Skeletal ossification of Middle Triassic pachypleurosaur Keichousaurus hui (Reptilia: Sauropterygia) revealed by zinc distribution

Tendency of overall elemental distribution

The elemental distribution in K. hui demonstrates a distinct enrichment of elements typically associated with bone, primarily calcium (Ca), phosphorus (P), and strontium (Sr) (Figs. 2–4, Figs. S1–S4). Conversely, manganese (Mn), sulfur (S), titanium (Ti), and potassium (K) exhibit higher concentration in the surrounding matrix and are relatively depleted in the skeletal structures (Fig. S5). Notably, Zn shows a strong correlation with skeletal elements in some specimens (Figs. 2E, 3E, 4E).

The skeletal structures exhibit significantly higher Ca content compared to the surrounding matrix, mineralogically consistent with apatite. However, the contrast in Ca-maps is not sufficiently pronounced to clearly delineate the skeletal outlines (e.g., Figs. 2B, 3B, 4B). P is primarily distributed in skeletons, thereby revealing the skeletal structures (e.g., Figs. 2C, 3C, 4C). Notably, all seven K. hui specimens consistently exhibit distinctly elevated Sr levels throughout the entire skeleton, providing exceptional anatomical clarity (e.g., Figs. 2D, 3D, 4D). The Sr-maps demonstrate a higher contrast compared to the Ca-maps and P-maps, suggesting their superior potential for distinguishing bones from the matrix.

The distribution of Zn in K. hui varies significantly among the seven specimens. In three specimens—XNGM WS-30-R43 (Fig. 2E), GMPKU-P-4316 (Fig. 3E), and XNGM WS-32-R18 (Fig. 4E)—Zn shows marked correlation with skeletal structures, but with distinct distribution patterns in each specimen. In contrast, the remaining four specimens show no elevated Zn levels in the skeletal regions (Figs. S1E, S2E, S3E, S4E). Given this variability, the Zn distribution in the three specimens displaying skeletal enrichment is described in greater detail below.

Elemental distribution in XNGM WS-30-R43, GMPKU-P-4316 and XNGM WS-32-R18

Elemental distribution in GMPKU-P-4316

GMPKU-P-4316 is a subadult individual exposed in ventral view (Fig. 3). The Zn-map shows that the distribution of Zn is closely related to the skeletal structure, with higher Zn content observed in specific bones such as the parietal, carpals, tarsals, proximal ends and shafts of the dorsal and caudal ribs. Notably, the proximal ends and shafts of dorsal and caudal ribs show distinct Zn enrichment compared to the rest of the ribs, forming a bilaterally symmetrical pattern along the dorsal and caudal vertebrae (Figs. 3E, 3F). Zn distribution around the tarsals is particularly noteworthy, showing a nearly circular pattern consistent with the bone shape (Figs. 5F, 5K, 6J, 6K). According to the complete outline of the tarsals in the light photo and Sr-map (since Sr can clearly delineate bone morphology, we use Sr-maps here to reveal the structure), high Ca and P concentrations are concentrated primarily in the centers of the astragalus and calcaneum (Figs. 5C, 5D, 5H, 5I, 6C, 6D, 6G, 6H). In contrast, Zn distribution is roughly complementary to Ca and P, extending slightly beyond the skeletal outline of the tarsals (Figs. 5L, 5M, 6L, 6M). The Zn-map further reveals the possible morphology of the fourth distal tarsal (dt4). In the left hindlimb, the dt4 is visible in the light photo, Sr-map, and Zn-map, appearing as a subround area between the astragalus and calcaneum (Figs. 5B, 5E, 5F), while being undetectable in the Ca-map or P-map (Figs. 5C, 5D). Notably, this subround Zn-enriched area is larger than those observed in the Sr-map and light photo (Figs. 5B, 5E, 5F). In the right hindlimb, the dt4 is only visible in the Zn-map (Fig. 6J, 6K) and not in other maps (Figs. 6C, 6D, 6E). Furthermore, the Zn content in the metatarsal close to the tarsal gradually diminishes from the proximal towards the distal end in both hindlimbs (Figs. 5F, 5K, 6J, 6K).

Preservation mechanisms of Ca, P, Sr, Zn within fossil bones

Changes in the elemental composition of fossil bones can occur during two distinct periods: first, during the organism’s lifetime, and second, during the fossilization process (Piga et al., 2011). Throughout fossilization, the preservation mechanisms of various elements differ due to a complex interplay of chemical, physical, and environmental factors (Parker & Toots, 1970; Pyzalla et al., 2008; Piga et al., 2011). Calcium hydroxyapatite (Ca₁₀(PO4)₆(OH)₂), the primary inorganic content of bones, has the flexibility of the lattice for replacements at every site within the mineral structure. This flexibility enables the replacement of Ca²⁺, PO₄³⁻, and OH⁻ ions with other ions, facilitating both the uptake and loss of elements within the bones (Miyaji, Kono & Suyama, 2005; Trueman et al., 2008).

