Osteological description

HMG-357 includes five vertebrae, eight ribs, a part of the left scapula and a clavicular arch. Judging from the anatomical positions and size of individual elements, it is safely assumed that this specimen originated from the pectoral region of a single individual. The neurocentral suture is completely closed in all vertebrae, indicating this individual was an “adult” or “old adult” sensu Brown (1981; also see “Discussion”). An Arabic numeral is painted on each element (but it does not match the anatomical order), and in the following description this number is employed to distinguish ribs, e.g. #15 for the rib with the painted number of “HMG 357-15”.

The articulated vertebrae (Figure 2A, B) represent the transition from the cervical to dorsal vertebrae, and the rib facets gradually shift from the centrum in cervicais to the tip of the transverse process in dorsals. The centra are amphicoelous (Figure 2C, D), and wider than tall and taller than long (Table 1), and there is no ventral notch on the ventral edge of the intervertebral facet. The foramina subcentralia are located in the shallow depressions on the ventral surface separated by a longitudinal ridge (Figure 2E). The neural spine slightly narrows dorsally in lateral view, and lacks the thickening at the anterior and posterior corners seen in the “pectoral” vertebrae of elasmosaurids (e.g. Sato et al., 2006).

All of the eight ribs (#7, 8, 10 to 15: Figure 3) are more or less damaged, but most of them can be matched with the corresponding vertebrae with reasonable confidence based on the fit of the head and rib facet, as well as the symmetry. We consider #11 as the anteromost right rib in the preserved series because of the shortness and the presence of the anterior process. #10 is slightly longer than #11, and likely represents the next right rib, because the partly damaged distal end retains the base of the anterior process. The articular facet is elongate and seems to match the facet on the second anteromost vertebra, although it may have been compressed slightly. #12 is longer than #10, and has an elongate articular facet which corresponds to the rib facets of the 3rd anterior vertebra; it is paired with #14 based on the characteristics of the constriction near the head, as well as the thickness of the shaft.

Table 1.

Vertebral measurements of HMG-357. Length (L), height (H), and width (W) of the centrum in mm. Subscripts “d” and “e” refer to distorted centrum and estimated length, respectively.

Figure 2.

Photograph (A) and interpretation (B) of the left lateral view of the five vertebrae of HMG-357 in articulation, and anterior (C), posterior (D) and ventral (E) views of the second vertebra. Hatching indicates damaged surface, and the open-circle-and-dot pattern marks matrix/adhesive in this and following figures. Abbreviations: rf, rib facet; tp, transverse process.

#15 and #7 are used for the histological study and described in detail. #7 is the best preserved rib in this specimen, although the distal end and the rim of the articular facet are slightly damaged. It is about 13 cm long from the rib facet to the broken distal end, and the longitudinal diameter of the damaged proximal end is about 25 mm. The shaft is gently curved, and the articulation facet is deeply concave to fit the convex facet on the transverse process, and there are ridges and gentle grooves on the surface near the head. #15 is heavily abraded and missing the rim of the head, the dorsal surface and the distal part, but paired with #7 based on the similarity of the ridge and groove near the ventral edge, as well as the concavity and inferred size of the articular facet. It is about 75 mm long, and the longitudinal diameter of the damaged proximal end is about 20 mm. #8 and #13 represent the posteromost pair belonging to the posteromost vertebra in the preserved series. The rib head is more massive than #7, the articular facet is deeply concave with a well denned rim, and the shaft is thicker.

Figure 3.

Photographs and interpretation of the ribs of HMG-357. A, B, dorsal (?) view of the 1st right rib (#11); C, D, dorso-posterior view of the 2nd right (?) rib (#10); E, F, posterior (?) view of the 3rd left rib (#12); G–L, proximal (G, H), anterior (I, J), posterior (K, L) views of the 4th left rib (#7); M–P, posterodorsal (M, N) and anterior (O, P) views of the 4th right rib (#15); Q, R, posterior view of the 5th left rib (#13).

Figure 4.

