A new giant nautilid species from the Middle Jurassic of Luxembourg and Southwest Germany

In comparison to other cephalopods such as ammonites and belemnites, nautilid shells are relatively rare fossils in Jurassic marine deposits and knowledge of their taxonomy is therefore still patchy. We describe herein a new species of Cenoceras, C. rumelangense, from the early Bajocian Humphriesianum Zone of Luxembourg and Southwest Germany. In Luxembourg, the type material occurs in the ‘Marnes sableuses d’Audun-le-Tiche’ unit, which yields other large-sized cephalopods, such as the largest known belemnite genus, Megateuthis. The new species reaches a remarkable size, with diameters up to 610 mm. Thus, it is amongst the largest known post-Triassic nautilids worldwide, together with Paracenoceras giganteum and Paracenoceras ingens from the Upper Jurassic. Additionally, we discuss some aspects of the taphonomy of these large shells deposited in a shallow marine setting.

The nautilids that evolved during and after the radiation of the Early Jurassic are difficult to classify due to their relative scarceness and their morphological similarity (Barroso-Barcenilla et al., 2016).Historically, they were mostly included in the basket genus Cenoceras, though later authors subdivided this group into several subgenera (e.g.Tintant, 1984).The fossil record of early Middle Jurassic nautilids in the north-western Tethys Realm is scarcely documented in the literature.Most species are based on poorly known type specimens first described in the nineteenth century (e.g.Crick, 1898;Foord & Crick, 1890;Morris & Lycett, 1850;Orbigny, 1843;Quenstedt, 1846Quenstedt, -1849;;Sowerby, 1816;Young & Bird, 1822;Zieten, 1830Zieten, -1833) ) and revisions of theses pioneering works have been scarce with a few exceptions (Enay et al., 1994).Aside from the historical works documenting nautilids from the Middle Jurassic of north-western Tethys, some other occurrences have been described from this area more recently.Particularly, Branger (2004) and Rulleau (2008) reported material from France.Exceptionally large-sized nautilid shells have been reported both from the Bajocian of Southwest Germany (Dietl, 1995(Dietl, , 2013;;Schweigert, 2021) and Luxembourg (Gross, 2005); these specimens, together with new unpublished findings, are revised herein and described as a new species.

Material and methods
We studied 18 specimens attributed to Cenoceras rumelangense n. sp., originating from the following collections: MNHNL: Musée national d'histoire naturelle, Luxembourg.SMNS: Staatliches Museum für Naturkunde, Stuttgart, Germany.MEP: Musée Eugène Pesch, Lasauvage, Luxembourg.MNM: Musée national des mines de fer, Rumelange, Luxembourg.Geolor: Museum of the French amateur geologist association Geolor, Thionville, France.Furthermore, for comparative reasons, we considered material described in the literature (Enay et al., 1994;Rulleau, 2008) and preserved in the following collections: MNHN: Musée national d'histoire naturelle, Paris, France.EPF: Espace Pierres Folles, Saint Jean-des-Vignes, France.Our specimens have been measured according to standard methods (e.g.Rulleau, 2008).For comparison purposes, we also included material attributed to Cenoceras gr.obesum (Sowerby, 1816) and Metacenoceras clausum (Orbigny, 1843), with data taken from the literature (Enay et al., 1994;Rulleau, 2008).The morphometric analysis (Figs. 1, 2) was performed on datasets referred to C. rumelangense, C. gr.obesum and M. clausum.First, we conducted an unconstrained ordination in the form of a principal component analysis (PCA) on all four descriptors in their absolute values (diameter, height, width, umbilical width) as well as on ratios of their descriptors (H/D, W/D, W/H, U/D).We chose a scaling 1 for the PCA representations because for the purpose of this study, we are more interested in interpreting the relationships amongst samples rather than amongst their morphometrical descriptors.Indeed, using this scaling, biplot distances amongst our samples are approximations of their Euclidean distances in multidimensional space (Legendre & Legendre, 2012).We concurrently evaluated the contribution of each eigenvalue to the correlations of the PCA (scree plot).Finally, we overlaid a dendrogram based on hierarchical agglomerative method using the weighted pair-group centroid method (WPGMC; Gower, 1967) on the PCA biplot to better interpret the distances between samples.We chose WPGMC because it avoids biases towards the over-represented C. rumelangense group (Legendre & Legendre, 2012).PCA analyses were run using the vegan (Oksanen et al., 2018) package in R (R Core Team, 2022).

