyce 2017; Tyborowski and Błażejowski 2019a, b), we identify MZ VIII Vr-71 as an indeterminate “plesiochelyid” resembling Craspedochelys spp.

Figures

Fig.2.The thalassochelydian turtle Craspedochelys? sp., MZVIII Vr-71,from the Kimmeridgian of Krzyżanowice.Central part of the carapace (Fig.6A:1), in external (A1, A2) and visceral (A3, A4) views.A1, A3, photographs; A2, A4, explanatory drawings.

Fig. 3. The thalassochelydian turtle Craspedochelys? sp., MZ VIII Vr-71, from the Kimmeridgian of Krzyżanowice, parts of the carapace and plastron likely belonging to a single individual.A. Anterior part of the carapace and plastron (Fig. 6A: 2) in dorsal (A1) and ventral (A2) views.B. Posterior part of the cara- pace and plastron (Fig. 6A: 3) in dorsal (B1) and ventral (B2) views.The fragments likely belong to the same individual as the material shown in Fig. 4A–C.

Fig. 4. The thalassochelydian turtle Craspedochelys? sp., MZ VIII Vr- 71, from the Kimmeridgian of Krzyżanowice. A. Anterior left part of the carapacial rim (Fig. 6A: 4) in external (A1), visceral (A2), and peripheral (A3) views. B. Fragment of the anterior right part of the carapacial rim (Fig. 6A: 5) in external (B1), visceral (B2), and peripheral (B3) views. C. Unidentified part of the carapacial rim in external (C1), visceral (C2), and peripheral (C3) views. D. Fragment of the posterior part of the carapace (Fig. 6A: 6).A–C likely belong to the same individual as the material shown in Fig. 3. The numbers indicate marginal scutes.

Fig. 5. The thalassochelydian turtle Craspedochelys? sp., MZ VIII Vr-71, from the Kimmeridgian of Krzyżanowice, shell fragments. A. Middle part of a second, fourth, or sixth costal in external view. B. Proximal part of a third or fifth costal in external view. C, D. Parts of unidentified costals in external view. E. Parts of two unidentified costals retaining a sutural contact in external view. F. Proximal part of a second, fourth, or sixth costal in external (F1) and visceral (F2) views. G. Probable part of the plastron in external (G1–G2), visceral (G3–G4),?posterior (G5), and?lateral (G6) views. Note that the position of the element within the shell is uncertain, so precise orientation is not possible. G2, G3, G5, and G6 are 3D models based on surface scans in the orthographic view with the Radiance Scaling shader enabled to optimize the lighting and present the raised?lateral edge (asterisk). H. Part of a plastron of a large individual in external (H1) and visceral (H2) view and natural cross-section (H3). Note that the fragments likely belong to various individuals.

Fig. 6. A. Approximate position of the identifiable fragments of the thalassochelydian turtle Craspedochelys? sp., MZ VIII Vr-71, within the shell, in dorsal (A1) and ventral (A2) views (schematic reconstruction of Craspedochelys spp. based on Anquetin et al. 2014, 2017, and TS personal observations, modified to fit the observed morphologies). Numbers represent the elements illustrated in the text: 1, Fig. 2; 2, Fig. 3A; 3, Fig. 3B; 4, Fig. 4A; 5, Fig. 4B; 6, Fig. 4D. Fragments likely belonging to the same individual are indicated by the same colour. B–E. Anterior edges of the carapace of Plesiochelys etalloni (Pictet and Humbert, 1857), NMS 8514/NMS 118 (B) and NMS 8727/NMS 116 (C), Craspedochelys jaccardi (Pictet, 1860), NMS 101 (D), and Craspedochelys picteti (Rütimeyer, 1873), NMS 9149/NMS 608 (E); all from the Kimmeridgian of Solothurn, Switzerland. The numbers indicate marginal scutes. Note the anteriorly protruding first marginal of Plesiochelys etalloni (B, C) and straight anterior edge in Craspedochelys spp. (D, E).

Fig. 1. Geographic distribution of Jurassic marine reptile fossil occurrences in Poland. See Table 1 for details. Silhouettes obtained from phylopic.org: Geosaurinae (Dmitry Bogdanov, CC BY 3.0), Ichthyosauria, Metriorhynchinae, and Teleosauroidea (Gareth Monger, CC BY 3.0), indeterminate reptile clade (public domain), Plesiosauroidea (Adam Stuart Smith, CC BY-SA 3.0), Pliosauridae (Nobu Tamura, CC BY-SA 3.0), and Testudinata (public domain).