Photograph and interpretation of the clavicular arch of HMG-357 m anterior (A, D), ventral (B, E), and dorsal or visceral (C, F) views. Abbreviations: cl, clavicle; frag, fragmentary; icf, interclavicular foramen; icl, interclavicle; 1, left; r, right.

In plesiosaurs, the transition from the cervical to dorsal vertebrae is gradual and they are conventionally distinguished based on the location of the rib facet relative to the neurocentral suture (e.g. Welles, 1943; Carpenter, 1999). The neurocentral suture is completely closed in HMG-357 and cannot be traced in the vertebra bearing #7 and #15 ribs, but the rib facet on the short transverse process is convex without any trace of suture, suggesting it is a dorsal vertebra. In addition, the damaged shaft of #7 indicates that it was much longer than more anterior ribs. Therefore, we interpret #7 and #15 as the first dorsal ribs of this plesiosaur.

The preserved portion of the clavicular arch is missing much of the original edge of the bone (Figure 4). It is fairly thick in the anteromedial portion near the large interclavicular foramen, but the element thins posterolaterally, and it is only 1 mm thick or so at the broken edge of the interclavicle and clavicles. The three bones are fused, but the suture is partially traceable. There are longitudinal sharp ridges on the interclavicle lateral to the interclavicular foramen. This portion of the arch is slightly asymmetrical.

Judging from the thickness of the anterior edge and concavity on the medial side, the preserved dorsal blade came from the left scapula (Figure 5). The distal end of the blade is damaged but indicates that the tip of the blade widens and did not simply taper.

Microanatomical and histological descriptions

CT images of HMG-357 rib # 7.—The CT value (X-ray transmissivity) of the matrix and bone did not differ considerably, and the contrast of the image was enhanced in Figure 6. The comparison with the structures observed in thin sections supports our interpretation of these images. The brightness of the bone is reversed in some places within the same image, i.e., the trabeculae appear brighter than the matrix in the anterior half of the rib but darker in the posterior half, likely reflecting the subtle difference in the relative transmissivity between the matrix and bone. The bright spots in spongiosa likely represent pyrite crystals (see “microscopic description” below).

The inner rib structure exhibits osteoporotic-like condition at the two sectioned points (i.e., rib head and midshaft). The medullary cavity is surrounded by extensive spongiosa, and a thin layer of compact cortical bone forms the solid surface.

Microscopic description.—The longitudinal section of a rib head (HMG-357 #15-1, Figure 7A) displays loose spongiosa, consisting of thin trabeculae (about 0.2 mm) oriented longitudinally to the articular facet. The outermost cortex at the ventral side of the rib exhibits nearly avascular primary bone with few secondary osteons. There is a mass of tissue at the ventral comer of the articular facet, which appears homogeneous and contains numerous fiber bundles, possibly related with tendon insertions (Figure 7D; Haines and Mohuiddin, 1968; Petermann and Sander, 2013). Bands of alternating light and dark tissue are present in the fiber bundles showing possible tempo and mode of calcification of existing soft tissue (see also Main et al., 2005). In the medullary region, the trabecular tissue is mostly secondary remodeled.

Figure 5.

Photograph and interpretation of the dorsal blade of the right (?) scapula of HMG-357 in lateral (A, C) and medial (B, D) views.

Figure 6.

CT images of the proximal (A) and midshaft (B) portions of the left (?) anteromost dorsal rib (#7) of HMG-357.

In the shaft sections, there is a large open medullary cavity surrounded by spongiosa with numerous erosion lacunae throughout the preserved portion (Figure 7B, C). Most parts of the cortex are heavily remodeled; especially, the ventral region is completely remodeled by dense Haversian tissue (Figure 7E).

A thin layer of primary bone tissue consisting of parallel-fibered bone is present in the outermost cortex of the posterior side (Figure 7G). A part of the peripheral bone tissue comprises mostly avascular tissue including possible lines of arrested growth (LAG) (Figure 7I), which resembles an external fundamental system (EFS) (e.g. Sander et al., 2006). The outermost edge of the ventral side includes numerous fiber bundles running toward the periphery between secondary osteons (Figure 7D). There are numerous erosion cavities in the deeper cortex; they are irregularly shaped and larger in the deeper area. Howship's lacunae are occasionally present in the wall of the erosion cavity, indicating active resorption by osteoclasts (Figure 7F). The trabeculae are about 0.1–0.3 mm thick and form a coarse network and are composed of secondary lamellar bone tissue (Figure 7H).

