The soft tissue and skeletal anatomy of two Late Jurassic ichthyosaur specimens from the Solnhofen archipelago

Geological setting

The ichthyosaur specimens examined here originate from the locality Eichstätt-Blumenberg of Bavaria, southeastern Germany. The strata are the early Tithonian (Late Jurassic) Eichstätt Plattenkalk, of the Kimmeridgian to early Tithonian Solnhofen Archipelago (Fig. 1). At the time of deposition, the area was situated around 34 degrees north (Munnecke, Westphal & Kölbl-Ebert, 2008), at the northern margin of the Tethys Ocean. The Plattenkalk deposits are carbonate successions of finely laminated limestone deposited in a series of subtropical, shallow to deeper basins within a huge carbonate platform (Kölbl-Ebert, Röper & Leinfelder, 2005). The carbonate platform was formed largely by sponge-microbial reefs and accumulated carbonate sands, and locally by small coral reefs (Koch, Senowbari-Daryan & Strauss, 1994; Munnecke, Westphal & Kölbl-Ebert, 2008). The bottom waters of the basins within the platform were primarily stagnant, and in some cases possibly characterized by an elevated salinity and microbial films leading to Plattenkalk deposition, which might be the cause of the excellent fossil preservation (Barthel, 1970; Gäb et al., 2020; Munnecke, Westphal & Kölbl-Ebert, 2008).

Systematic paleontology

Ichthyosauria de Blainville, 1835

Parvipelvia Motani, 1999

Thunnosauria Motani, 1999

Ophthalmosauridae Baur, 1887

Aegirosaurus Bardet & Fernández, 2000

Aegirosaurus sp.

Referred specimen: Nearly complete skeleton JME-SOS-08369 (Figs. 2, 5, 6, 7, Table 1). The specimen is housed in the Jura-Museum Eichstätt, Eichstätt, Germany.

Locality: The visitor quarry in Eichstätt-Blumenberg.

Age: Early Tithonian (eigeltingense horizon), Late Jurassic.

Taxonomic assignment

The nearly complete specimen JME-SOS-08369 can be referred to Ophthalmosauridae with confidence, based on the extensive lateral exposure of the angular (Fernández & Campos, 2015; Moon, 2017). When compared to Aegirosaurus leptospondylus, JME-SOS-08369 shares the following diagnostic characters (Bardet & Fernández, 2000): a snout not markedly demarcated from the skull, delicate and small, densely packed teeth strongly anchored in a dental groove, an enamel crown that is minutely ridged or smooth, a medium sized orbit (orbital ratio 0.22, in between the two Aegirosaurus specimens), a jugal extending anteriorly to the orbit, and a long and very slender rostrum. The skull and total body length fall within the original description of the species (Bardet & Fernández, 2000). However, the status for more than half of the diagnostic characters cannot be accessed in JME-SOS-08369, including the shape and size of the naris, orbit and sclerotic plates, traits in the postorbital/cheek region, as well as all diagnostic postcranial characters. JME-SOS-08369 differs from Aegirosaurus leptospondylus in one aspect: it seems to have a maxilla with relatively long lateral exposure, whereas in Aegirosaurus leptospondylus the exposure is small (Bardet & Fernández, 2000) (L. L. Delsett, 2015, personal observation, SNSB-BSPG 1954 I 608). In conclusion, based on the available data, we refer JME-SOS-08369 to Aegirosaurus sp, pending further review.

Taphonomy JME-SOS-08369

The specimen (Fig. 2) is articulated and preserved in its entire length, missing only the dorsal portion of the caudal fin. Its total length as preserved is 1,610 mm from the tip of the snout to the tip of the ventral caudal fin lobe measured in a straight line. During or after death, the animal landed dorsolaterally on the sea floor. Preservation of the skeletal elements is rather poor. The skull (Fig. 5) is compressed, and is seen from the right side Its total anteroposterior length is 440 mm. The jaw consists of the tooth-bearing right premaxilla and maxilla seen in lateral view, as well as the right dentary, surangular, splenial, and angular. Not all sutures in the skull are well preserved, and the accurate shape of certain elements, such as the maxilla, is not entirely clear. The posterior portion of the left lower jaw elements are preserved and seen in medial view, and a small portion of the right splenial is preserved in lateral view. The lower jaw extends slightly beyond the upper jaw anteriorly, but it is not known whether this is an actual underbite or a taphonomical artifact. Anterior to the orbit and in the posteriormost portion, the skull has suffered considerable distortions, and only a partial postorbital is preserved. The preserved soft tissue is described below.