Baden-Württemberg, Germany
In Baden-Württemberg, Middle Jurassic deposits occur at the foothill of the Swabian Alb Mountains and in its foreland.The Bajocian of Swabia is made up of sandy limestones, clays with intercalated limestone beds and ferruginous oolites.The debris of the younger Upper Jurassic limestones mostly covers these deposits.Since the Bajocian rocks do not have any economic value, active clay pits or quarries do not exist and even in the past, only locally some parts of the section were temporarily quarried for building or pavement stones.The Bajocian of Swabia is subdivided into several formations (Bloos et al., 2006), the Wedelsandstein Formation at the base (formerly 'Braunjura Gamma'), followed by the Ostreenkalk and the Hamitenton formations (formerly 'Braunjura Delta').In the southwestern part of the Swabian Alb, iron oolites become predominant and form the Gosheim Formation (Dietze et al., 2015).Towards the east and in adjacent Franconia, the Bajocian is represented by the highly condensed, partly iron-oolitic Sengenthal Formation (Dietze et al., 2017).The rich fossil content was described in classical monographs by Quenstedt (1856Quenstedt ( -1857) ) and Oppel (1857).Subsequently, Engel (1908) provided faunal lists of each formation ('Braunjura Gamma' and 'Braunjura Delta' correspond to the Bajocian).Ammonites, besides belemnites, bivalves, brachiopods, serpulids, and gastropods dominate the rich invertebrate fauna.In the middle Swabian Alb, an almost continuous section was exposed at Gruibingen during the construction of the motorway A8.At that occasion, the fossil content of the entire Middle Jurassic section was documented, however, with a clear focus on ammonite biostratigraphy (e.g.Dietl, 1988Dietl, , 2006Dietl, , 2007Dietl, , 2013)).
In contrast to Luxembourg, the Bajocian rocks of Swabia were deposited in a deeper marine shelf setting; scleractinian corals are restricted to a few locally developed thin horizons where they formed coral meadows (Bosch et al., 2013).The Ostreenkalk Formation is characterized by an alternation of clays, marly limestones and intervals of iron-oolitic marls and limestones containing abundant oysters (Actinostreon, Liostrea) and other thick-shelled bivalves (Ctenostreon) as well as giant belemnite rostra (Megateuthis).Stephanoceratids and other age-diagnostic ammonites date this formation into the Humphriesianum Zone.Nautilids, mostly represented by species of Cenoceras and Metacenoceras, are generally rare in the Bajocian of Swabia, but this may be due to the small outcrops and the focus on the collection of ammonites as the most attractive and iconic fossils within these beds.In the Ostreenkalk Formation of the Gruibingen section, the herein reported giant specimen was the only nautilid record.
Genus Cenoceras Hyatt, 1883 Type species.Nautilus orbignyi Prinz, 1906(= Nautilus intermedius J. Sowerby, 1816in Orbigny 1843, pl. 27), by original designation.Remarks.There is great confusion about the nature, type horizon and stratigraphical range of Nautilus intermedius J. Sowerby, 1816, which has been controversially interpreted by numerous authors (for a recent overview see Țibuleac et al., 2020).Tintant (1984), who tried to trace the illustrated specimen in the Orbigny collection, reviewed the Early Jurassic representatives of the genus Cenoceras.He assigned a specimen of this collection originally labelled as Nautilus striatus to Sowerbyʼs species Nautilus intermedius; however, this specimen has a much wider umbilicus than the specimen illustrated on Orbignyʼs plate 27, to which Hyatt (1883) explicitly referred when erecting his new genus Cenoceras.Tintant (in Enay et al., 1994) suspected that specimens were mixed up and that the illustration of plate 27 shows a chimaera by artificial combination of characters from various specimens of species of different ages and localities.Finally, he agreed that Nautilus intermedius sensu Orbigny was conspecific with Nautilus intermedius J. Sowerby, 1816.However, interpreting the specimen illustrated on Orbignyʼs plate 27 as a non-existent taxon would make Hyattʼs genus Cenoceras invalid, since Tintant did not formally designate a type for Nautilus orbignyi amongst Orbignyʼs Nautilus intermedius specimens.Treating N. orbignyi Prinz as a "nomen nudum" as suggested by Tintant (1984) is impossible, because its holotype is by definition the illustrated specimen of Orbignyʼs plate 27.