The marrow spaces of HMG-357-# 15 are filled with minerals. Dark irregular spots are observed on the trabecular surfaces and form thick black layers in places (e.g. Figure 7F). They are opaque in transmitted light, but exhibit metallic luster in reflected light, likely representing pyrite crystals commonly occurring in fossilized bones from the Yezo Group (e.g. Kaim et al., 2008).

Justification of taxonomic identification

HMG-357 is identified as a polycotylid based on the morphology of the vertebrae and clavicular arch. Preserved bones of HMG-357 are comparable in size to the previously described Yezo polycotylid specimens which represent relatively small adult individuals (e.g. Sato and Storrs, 2000; Echizenya and Minoura, 2005).

The vertebrae of HMG-357 shows typical features of polycotylid “pectorals,” such as the short centrum with oval articular faces without a ventral notch, a pair of foramina subcentralia separated by a ridge, large zygapophyses, and the neural spine which narrows dorsally (e.g. Williston, 1903; Adams, 1997; Frey et al., 2017). In contrast, more common elasmosaurid “pectoral” vertebrae are characterized by the square neural spine in which the anterior and posterior comers are thickened and the intervertebral facet bears a ventral notch, and they lack the sharp ridge on the ventral surface of the centrum separating the paired depressions hosting the foramina subcentralia (e.g. Sato et al., 2006). Morphology of the clavicular arch is quite variable among the elasmosaurids (e.g. Welles, 1962), but an elongate interclavicular foramen between the sharp longitudinal ridges as in the HMG specimen has not been reported in any known elasmosaurid taxa. Much less information is available for the Cretaceous pliosauroids because of the paucity of the postcranial materials, but HMG-357 can be distinguished from the two genera with comparable specimens: Brachauchenius which lacks foramina subcentralia (e.g. Williston, 1903; Albright et al., 2007a), and Kronosaurus whose cervical and anterior dorsal centra are taller than wide (Romer and Lewis, 1959). The clavicular arch has been described only in a few pliosauroids (e.g. Peloneustes in Andrews, 1913), but we are not aware of any with an interclavicular foramen and surrounding ridge.

Figure 7.

Histological thin sections of the right (?) anteromost dorsal rib (#15) of HMG-357. A, longitudinal section of the proximal portion under natural transmitted light, with square D indicating the location of the close-up image in D of this figure; B, C, transverse sections (B is a few millimeters more distal than C) under natural transmitted light, with squares E to I indicating locations of close-up images in E to I of this figure; D–I, close-ups under polarized microscope showing fiber bundles and the bands of alternating light and dark tissue (D), Haversian tissue (E), Howship's lacunae (arrow heads, F), primary parallel-fibered bone tissue (arrow head, G), secondary lamellar bone tissue (H), and avascular bone tissue with possible lines of arrested growth (arrow head, I).

Characteristics of the interclavicular arch of HMG-357 distinguish it from non-polycotylids, but they are fairly common among polycotylids for which detailed comparison is available. HMG-357 shares the presence of a large interclavicular foramen with ridges on the sides with Dolichorhynchops osborni, D. bonneri, Eopolycotylus rankini, and Trinacromerum bentonianum (Williston, 1903, 1908; Adams, 1997; Albright et al., 2007b; O'Keefe, 2008); the taxa with the large foramen range from the Cenomanian to Santoni an or Campanian. In the Campanian—Maastrichtian D. herschelensis Sato, 2005, however, the area for the interclavicular foramen is mostly ossified, leaving very small openings and weak lateral ridges. The interclavicle is poorly preserved or unknown in many polycotylid taxa and it is hard to assess the effect of ontogenetic variability, but the presence of well defined foramen in the adult individual HMG-357 and other early polycotylids suggests that the closure of this foramen in D. herschelensis likely represents an autapomorphic condition.