Description of JME-SOS-08369 and comparison to species within Ophthalmosauridae

Premaxilla

The premaxilla is slender, laterally convex, and increases in dorsoventral height posteriorly. Several anteroposteriorly elongated foramina are situated in the anteriormost portion of the element. Posterior to them is the fossa premaxillaris which extends to the posterior end of the premaxilla, as in Palvennia hoybergeti and Ophthalmosaurus icenicus (Delsett et al., 2018; Moon & Kirton, 2016). In contrast, Sveltonectes insolitus bears a continuous groove (Fischer et al., 2011).

Maxilla

In lateral view, the maxilla extends significantly anterior to the anteriormost border of the naris and possesses a processus narialis. The maxilla is exposed for 40% of the total length of the premaxilla and maxilla, similar to the exposure in Undorosaurus? kristiansenae (38%) and Janusaurus lundi, and in contrast to Aegirosaurus leptospondylus and Sveltonectes insolitus, which have small maxillae relative to the other elements (Bardet & Fernández, 2000; Druckenmiller et al., 2012; Fischer et al., 2011; Roberts et al., 2014). The maxilla is dorsoventrally taller than in Undorosaurus? kristiansenae (Druckenmiller et al., 2012).

Jugal

Anteriorly, the jugal only slightly exceeds the orbital margin, as is also the case for Aegirosaurus leptospondylus (Bardet & Fernández, 2000) and Ophthalmosaurus icenicus (Moon & Kirton, 2016). Its anterior portion is of similar height throughout, as in Athabascasaurus bitumineus, but is not as slender as in Leninia stellans (Druckenmiller & Maxwell, 2010; Fischer et al., 2013). The anterior portion of the jugal is more strongly curved than the jugal of Undorosaurus? kristiansenae and Leninia stellans, in which it is nearly straight (Druckenmiller et al., 2012; Fischer et al., 2013). The curvature of the element is more similar to Palvennia hoybergeti, but with a shorter anterior reach (Delsett et al., 2018).

Angular

The angular is extensively exposed laterally as in Aegirosaurus leptospondylus and most other ophthalmosaurids (Fernández & Campos, 2015), in contrast to the short anterior reach in Palvennia hoybergeti (Delsett et al., 2018). The element reaches approximately as far anteriorly as the surangular, which is slightly anterior to the posterior portion of the premaxilla. The angular increases in dorsoventral height in the posteriormost portion, ventral to the orbit.

Dentary

The dentary increases only slightly in dorsoventral height posteriorly, resulting in a relatively slender element, which is more similar to Aegirosaurus leptospondylus than to Palvennia hoybergeti and Ophthalmosaurus icenicus (Bardet & Fernández, 2000; Delsett et al., 2018; Moon & Kirton, 2016).

Dentition

On the right side, 18 teeth are preserved in the maxilla, 17 in the premaxilla and approximately 20 in the dentary, which means that the specimen has more maxillary teeth than reported for Aegirosaurus leptospondylus (Bardet & Fernández, 2000). The teeth sit closely together in a groove that does not seem to be partitioned into alveoli. The teeth are small, straight, and gracile, more similar to Keilhauia nui and Aegirosaurus leptospondylus (Bardet & Fernández, 2000; Delsett et al., 2017) (SNSB-BSPG 1954 I 608, L. L. Delsett, 2015, personal observation) than to Ophthalmosaurus, Undorosaurus? and Brachypterygius extremus (Druckenmiller et al., 2012; Kirton, 1983; Moon & Kirton, 2016). The margin of the enamel is straight, and the enamel does not have deep ridges.

Pectoral girdle

Three incomplete, slender elements are interpreted as the remains of one or two clavicles, and possibly the interclavicle (Fig. 2). A larger, flat element is interpreted as the coracoid, but information on notches and facets are unavailable.