Of course, it is well-known that the illustrations of Orbignyʼs plates are highly idealized and standardized (the apertures always pointing to the left), but he provided measurements of a concrete specimen, most likely of the illustrated one.Amongst the 15 surviving specimens in the Orbigny collection labelled as Nautilus intermedius, the closest morphological resemblance is with a specimen (MNHN-F-A02204) from the Toarcian of Avallon (Fig. 4) (Tintant in Enay et al., 1994).Tintant determined this specimen as Cenoceras astacoides (Young & Bird, 1822), a species originally described from the Toarcian of Yorkshire, United Kingdom (Howarth, 1962).MNHN-F-A02204 is the sole specimen of the surviving Nautilus intermedius material in Orbignyʼs collection that originates from the vicinity of a locality cited by dʼOrbigny (Tintant in Enay et al., 1994).Hence, we conclude that MNHN-F-A02204 was indeed the specimen on which Orbignyʼs plate 27 is based and which was subsequently assigned to a new species by Prinz, (1906), Nautilus orbignyi.Nautilus orbignyi Prinz, 1906 is here considered a subjective junior synonym of Nautilus astacoides Young & Bird, 1822 and, more importantly, the widely used genus Cenoceras can still be used in the strict sense of Tintant.We concur with Tintant (1984) that Nautilus intermedius Sowerby, 1816 is a species from the Sinemurian.The illustration of the holotype, however, is a very schematic drawing.Obviously this holotype is untraceable, because Foord (1891, p. 192) referred to several specimens stored in the collection of the British Museum but explicitly not to Sowerby's illustrated specimen.Even the type locality of Sowerbyʼs Nautilus intermedius, Keynsham in Somersetshire, was only cited with a question mark as a doubtful occurrence of this species.Tintant (in Enay et al., 1994) illustrated two nautilids of Orbignyʼs collection (MNHN-F-R09436 and MNHN-F-B48672) that he identified as Cenoceras intermedium (Sowerby, 1816), however, these specimens were originally labelled as Nautilus striatus Sowerby and did not belong to the series of N. intermedius in Orbignyʼs collection.Thus, replacing the type species of Cenoceras Hyatt, 1883 by Nautilus intermedius Sowerby, 1816 without legitimation by the International Commission of Zoological Nomenclature was a clear violation of the principles of zoological nomenclature (ICZN, 1999) and cannot be accepted.
Type locality: Rumelange, abandoned quarry site ʽBerenskaul-Waisskaul' on the Luxembourgish-French border (Coordinates 49°27′12ʺN and 6°00′41ʺE; Bintz et al., 1973, p. 159) (Fig. 3).Type horizon: 'Marnes sableuses d' Audun-le-Tiche' (dom4), lower Bajocian, Humphriesianum Zone.Diagnosis: Giant-sized, narrow-umbilicate species of Cenoceras with strongly inflated subtrapezoidal whorl section, generally higher than wide; longitudinal striation on the flattened venter and on the outer third of the flank.Description: The holotype (Fig. 5) is a steinkern with a few preserved remnants of the shell.The body-chamber occupies half of the last whorl.The last septa are more closely spaced, indicating that the specimen reached the adult age.The venter is distinctly flat and is ornamented with dense longitudinal striae that are particularly well developed on the body-chamber and which extend on the outer third of the flanks.The section of the bodychamber is markedly subtrapezoidal, compressed (compression index is 0.81 on the body-chamber and 0.77 on the phragmocone); the compression is more marked in younger whorls and less strong on the body-chamber but does not affect the lateral dimensions.The flanks are slightly rounded on the inner third (located towards the umbilicus) and distinctly flattened to slightly concave  Ontogeny: Growth pattern lacks principal changes of coiling or other modifications during ontogeny, except the inflated whorl section gradually changing from being rounded in the juvenile stage to subtrapezoidal in the medium and adult stages.In the latest stage, the chambers and suture lines are narrowing thus indicating adult size.Ornamentation: The almost uncompressed specimen from Swabia (Fig. 9) is the only one with a completely preserved calcitic replacement shell.The adult shell is ornamented with dense longitudinal striae both on the venter and on the outer third of the flank.The aperture does not exhibit any apertural modifications and crosses the venter in a straight way.Due to the shell preservation, the beginning of the body-chamber and the suture lines of the phragmocone cannot be traced.Measurements and morphometrics: Various measurements and parameters of C. rumelangense specimens are reported in Table 1.We compared the data with measurements and parameters published for two coeval