Significance of the occurrence of Turonian polycotylids in Japan

Early polycotylid remains have been reported from the upper Lower Cretaceous of North America and Australia (Druckenmiller and Russell, 2009), although phylogenetic relationships within the group and referral of certain genera at higher taxonomic levels conflict in recent studies (e.g. Albright et al., 2007b; Druckenmiller and Russell, 2008; Ketchum and Benson, 2010; Sato et al., 2011). Diverse polycotylids are known from the Cenomanian to Campanian, mostly from the Western Interior Seaway (WIS) of North America but also from Asia and northern Africa, with the latest ones from the Campanian-Maastrichtian of North America and the Southern Hemisphere (summarized in Druckenmiller and Russell, 2009; Schmeisser McKean and Gillette, 2012).

The fossil record of Late Cretaceous polycotylids is geographically and stratigraphically sparse and uneven outside the WIS, and the polycotylid remains from the Japanese Cretaceous are important to form an understanding of the global distribution of the group. For example, the Cenomanian occurrence of an indeterminate polycotylid (Sato and Storrs, 2000) and the Turonian age of HMG-357 indicate that polycotylids crossed the Cenomanian-Turonian (C/T) boundary in the northwestern Pacific as they did in the Western Interior Seaway, so that their survival was not a local phenomenon restricted to the latter. No other Late Cretaceous polycotylid localities in the Pacific have a stratigraphic range across the boundary.

In the Yezo Group of Hokkaido and Sakhalin, C/T boundary events such as faunal turnover among shelled molluscs (ammonoids, inoceramids) and some microfossils, as well as an excursion of the carbon isotope ratio, are recognized, and their causal relationship with an oceanic anoxic event (OAE2) has been debated (e.g. Toshimitsu et al., 2003; Takashima et al., 2004; Yazykova et al., 2004; Takahashi, 2005; Kurihara et al., 2012). Meanwhile, some groups appear to have been less affected. There are no marked extinction events among radiolarians and dinoflagellate cysts (Yazykova et al., 2004; Kaneko and Hirano, 2005), and the coleoid cephalopods Longibelus matsumotoi and Conotheuthis hayakawai occur in Cenomanian and Turonian strata (Tanabe el al., 2017). Sato and Tanabe (1998) reported the occurrence of a Cenomanian polycotylid skeleton which contains many jaw apparati of ammonoids as its stomach contents. Most ammonoid species went extinct at the C/T boundary just to be replaced with other new species in the Turonian (Toshimitsu et al., 2003). Although the number and quality of Yezo plesiosaur specimens are not sufficient to analyze changes at lower taxonomic levels as in these invertebrates, they do demonstrate the survival of the clade across the C/T boundary, thanks to the continuous stratigraphic sequence with fine biostratigraphic resolution.

Ontogenetic stages and aquatic adaptation inferred from bone histology

Ontogeny affects morphology and various other aspects of an animal, and the importance of assessing the growth stage of studied specimen(s) cannot be overemphasized. However, growth stages are only roughly estimated for large fossil vertebrates, because their remains are often fragmentary and/or rare, and it is hard to document an ontogenetic sequence. Here we assess the growth stage of HMG-357 based on the osteological and histological features.

Unlike dinosaurs for which various criteria to distinguish ontogenetic stages have been proposed (summarized in Hone et al., 2016), there are very few previous studies for plesiosaurs. Brown (1981, p. 255) distinguished “juvenile”, “adult”, and “old adult” individuals of plesiosaurs based on osteological characters. An individual is “juvenile” if the neural arches and centra are not fused; it is “adult” if they are fused, and “old adult” if the individual exhibits additional characters indicative of advanced ossification. As Brown himself clearly indicated, these terms should be always used with quotation marks to indicate their special meanings which do not necessarily correspond to the biological age or maturity. Unfortunately this practice is not strictly followed, and it is sometimes unclear whether the distinction of juvenile and adult is based on the criterion set by Brown or not (e.g. Wiffen et al., 1995; Liebe and Hurum, 2012).