Humerus

Two partial humeri (Fig. 2) with associated distal elements (Fig. 6B) are preserved. The right humerus is broken and only visible in cross-section. Six forelimb elements with unknown identity are preserved on top of, and in immediate proximity to the humerus, and eleven elements from the left forelimb are preserved partly disarticulated close to the left humerus. All elements are rounded in dorsal or ventral view. Based on their size and position, it is assumed that the elements originate from the proximal portion of the limb. If this is correct, it means that they differ from Aegirosaurus leptospondylus (Bardet & Fernández, 2000) (SNSB-BSPG 1954 I 608, L. L. Delsett, 2015, personal observation), where the distal elements on the forelimb are almost rectangular, except for the distalmost third of the limb. Rectangular forelimb elements are also found in Platypterygius australis, Platypterygius hercynicus, Undorosaurus? kristiansenae and Sveltonectes insolitus, whereas Ophthalmosaurus icenicus and Palvennia hoybergeti possess oval and circular elements (Delsett et al., 2018; Druckenmiller et al., 2012; Fischer et al., 2011; Kolb & Sander, 2009; Moon & Kirton, 2016; Zammit, Norris & Kear, 2010).

Vertebral column

Most of the vertebral column is covered in fossilized soft tissue, and the centra are flattened and partly decomposed (Figs. 2, 7B, 7E). The best preserved are the posterior dorsal and the caudal vertebrae. In JME-SOS-08369, the vertebrae are larger anterior to the tail bend than posterior to it. A complete series of postflexural vertebrae is preserved, with approximately 70 vertebrae. The number of postflexural centra and their shape is similar to the isolated tail JME-SOS2183 described below. Two neural arches are interpreted as those of the atlas-axis and cervical vertebra 3.

The ribs are preserved articulated (Fig. 2). Most ribs do not preserve the proximal portion, but where available, they are bicipital. They possess longitudinal grooves along the entire length, giving a figure eight cross section, as is found in the majority of ophthalmosaurids, but different from Acamptonectes densus, Mollesaurus periallus, and three specimens (PMO 222.667, PMO 224.252, PMO 222.670) from Spitsbergen (Delsett et al., 2017; Delsett et al., 2019; Fernández, 1999; Fischer et al., 2012).

The abdomen (Figs. 6A; 7C, 7D) preserves the remains of an almost complete set of gastralia close to their position in life, probably from the left side, directed slightly posteriorly. Gastralia from the right side are situated on top of the ribs, perpendicular to the direction of the ribs. Their anteroposterior width is less than 5 mm, which is less than half the width of the ribs, and the thickness is uniform throughout. Gastralia in ophthalmosaurids are described in Ophthalmosaurus icenicus and specimens from the Slottsmøya Member of Spitsbergen (Delsett et al., 2018; Delsett et al., 2017; Moon & Kirton, 2016; Roberts et al., 2014). In all of these, the gastralia are circular to subcircular in cross-section and with a smaller diameter compared to the thoracic ribs, as in the specimen described here. The preservation of a nearly complete gastral basket in JME-SOS-08369 is unique among ophthalmosaurids.

Pelvic girdle

Two ilia (Figs. 2, 6C) are preserved, and a few elements that might originate from the hindlimbs are scattered nearby. The ilium is approximately 40 mm in dorsoventral height, slender and straight. Aegirosaurus leptospondylus and Ophthalmosaurus icenicus resemble this specimen in having ilia with a similar anteroposterior width overall, but their ilia are more curved (Bardet & Fernández, 2000; Moon & Kirton, 2016). In contrast, the ilia of Keilhauia nui, Athabascasaurus bitumineus, and Undorosaurus? kristiansenae all have one anteroposteriorly expanded end (Delsett et al., 2017).

Ichthyosauria de Blainville, 1835

Referred specimen JME-SOS2183 (Fig. 3, Table 1). The specimen is housed in the Jura-Museum Eichstätt, Eichstätt, Germany.

Locality The Eichstätt-Blumenberg area. The exact location is not known.

Age early Tithonian, Late Jurassic.

Taxonomic assignment

JME-SOS2183 (Fig. 3) only consists of a tail. There are no diagnostic characters in the tail for Ophthalmosauridae, and the specimen is thus referred to Ichthyosauria indet. Based on its strong similarity to other, more complete specimens from Solnhofen, it most likely belongs in Ophthalmosauridae.