Despite some inevitable loss during fossilization, Ca and P are highly tenacious and stable, remaining the predominant constituents in fossil bones (Pan & Fleet, 2002). For instance, in the bones of the Middle Triassic marine reptile Mixosaurus panxianensis, Ca accounts for 68.86 wt% to 71.11 wt%, and P ranges from 6.67 wt % to 8.05 wt % (Wang et al., 2024). A similar pattern is ubiquitous across diverse fossil taxa (Li et al., 2020; Liu et al., 2024), and this trend is also observed in seven specimens of K. hui analyzed in this study (e.g., Figs. 2B–2C, 3B–3C, 4B–4C). The enrichment of Sr in fossils occurs both during the organism’s lifetime and through the fossilization process. Sr concentrations are higher in seawater than that in freshwater, resulting in marine organisms typically exhibiting higher Sr content that is strongly correlated with the Sr levels in their environment (Rosenthal, Eves & Cochran, 1970). Sr incorporates into the bones during fossilization through substitution processes, leading to further enrichment, reflected by the elevated Sr content observed in the seven K.hui specimens (e.g.,Figs. 2D, 3D, 4D) (Gueriau, Jauvion & Mocuta, 2018).

However, the preservation of Zn in fossil bones presents significant challenges, as some degree of Zn loss is inevitable during the fossilization and preparation processes. Firstly, in living organisms, Zn is present in much lower concentrations and is unevenly distributed within bones compared to the abundant endogenous Ca and P (Honda et al., 1984; Goodwin et al., 2007; Wang et al., 2024). Secondly, diagenetic processes, particularly demineralization, are likely responsible for postmortem Zn loss (Dean et al., 2023). Micro-XRF scans of Mixosaurus panxianensis have revealed a pattern of Zn diffusion, with Zn concentrations gradually decreasing from the trunk region toward the surrounding matrix (Wang et al., 2024). Finally, the fossil preparation process can also alter the original distribution of Zn, despite efforts to minimize such impacts. Physical friction during matrix removal can damage the original Zn signal, and the use of solvents such as water or acetone during preparation can lead to Zn loss. Notably, Zn²⁺ exhibits high solubility in water, increasing the likelihood of its removal during preparation (Mou, 1999). As a result, among the seven K. hui specimens analyzed, only three retained a relatively intact original Zn signal (Figs. 2E; 3E; 4E). The remaining specimens failed to preserve this signal due to factors related to fossilization or preparation.

Origin of Zn in Keichousaurus hui

The robust correlation between Zn intensity and bone morphology in specimens XNGM WS-30-R43, GMPKU-P-4316, and XNGM WS-32-R18 suggests that the observed Zn enrichment in these fossilized bones is endogenous (Figs. 2E, 3E, 4E). This interpretation is supported by several lines of evidence. Firstly, unlike other bone-related elements such as Ca, P, and Sr, which are uniformly distributed throughout the skeletal structures, Zn shows selective accumulation in specific skeletal regions. For example, Zn is concentrated in areas like the periphery of tarsal bones in K. hui (Figs. 5 and 6). Zn exhibits an uneven distribution within the bones of extant and extinct species, predominantly localizing in regions between mineralized and unmineralized zones of osteons, as well as at the ossification front where active calcification occurs (Bradley, Moger & Winlove, 2007; Anné et al., 2014). For instance, synchrotron microfocus X-ray analysis of the forelimbs of one-day-old mice supports this pattern, demonstrating Zn enrichment across all skeletal elements, including both cartilage and bone (Fig. 8D). Peak concentrations are observed at the developing edges of bones, a distribution pattern strikingly similar to that observed in K. hui (Fig. 8D) (Anné et al., 2017). Secondly, Zn content in the surrounding pelitic limestone is markedly lower than that detected in the bones of the three specimens, effectively ruling out exogenous Zn. And there is no evidence to suggest that Zn could have been transferred by aqueous fluids in this system. Mineralogical experiments indicate that substituting exogenous Zn for Ca in hydroxyapatite is relatively challenging (Miyaji, Kono & Suyama, 2005). This contrasts with the physiological incorporation of Zn into bone, suggesting that Zn from groundwater is unlikely to enter bones through substitution processes. Finally, the three Zn-enriched specimens can be compared with the Mixosaurus panxianensis specimen that shows evident Zn loss (Wang et al., 2024), revealing different Zn distribution characteristics: K. hui displays discrete Zn zoning with sharp compositional boundaries (e.g., Fig. 4E), whereas M. panxianensis exhibits diffuse boundaries (Wang et al., 2024). This sharp contrast indicates that the absence of Zn in other skeleton regions is not likely to be attributed to post-depositional loss processes, given that the element loss tends to lead to the diffuse distribution as shown in M. panxianensis.