PCA biplot for absolute morphometric descriptor values (scaling 1)
Although the three groups seem separated in the biplot (Fig. 1), the eigenvalues showed that the first axis was explaining much more than the following axes for the absolute values PCA.This means that the clausum and obesum groups are closer to each other than they are to the rumelangense group.This is also supported by the clustering analysis (Fig. 1).Assuming that there are indeed three groups, the clustering analysis puts specimen EPF Baj74 (in Rulleau, 2008, p. 52) with the rumelangense group, not with the obesum group in which it was originally classified (Fig. 1).Also, the specimens attributed to the rumelangense group, MNHNL BM370 (sample 19) and MNHNL BM783 (sample 25), were clustered together with the obesum group.

PCA biplot for ratios of morphometric descriptor values (scaling 1)
Three groups are separated in the biplot (Fig. 2), the eigenvalues showed that the three first axes are relevant (see inset Fig. 1) for the ratio values PCA.Clustering results overlay well with PCA results and original grouping, aside from specimen EPF Baj74 (sample 13), which is associated with the rumelangense group again (Fig. 2).
For the PCA on absolute values, the proportion explained by the principal components 1 and 2 were 94% and 4%, respectively.For the PCA on absolute values, the proportion explained by the principal components 1 and 2 were 65% and 21%, respectively.
Remarks.In first view, the herein described Cenoceras seems to fit well with the description and illustration of Nautilus obesus Sowerby, 1816 (p. 124).However, the provided illustration is very schematic and, more important, the original of the illustrated specimen could not be traced (Foord & Crick, 1890).Thus, Foord and Crick (1890) re-interpreted this species based on further material.Interestingly, they noted that in one of the studied specimens they had assigned to Sowerbyʼs species, the shell is preserved and it is said to have a shell showing only growth lines.In contrast, in our material with preserved shell, as mentioned above, a strong longitudinal striation is present at least in the ventral and ventrolateral part of the conch, most likely expanding to the entire conch in the juvenile stage as in other Cenoceras (s.str.)species.We therefore suspect that at least some of the material of Nautilus obesus sensu Foord & Crick does not represent Cenoceras s.str.but Metacenoceras Tintant, 1984 (Metacenoceras, previously considered a subgenus of Cenoceras was given full genus rank in Dietze et al., 2021. p. 48).This is corroborated by the fact that Foord and Crick (1890) included Orbignyʼs Nautilus lineatus in the synonymy of N. obesus Sowerby.For Nautilus lineatus Sowerby in Orbigny, Tintant (in Enay et al., 1994) introduced the new species Cenoceras (Metacenoceras) moutierense pointing out the differences with the incomplete specimen illustrated by Foord & Crick (1890, fig. 11) as Nautilus obesus.Although the determination of the latter was said to go back to the times of J. Sowerby, it is unknown whether it belonged to the type series.Moreover, this specimen has not preserved its shell and the entire body-chamber is missing.Although a conspecificity of our specimens with Sowerbyʼs original Nautilus obesus cannot be ruled out, this is unprovable due to the apparent loss of the type, and consequently, we here regard Nautilus obesus J. Sowerby, 1816 as a nomen  2008) illustrated several large-sized nautilids from the Humphriesianum Zone of France as Cenoceras ex gr.obesum.His material cannot be assigned to Sowerbyʼs species for the same reasons.Comparing Rulleauʼs measurements with those of the herein described material, however, clearly indicates that his material differs significantly from ours and must represent another species, with the exception of specimen EPF Baj74, whose measurements fall into the variation of C. rumelangense as shown by our cluster analysis.Pending a verification of the original specimen in the Espace Pierres Folles collection, we tentatively include this specimen in our new species, with a question mark.We furthermore consider it likely that the specimens illustrated by Grulke (2016, p. 139) from the upper Bajocian of Frogden Quarry, Sherborne represent large-sized specimen of C. rumelangense.If confirmed, these records would represent the largest specimens known, with 700 resp.750 mm of diameter (Grulke, 2016, p. 139).
Occurrence: Cenoceras rumelangense n. sp.occurs in the Humphriesianum Zone (lower Bajocian, Middle Jurassic) of Luxembourg and Southwest Germany (Baden-Württemberg). Unpublished data suggest a possible occurrence in coeval strata of North Switzerland (personal observation by GS).Clustering analyses performed in this study furthermore hints at the occurrence of the taxon in the Rhône Valley, France (specimen EPF Baj74 in Rulleau, 2008, p. 52).Unstudied specimens from Sherborne, UK, that were illustrated by Grulke (2016) might also belong to C. rumelangense and thus potentially extend the geographic occurrence towards the Northwest of Europe.