Description and comparison JME-SOS2183

The specimen preserves the peduncle, and the entire caudal fin posterior to it (Fig. 3). As the nearly complete specimen JME-SOS-08369, it preserves a soft tissue outline, only described previously in ophthalmosaurids in the neotype of Aegirosaurus leptospondylus and in a specimen lost in WWII (“Tail Munich State Museum” TMSM), both from Solnhofen (Bardet & Fernández, 2000; Bauer, 1898; Merriam, 1908). The soft tissue outline of the caudal fin is better developed in JME-SOS2183 than in JME-SOS-08369 and demonstrates that the tail had the classical lunate shape in lateral view as in other Jurassic ichthyosaurs, such as Stenopterygius specimens from of Holzmaden. As in the nearly complete specimen JME-SOS-08369, the peduncle of JME-SOS2183 is dorsoventrally narrow and widens dorsally and ventrally into the caudal fin. The dorsal lobe tip is dorsoventrally shorter and anteroposteriorly longer than the ventral portion. The tip of the dorsal lobe is also angled more posteriorly compared to the tip of its ventral lobe. This might however be a taphonomic artefact of preservation and/or preparation, as a crack extends through the matrix and fossil at this point.

Vertebral column

Six caudal vertebrae are preserved articulated anterior to the tail bend and are significantly larger (heights 14–18 mm and length 7–8 mm) than those posterior to the bend. The posterior caudal vertebrae are amphicoelous, and equally anteroposteriorly long dorsally and ventrally. The vertebrae forming the tail bend are not preserved.

Approximately 70 small postflexural vertebrae are preserved in articulation. This is similar to the nearly complete specimen JME-SOS-08369 and to the TMSM (Bardet & Fernández, 2000; Bauer, 1898; Merriam, 1908). The neotype of Aegirosaurus leptospondylus preserves the entire tail, with 61 postflexural centra (Bardet & Fernández, 2000). The dorsoventral height and anteroposterior length of all the postflexural vertebrae in JME-SOS2183 are similar, giving a quadratic outline in lateral view (height = 5 mm, length = 5 mm). Consequently, they are anteroposteriorly longer than those anterior to the tail bend, as in the TMSM specimen (Merriam, 1908). The exception are the miniature vertebrae at the very end of the tail. This is similar to some specimens of Ophthalmosaurus icenicus (Moon & Kirton, 2016), but see (Buchholtz, 2001). Height is also slightly larger than length in the postflexural vertebrae of Acamptonectes densus (Fischer et al., 2012: SNHM 1284-R, fig. 8), Sveltonectes insolitus (Fischer et al., 2011), and Arthropterygius (Maxwell, 2010).

Results–characterization of soft tissue

Several different types of soft tissue are preserved in the nearly complete specimen JME-SOS-08369, based on observation in UV, regular and low angle light (Figs. 2, 4, 7, 8). Here, we discuss two types (type I and II, Table 2). In addition, the abdomen and pelvic girdle regions especially preserve additional types of soft tissue that should be considered for analysis. In these regions, soft tissue appears to be preserved with several different colours and textures and might represent muscles, inner organs, connective tissue, or stomach content. A rectangular lump of fossilized soft tissue in the approximate anteroposterior midpoint of the dorsal outline might represent the remnants of a folded dorsal fin (Fig. 7B). No scale impressions are observed in the skin of any of the specimens.

Type I is a thick, outer phosphatic layer, preserved with an off-white color, with yellowish fluorescence (Figs. 7A, 7B, 7E, 7F). It can be observed along the dorsal outline of the specimen starting from the posterior end of the skull, posterior to the pelvic girdle, and in the ventral portion of the tail. During preparation, an additional soft tissue type layered with Type I was observed (Radecker, 2014), but this layer is now preserved only in a few smaller patches. Type II is the most common soft tissue preserved in JME-SOS-08369 (Figs. 2, 7A, 7D–7F). It is an amorphous and light yellow substance infilling large parts of the specimen, covering many of the skeletal elements. Both type I and type II soft tissue are preserved as incrustation pseudomorphs (epimorphs), with no sign of subcellular preservation.