Hence, the Zn enrichment is more likely to be endogenic, i.e., the Zn was accumulated when these K. hui were still alive. The distribution of Zn reflects the original organism’s chemical remains rather than taphonomic artifacts added during fossilization, despite the preservation mechanism remains unclear.

The different ossification patterns in juvenile and subadult individuals

The distribution patterns of Zn, especially the enrichments in the XNGM WS-30-R43 and GMPKU-P-4316 (Figs. 2–3), demonstrate active ossification in K. hui. XNGM WS-30-R43, a juvenile specimen, exhibits higher Zn levels predominantly in the axial skeleton compared to the limbs, suggesting that the axial bones develop more rapidly or earlier than the limb bones at this ontogenetic stage (Fig. 2E). GMPKU-P-4316, a subadult individual, displays high Zn levels in scattered areas such as the carpals, tarsus, parietal, proximal ends of dorsal ribs, and caudal ribs (Fig. 3), suggesting active ossification in these regions during the subadult stage. Particularly, Zn enrichment in the parietal bone indicates that this region was actively ossifying, with the parietal opening gradually closing (Figs. 3E, 3F). This pattern aligns with observations in certain extant reptiles, where closure of the parietal fontanelle is markedly delayed (Maisano, 2001). Similarly, the high Zn content in the ends and shafts of dorsal ribs implies continued ossification in these regions (Figs. 3E, 3F). These findings align with previous descriptions of subadult K. hui by Lin, Rieppel & Field Museum of Natural History (1998) and Qin, Yu & Luo (2014), which highlight rapid ossification of sexually dimorphic limb elements, expansion of dorsal rib ends, and closure of the parietal opening. The observed Zn distribution in GMPKU-P-4316 corroborates these developmental patterns.

In subadult specimens, Zn enrichment is notably absent in the neck and certain cranial regions, with a sharp demarcation line separating these areas from others, suggesting an abrupt rather than gradual transition (Fig. 3E). Microscopic observations revealed that the specimen had been somewhat excessively prepared, potentially resulting in the loss of original Zn signals. Alternatively, Zn may not have been preserved in these regions during fossilization, as discussed earlier, due to the challenges associated with its preservation. Consequently, it cannot be definitively concluded that the neck of subadult individuals lacked Zn enrichment. It is plausible that the neck, like the tail, underwent active ossification, consistent with previous allometric growth analyses (Xue et al., 2015). Unfortunately, this detail is not reflected in the Zn distribution maps, despite meticulous fossil preparation efforts.

The Zn-maps of XNGM WS-30-R43 and GMPKU-P-4316 reveal distinct ontogenetic patterns of ossification, indicating rapid development in both juvenile and subadult individuals (Fig. 7A). In the juvenile period, ossification mainly occurred in the axial bones and the limbs, while the actively ossifying bones in subadult individual are in the limbs (Figs. 2E, 2F, 3E, 3F).

Ossification patterns of tarsus in Keichousaurus hui and comparison with extant species

The ossification pattern of the tarsus in K. hui, as indicated by the distribution of Zn, demonstrates directionality from the centre to the periphery (Fig. 7B). This pattern aligns closely with the highly conserved ossification sequence observed in extant reptiles, which exemplifies typical endochondral ossification (Rieppel, 1991; Maisano, 2001). In this pattern, the ossification center of the tarsal bones initially appears as relatively small regions within larger cartilage primordia and subsequently expands radially from the centre to the periphery of each tarsal element (Hemo, Gigi & Wientroub, 2019; Maisano, 2000; Maisano, 2002). In Cyrtodactylus pubisulcus (Gekkonidae), the ossification centers similarly originate within the cartilage primordia and undergo gradual expansion over time, and the tarsal ossification sequence begins with the astragalus, followed by the calcaneum, and then the dt4. In terms of size, the tarsal elements follow the order astragalus >calcaneum >dt4 (Rieppel, 1992) (Fig. 8F).