On the occurrence of 'giant' nautilids in the Jurassic
'Gigantism' is an ecologically important trait associated with competitive superiority; although it has been studied in specific cases, the general conditions at the origin of a very large body size remain still obscure (Vermeij,  (2015), a giant species must fulfil the following three criteria: (1) species of the respective size should not have occurred all the time.There should be a low number of giant species; (2) adults of giant species have to be much larger than the average of their group; (3) the largest specimen(s) of the respective species should not show strong signs of pathology or abnormal growth.Pathological gigantism (Manger et al., 1999) can be ruled out in the case of C. rumelangense, as none of the studied specimens shows any abnormalities in growth that would hint at a pathological development.In addition, there are several specimens that reach 500 mm diameter or more (Table 1), making the species outstanding by size not only for the genus Cenoceras but also in general for Mesozoic nautilids.
Living nautilids, namely Nautilus pompilius, N. vitiensis, N. samoaensis, N. vanuatuensis only rarely reach diameters larger than 200 m (Barord et al., 2023), N. belauensis and the questionable taxon N. repertus being credited with the largest shell diameters reaching up to 240 mm (Saunders, 1987).Compared to living nautilids, Mesozoic taxa reach considerably larger dimensions.Large-sized nautilids (> 300 mm) occasionally occur throughout the Jurassic.Already Zieten (1831) reported large-sized specimens from the Lower and Middle Jurassic of Swabia with diameters of "anderthalb Schuh" (c.430 mm) and named them Nautilites giganteus Schübler.The collection of the National Museum of Natural History Luxembourg contains a large nautilid (330 mm; MNHNL n° SI340), ascribed to the genus Cenoceras from the 'Marnes et calcaires de Strassen' unit (Sinemurian, Lower Jurassic).The Stuttgart Natural History Museum houses a specimen of Hemicenoceras arare (Dumortier, 1869) from the early Pliensbachian of Swabia with a diameter of 350 mm (SMNS 70673).In the Middle Jurassic, Branger (2004) reported a large specimen (440 mm) of Cenoceras mazardrikense from the Callovian of western France.The specimens of C. rumelangense reported herein are thus considerably larger than these previously reported 'giant' nautilids.
The genus Paracenoceras Spath, 1927 is also known for some giant species, such as Paracenoceras giganteum (Orbigny, 1843) and P. ingens Tintant et al., 2002 from the Kimmeridgian (Upper Jurassic).Large-sized specimens of Paracenoceras spp. with diameters up to 500 mm are cited from southern Germany, Pomerania and France (Loesch, 1914;Kuhn, 1936;Tintant in Enay et al., 1994;Schweigert, 2021;Schlampp, 2022).Thurmann and Étallon (1861) noticed a diameter of 7 dm for Nautilus giganteus.Tintant et al. (2002) reported that some paracenoceratids from Switzerland could reach 600 mm or even more.However, the largest specimen measured by Tintant et al. (2002) was a Paracenoceras ingens of 522 mm diameter.To our knowledge, nautilids from the Cretaceous and younger ones did not reach diameters over 450 mm.Thus, the specimens of C. rumelangense n. sp. reported herein (Tab. 1) appear to be the largest post-Triassic nautilid documented so far in the scientific literature.