The remnants of soft tissue in the isolated tail JME-SOS2183 are preserved around the entire specimen and was likely enhanced by the addition of some sort of stabilizer/lacquer many years ago (A. Radecker, 2017, personal communication). Type I soft tissue preservation as seen on the leading edge of the caudal fin differs from that on the posterior margin, by being thicker and more robust (Figs. 3A, 3B). A difference in colour might indicate a different tissue composition On the dorsal leading edge, this tissue is segmented.

XRD analyses of material in two samples (JME-SOS-08369-B1 and B2, Fig. 4) that appear off-white in regular light and fluorescent light yellow in UV light, match apatite and is likely referable to type I soft tissue. The same color in UV light is present in the isolated tail JME-SOS2183 in the leading edge of the dorsal lobe of the tail, corresponding to the position of the analyzed samples from JME-SOS-08369, probably representing the same type of mineralized soft tissue. The light beige matrix present on all analyzed samples (JME-SOS-08369-B1, B2 and B3) fits calcite, as expected for limestone.

XRD cannot distinguish between the various members of the apatite group, and chemical data is required to confirm the identification. The semi-quantitative analyses (EDS) (Fig. 8) confirm that the matrix is calcite containing minor amounts of Mg and a silicate mineral, but no phosphates. The material identified via XRD as apatite contains P, Ca, O and F as the main constituents, with minor Na and S, and is consistent with fluorapatite. Apart from Ca and O these elements were only observed in the off-white-colored material. The analyzed samples also have some black material, which is found to be iron oxide with minor Mn. The humeri were covered with a late encrustation of iron hydroxide when excavated (Radecker, 2014).

To determine whether the phosphorous could result from an internal source, we calculated whether the soft tissues in a similar-sized cetacean could produce a sufficient amount of phosphorous for the observed amount of phosphatized soft tissue in the nearly complete specimen JME-SOS-08369 (Supplemental Information). The calculations on the amount of phosphorous in the soft tissue (blubber, skin, muscles) for an ichthyosaur this size, yield a total of 32 g phosphorus. The density of fluorapatite is 3.2 g/cm³, which means the decomposition of the specimen could create 54 cm³ pure fluorapatite. Based on its outline (Fig. 2), the area of the specimen is 6,560 cm², which means that if the apatite formed a single solid crystal under the fossil, it could be 0.8 mm thick. The apatite is not a solid crystal, but will form on the particles within the sediment and pores around the specimen. Consequently, the zone containing apatite could be thicker than 0.8 mm just from phosphorus derived from the soft tissue, even without considering the phosphate derived from the later decomposition of the skeleton. It should be highlighted that by using ideal fluorapatite in our calculation, the amount of apatite that can be formed is a minimum value. For example, the incorporation of carbonate groups for phosphate groups in the apatite structure, would result in more apatite formed, as the content of P in carbonate apatite is smaller than in fluorapatite. That is, the more carbonate that substitutes for phosphorous in the apatite, the more apatite can be formed from the same amount of P. The incorporation of carbonate for phosphorous in apatite also leads to a reduction in density, which means that for the same weight of apatite formed, carbonate-rich apatite will have a larger volume than the same weight of fluorapatite. Consequently, with carbonates substituting for some P in the apatite structure, two effects will increase the volume of apatite that forms.

Soft tissue

Type I soft tissue, found along the margins of the body of the nearly complete specimen JME-SOS-08369 and surrounding the tail in JME-SOS2183, is interpreted as smooth epidermis and dermis, possibly including connective tissue in the tail. Previous analysis found that the dermis and epidermis were multi-layered, as in modern tetrapods (Lindgren et al., 2018), which fits with the observation during preparation of at least two layers of soft tissue (Radecker, 2014).

In the dorsal portion of the tail of the SM neotype of Aegirosaurus leptospondylus, Bardet & Fernández (2000, p. 508) observed minute, rounded ‘scale structures’ interpreted as scales covering the skin, but did not figure or elaborate on them. This contrasts with the specimens described here, which are in line with other recent studies, where ichthyosaur skin is interpreted to be smooth (Lindgren et al., 2018).

Type II soft tissue resembles mostly the light yellow, amorphous substance also observed by Lindgren et al. (2018) in a Stenopterygius specimen (MH432) and interpreted by them as adipocere from decomposed blubber. Under conditions with little or no oxygen, microorganisms produce polyhydroxy fatty acids from fatty tissues from a surplus of organic compounds, commonly known as adipocere (Schoenen & Schoenen, 2013).