Similarly, in early postnatal laboratory mice, ossification centre is also in the cartilage primordia and gradually expanded (Patton & Kaufman, 1995) (Fig. 8G). Notably, in the astragalus and calcaneum of GMPKU-P-4316, Zn exhibits higher concentrations in the peripheries compared to the centres (Figs. 5–6), suggesting that the peripheries were actively undergoing ossification while the centres may have already completed ossification. Additionally, the concentrations of Ca and P in the centres of the astragalus and calcaneum exceed those in the peripheries, complementing the distribution pattern of Zn (Figs. 5L, 5M, 6L, 6M). This further supports the inference that the degree of calcification is more advanced in the centers than in the peripheries. A partial extension of Zn is observed beyond the supporting bony elements, distributed along the sub-round bone margin in hindlimb (Fig. 5N, 6N), which may represent biological residues from the bone development process. Furthermore, the right hindlimb of GMPKU-P-4316 displays a subcircular area in the Zn-map (Fig. 6J) between the astragalus and calcaneum, which is absent in the light photo, Ca-map, P-map, and Sr-map (Figs. 6B–6E). This subcircular region in the Zn map may correspond to the biological remains of the dt4, which appears not to have ossified yet. By comparing the elemental distributions between the centers and peripheries, we can gain deeper insights into the ossification process of the tarsus in K. hui. This analysis supports the conclusion that the ossification pattern of the tarsus in K. hui follows a gradual progression from the center to the periphery (Fig. 7B).

According to the element maps, the astragalus exhibits a larger relative area enriched in Ca and P compared to the calcaneum, indicating that the extent of ossification is more advanced in the astragalus than in the calcaneum (Figs. 5L, 5M, 5N, 6L, 6M, 6N). Morphological observations further reveal that the irregular shape of the calcaneum suggests incomplete ossification. Therefore, the sequence of tarsal ossification in K. hui can be inferred as astragalus, calcaneum, and dt4, which aligns with the ossification pattern observed in certain members of Gekkonidae (Rieppel, 1992).

The ossification of ribs in Keichousaurus hui

Pachyostosis, a non-pathological bone hypertrophy, is observed in a wide array of taxa adapted to marine environments, irrespective of their phylogenetic relationships (Houssaye, 2009). The increase in skeletal mass plays the functional role of ballast for buoyancy control and hydrostatic regulation of body trim (Ricqlès & de Buffrénil, 2001; Houssaye, 2009). The Middle Triassic pachypleurosaurs, which inhabited shallow epicontinental marine habitats and intraplatform basins, display pachyostosis in most elements of the presacral axial skeleton (Sues & Carroll, 1985; Rieppel & Kebang, 1995; O’Keefe, Rieppel & Sander, 1999). In K. hui, pachyostosis is particularly evident in the dorsal region, where the ribs and vertebrae are thickened (Young, 1958).

In the juvenile individual (XNGM WS-30-R43), the shafts of dorsal ribs exhibit elevated Zn content (Figs. 2E, 2F), suggesting active ossification in these regions. Similarly, the subadult individual (GMPKU-P-4294) shows Zn enrichment in the shafts of dorsal ribs, though the areas with high Zn content are less extensive compared to the juvenile (Figs. 3E, 3F). Notably, pachyostosis is not apparent in either the juvenile or subadult specimens (Figs. 2A–2D; 3A–3D). The adult individual exhibits significant Zn enrichment in the rib shafts, suggesting that ossification is still ongoing, while pachyostosis is clearly evident (Fig. 4E).

Previous morphological studies have indicated that pachyostosis is absent in K. hui embryos (Young, 1958). Our observations of juvenile and subadult individuals also show no clear signs of pachyostosis, but the pachyostosis is apparent in adult individual. In addition, the distribution of Zn highlights active ossification in the shafts of the dorsal ribs in juvenile, subadult and adult individuals. Integrating these morphological features with Zn distribution patterns, we infer that pachyostosis in K. hui develops progressively from embryonic stages to adulthood.

The advantage of Zn revealing the ossification information in fossils

Zn, preserved in situ over deep time, enables the identification of active ossification sites related to ontogenetic stages, particularly in non-adult extinct organisms. Moreover, the macroscopic enrichment of Zn can elucidate the ossification pattern of different ontogenetic individuals, providing more information about growth pattern.

Comparative anatomy and bone histology techniques have been used to reconstruct the ossification patterns in fossil marine reptiles (Klein & Sander, 2008; Scheyer, Klein & Sander, 2010). However, comparative anatomy as direct investigation is usually hampered by preservative conditions and the number of specimens (Horner & Weishampel, 1996). Paleohistology which can provide an independent assessment of the age or ontogenetic stage of extinct species based on developmental records in skeletal tissues, is a destructive method and unsuitable for rare fossil specimens (Klein & Sander, 2008). Hence, XRF mapping of Zn presents a novel, non-destructive method for addressing ossification, especially in precious specimens, offering additional insights into ossification patterns in fossils.