Taphonomical remarks
All of the herein studied specimens of Cenoceras rumelangense n. sp.show an extensive overgrowth of epifauna.Amongst the most common organism are serpulids (Dorsoserpula lumbricalis), small oyster shells (Liostrea) and bryozoans.Other bivalves, such as Eopecten abjectus (Phillips, 1829) are also present in some specimens.Typically, the epifauna is present on both flanks of the shell.An overview about benthic fauna from the 'Marnes sableuses d' Audun-le-Tiche' unit in Rumelange is given by Fayard et al. (2005).
In the specimen from Swabia (SMNS 62245), epizoans are represented by serpulids (Dorsoserpula lumbricalis [Schlotheim, 1820]) and bryozoans (Berenicea, Kololophos), whereas oysters are missing.Only the venter of the adult body-chamber is free of epifauna.This strongly points to a post-mortem overgrowth, when the empty shell was in an upright position at the seafloor forming a benthic island in a muddy substrate, otherwise the ventral area of the shell would have been overgrown as well.Similar observations were made in Lower Jurassic representatives of Cenoceras (Evans & King, 2019).

Conclusions
The study of previously undescribed giant nautilids from the lower Bajocian Humphriesianum Zone of Luxembourg and Southwest Germany allowed the identification of a new taxon of the genus Cenoceras, introduced herein as Cenoceras rumelangense n. sp.Clustering analysis support the distinction between C. rumelangense and coeval taxa Metacenoceras clausum (Orbigny, 1843) and Cenoceras gr.obesum (Sowerby, 1816sensu Rulleau, 2008).C. obesum (Sowerby, 1816) is treated as nomen dubium.A comparison with measurements from the literature suggests that C. rumelangense is amongst the largest post-Triassic nautilids, together with Paracenoceras spp.from the Upper Jurassic.These large-sized shells also provided excellent support for encrusting organisms (serpulids, bryozoans, oysters and other bivalves), thus acting as so-called benthic islands.
Watrinelle †, Eugène Pesch †, Jo Simon, Robert Haas and Jean Valentiny †.The management of the Cimalux quarry in Rumelange and especially the site manager Bérangère Barthe are thanked for granting access to the quarry and encouraging research on the fossils that occur therein.We also thank Frédérique Humbel and Raffaella Campobasso for providing access to the collections of the Museum Pesch in Lasauvage and the National Mining Museum in Rumelange, respectively.Paul Braun (Musée national d'histoire naturelle, Luxembourg) provided valuable assistance with photographs of the Luxembourgish specimens.A special recognition also goes to Svetlana Nikolaeva (Natural History Museum, London) and a second anonymous reviewer whose remarks contributed to enhance the manuscript.

Fig. 1
Fig. 1 PCA biplot for absolute morphometric descriptor values (scaling 1).Clustering results are overlaid and represented by grey lines.Colour/shapes represent a priori grouping to species level

Fig. 2
Fig. 2 PCA biplot for ratios of morphometric descriptor values (scaling 1).Clustering results are overlaid and represented by grey lines.Colour/shapes represent a priori grouping to species level

Fig. 3
Fig. 3 Stratigraphic position (marked by a black arrow) and geographic location of the type locality (marked by a red circle) nearby Rumelange.Map of Southwest Luxembourg and stratigraphic log: Service géologique du Luxembourg (www.geolo gy.lu)