Brown's criterion has been widely used in the description of plesiosaurian specimens, probably because it is so practical; vertebrae are commonly found and easily identified as plesiosaurian. At the same time, it is quite approximate, in particular for the distinction of “adult” and “old adult” specimens. The ontogenetic sequence of ossification remains unknown for plesiosaurs, and it is hard to judge if a particular character of ossification appears earlier or later than other features. In the first place, the sequence of the closure of neurocentral sutures within the vertebral column (e.g. Brochu, 1996; Hone et al., 2016) has not been established to show if there is a consistent pattern within the Plesiosauria or subgroups, so even the distinction of “juvenile” and “adult” is challenged if the suture is partially closed within an individual. It is practically impossible to distinguish “adult” and “old adult” individuals for the fragmentary specimens in which additional characters indicative of advanced ossification cannot be confirmed.

In the case of HMG-357, the closure and complete obliteration of the neurocentral suture indicates it does not represent a “juvenile” sensu Brown. The interclavicle and clavicles are fused, leaving partially obliterated sutures, and this may be taken as evidence of advanced ossification at an “old adult” level. These bones are fused but sutures are clearly visible at least in one “juvenile” Trinacromerum specimen (the holotype of T. “willstoni” Riggs 1944; T. bentonianum in Carpenter, 1996), and they are at least partially visible in several polycotylid specimens in which neurocentral sutures are closed (e.g. Sato, 2005; Albright et al., 2007b).

The parallel-fibered avascular bone observed in HMG-357 (Figure 7G) can be interpreted as a sign of slow periosteal growth immediately before death. Woven-fibered tissue with plexiform or radial vascular canals have been observed in the juvenile and subadult individuals of various non-polycotylid plesiosaurs and “pliosaurid” (see Introduction) plesiosaurs (Wiffen et al., 1995; Fostowicz-Frelik and Gazdźicki, 2001; Liebe and Hurum, 2012), suggesting relatively rapid bone deposition (Huttenlocker et al., 2013). Assuming that rapid deposition of bone tissue also occurred in an early ontogenetic stage of polycotylid plesiosaurs, the peripheral bone tissue in HMG-357 may indicate slowed deposition at a late ontogenetic stage of the individual.

Given that the degree of remodeling is mostly related with skeletal maturity in tetrapods such as dinosaurs and terrestrial mammals (Klein and Sander, 2008; Mitchell et al., 2017), the extensive remodeling of the rib may suggest that HMG-357 reached skeletal maturity. In addition, although obscured by the extensive remodeling, the presence of the peripheral bone tissue which resembles an EFS implies an effective cessation of any significant periosteal growth of the bone in HMG-357 (e.g. Erickson, 2005; Woodward et al., 2011).

The extensive spongiosa and thin compacta in the studied HMG ribs represent an osteoporotic-like condition (Ricqlès and Buffrénil, 2001), which has been reported in highly aquatic amniotes including extinct taxa such as ichthyosaurs (Houssaye et al., 2014; Houssaye et al., 2016; Nakajima et al., 2014) and mosasaurs (Houssaye et al., 2013). Prominence of the erosion spaces including the Howship's lacunae suggests ongoing osteoclast activity and intense resorption activity. In extant taxa, active swimmers such as cetaceans and pinnipedians display similar microanatomical features (notably the spongy inner organization) requiring efficient swimming abilities (e.g. manoeuvrability, speed) and relying on hydrodynamic buoyancy and body trim control (Ricqlès and Buffrénil, 2001; Houssaye, 2009). Therefore, these results on HMG-357 suggest that this polycotylid plesiosaur had acquired highly aquatic adaptation in bone microstructure.

In summary, osteological features indicate that HMG-357 represents an “old adult” sensu Brown (1981), in which the periosteal deposition of bone tissue was probably slowed by maturity at least in this part of the skeleton. The microscopic features also exhibit an osteoporoticlike condition, suggesting that polycotylid plesiosaurs were highly aquatic animals.