The EDS and XRD analyses (Fig. 8) show that the soft tissue in the tail of the nearly complete specimen JME-SOS-08369 is phosphatized, likely by francolite, as is common in vertebrate soft tissue from Solnhofen (Barthel, Swinburne & Morris, 1990). As most of the fossilized soft tissue in vertebrates from the Solnhofen area preserve subcellular details, Wilby (1993) and Wilby, Briggs & Viohl (1995) suggested a supersaturated (probably sedimentary) source of phosphorus, because of depleted levels of sedimentary phosphate adjacent to the fossils (Briggs, 2003; Briggs & Wilby, 1996). This contrasts with the phosphatization in the Santana Formation (Brazil), which is interpreted as authigenic mineralization due to decomposition of remains containing phosphate (Briggs, 2003; Briggs et al., 1993; Martill, 1988). JME-SOS-08369 shows no sign of subcellular preservation. The phosphate mineralization is instead observed as incrustation pseudomorphs (epimorphs), resulting from a process by which the soft tissue is coated by apatite, and the encased tissue decomposes. The encasing phosphate remains intact as a thin sheet and retains the outer shape of the original tissue.

Source of the phosphorous

The calculations showed that skin, muscular tissue and blubber could provide sufficient P for the formation of a 0.8 mm thick layer of fluorapatite, i.e., sufficient for the observed preservation of mineralized epimorphs of soft tissue. The other main component of apatite is Ca, found in abundance throughout the host rock (Fig. 8). The skeleton is a possible phosphorous source. However, the apatite layer is not concentrated around the skeletal elements, but rather through the entire area covered by the specimen, indicating that the soft tissues are likely a major phosphorous source.

We conclude that an external phosphorous source is not necessary for this type of preservation to occur. In addition to muscles, we argue that the ichthyosaur possessed blubber, based on our observations of Type II soft tissue, in accordance with previous work (Lindgren et al., 2018). Thus, the phosphorus in the mineralized soft tissue is most likely derived from the most local source, the decaying organic matter of the ichthyosaur itself. A local source of phosphorus rather than external is further supported by the detailed investigation of the specimen and the element mapping (Fig. 8), which show phosphorus only in the area where the fossil is preserved. If the phosphorus source was external, it seems unlikely that apatite only formed where the fossil is preserved. However, as analyses of modern whale-falls have shown that there is an exchange of chemical components between decaying carcasses and the surrounding sediments, an external phosphorous source cannot be completely excluded. The observed preservation of the ichthyosaur is possible because of a short post-mortem floatation with rapid burial and/or overgrowth by microbial films.

Phosphatization of soft tissue in a carbonate-dominated environment, as in the Solnhofen area, has been hypothesized to be dependent on an anoxic environment with low pH (Briggs, 2003; Briggs & Wilby, 1996). In larger vertebrate carcasses, another possible scenario is that the skin creates a local closed system and an elevated concentration of phosphorous during decay. This creates a local fall in pH, shifting the equilibrium away from favouring precipitation of carbonate, to francolite (Briggs, 2003; Briggs & Wilby, 1996; Gueriau et al., 2020).

The source of fluorine in francolite can be explained by the bioapatite presence in the animal, as fluoride incorporation is prevalent although typically low ow with higher concentration close to veins and tissue (Li & Pasteris, 2014). Francolite is known to form by precipitation from anoxic porewater, replacement of carbonate, bacterial processes, post-mortem precipitation of francolite in phosphorus-rich bacterial cells; and possibly by precipitation from the water-column at the sediment water interface (McArthur, 1985).

The minor amounts of Na and S in fluorapatite can be explained by the coupled substitution Ca²⁺ + P⁵⁺ → Na⁺ + S⁶⁺, typical for apatite-group minerals, where Na and S originate from the sea water. In addition, iron was observed through SEM-EDS as well as during preparation, where iron oxide covered some of the elements (Radecker, 2014). This fits well with iron-rich bands observed in SEM-EDS, which are not present in the areas with apatite. The iron oxides and hydroxides are also most likely late weathering products, after the deposits became subaerial in the Cretaceous-Paleogene.