Non-technical Summary
In this study, we performed further microstructural studies of the eggshells of elongatoolithid eggs from China. We found that more diverse calcite crystal morphologies exist in these eggshells than were previously known. While excluding pathological structures, we also found evidence that some eggshells previously identified as elongatoolithid actually do not belong to this group. This study provides important data for the comparison of related fossil and extant eggshells.
Introduction
Electron backscatter diffraction (EBSD) is a powerful tool for in situ quantitative analysis of crystalline materials. An EBSD detector is attached to a scanning electron microscope (SEM) to obtain a Kikuchi pattern, allowing automatic analysis of grain orientations, local textures, point-to-point orientation correlations, and phase identification and distribution (Schwarzer et al. Reference Schwarzer, Field, Adams, Kumar, Schwartz, Schwartz, Kumar, Adams and Field2009). Geologists also apply this technique in different fields in the Earth sciences (Prior et al. Reference Prior, Mariani, Wheeler, Schwartz, Kumar, Adams and Field2009), including paleontology. This method has been applied by paleontologists in studies of biocrystalline materials like coral reefs (Cusack et al. Reference Cusack, England, Dalbeck, Tudhope, Fallick and Allison2008b; Cusack Reference Cusack2016) and mollusk shells (Cusack et al. Reference Cusack, Dauphin, Chung, Pérez-Huerta and Cuif2008a; Pérez-Huerta et al. Reference Pérez-Huerta, Dauphin, Cuif and Cusack2011). Sauropsid eggshells are typical biocrystalline materials (Mikhailov Reference Mikhailov2019) and therefore are well-suited for EBSD study. Dalbeck and Cusack (Reference Dalbeck and Cusack2006) first studied modern avian eggshells with EBSD to explore the relationship between the distribution of trace elements and the crystallographic structures of the eggshell, and this technique was soon applied to eggshells of non-avian dinosaurs (Grellet-Tinner et al. Reference Grellet-Tinner, Sim, Kim, Trimby, Higa, An, Oh, Kim and Kardjilov2011; Trimby and Grellet-Tinner Reference Trimby and Grellet-Tinner2011; Moreno-Azanza et al. Reference Moreno-Azanza, Mariani, Bauluz and Canudo2013, Reference Moreno-Azanza, Bauluz, Canudo, Gasca and Torcida Fernández-Baldor2016, Reference Moreno-Azanza, Bauluz, Canudo and Mateus2017; Eagle et al. Reference Eagle, Enriquez, Grellet-Tinner, Pérez-Huerta, Hu, Tütken and Montanari2015; Choi et al. Reference Choi, Han and Lee2019, Reference Choi, Barta, Moreno-Azanza, Kim, Shaw and Varricchio2022a; Choi and Lee Reference Choi and Lee2019; Kim et al. Reference Kim, Choi, Kim and Lee2019; Oser et al. Reference Oser, Chin, Sertich, Varricchio, Choi and Rifkin2021; Han et al. Reference Han, Yu, Zhang, Zeng, Wang, Cai and Wu2023), other birds (Grellet-Tinner et al. Reference Grellet-Tinner, Murelaga, Larrasoaña, Silveira, Olivares, Ortega, Trimby and Pascual2012, Reference Grellet-Tinner, Spooner and Worthy2016, Reference Grellet-Tinner, Lindsay and Thompson2017; Jain et al. Reference Jain, Bajpai, Kumar and Pruthi2016; Pérez-Huerta and Dauphin Reference Pérez-Huerta and Dauphin2016; Dauphin et al. Reference Dauphin, Luquet, Perez-Huerta and Salomé2018; Choi et al. Reference Choi, Hauber, Legendre, Kim, Lee and Varricchio2023), and other sauropsids (Choi et al. Reference Choi, Han, Kim and Lee2018, Reference Choi, Kim, Paik, Park, Jung and Xu2022b,Reference Choi, Kim, Kim, Kweon, Lee, Zhang and Varricchioc; Choi Reference Choi2020; Moreno-Azanza et al. Reference Moreno-Azanza, Díaz-Berenguer, Silva-Casal, Pérez-García, Badiola and Canudo2021; Xu et al. Reference Xu, Xie, Zhang, Choi, Kim, Gao, Jin, Jia and Gao2022; Wu et al. Reference Wu, Tseng, Tsao, Chiang, Tai, Hsieh, Yu and Juang2023). For fossil eggshells, EBSD can be used to examine the extent of diagenesis to exclude nonbiological structures (Eagle et al. Reference Eagle, Enriquez, Grellet-Tinner, Pérez-Huerta, Hu, Tütken and Montanari2015; Moreno-Azanza et al. Reference Moreno-Azanza, Bauluz, Canudo, Gasca and Torcida Fernández-Baldor2016). Shape and size of calcite grains can be directly displayed in Euler and grain boundary (GB) maps, allowing identification of structures that could be neglected with a traditional method like polarized light microscopy. The usefulness of EBSD examination has been shown in some ootaxa, and it also has potential in the parataxonomic study of eggshells.
Elongatoolithid eggs, represented by Elongatoolithus (Zhao Reference Zhao1975; Zhao et al. Reference Zhao, Wang and Zhang2015), include many ootaxa from Asia and North America. These eggs are characterized by their elongated shape and ornamented eggshell with two structural layers (mammillary layer and continuous layer) and radially arranged paired eggs in layered rings (Zhao Reference Zhao1975; Yang et al. Reference Yang, Wiemann, Xu, Cheng, Wu and Sander2019). Mikhailov (Reference Mikhailov1997b) refined the diagnosis of the eggshell of Elongatoolithidae, which has a ratite morphotype, angusticanaliculate pore system, linearituberculate ornamentation on the equatorial part of the egg, and dispersituberculate ornamentation on the poles. The continuous layer of elongatoolithid eggshell lacks an external ultrastructural zone; therefore, the range of the continuous layer also represents the range of its squamatic zone (Mikhailov Reference Mikhailov1997b).
Several problems exist in the study of parataxonomy of elongatoolithid eggs. An important question is whether this group is monophyletic. Taxonomically, three elongatoolithid oogenera (Elongatoolithus, Macroolithus, and Macroelongatoolithus) have been related with oviraptorosaurs based on strong evidence (Sato et al. Reference Sato, Cheng, Wu, Zelenitsky and Hsiao2005; Weishampel et al. Reference Weishampel, Fastovsky, Watabe, Varricchio, Jackson, Tsogtbaatar and Barsbold2008; Pu et al. Reference Pu, Zelenitsky, Lü, Currie, Carpenter, Xu and Koppelhus2017), while the oogenus Paraelongatoolithus is considered to be similar to the eggs of dromaeosaurs (Grellet-Tinner and Makovicky Reference Grellet-Tinner and Makovicky2006; Wang et al. Reference Wang, Wang, Zhao and Jiang2010a; Choi and Lee Reference Choi and Lee2019), which clearly makes the oofamily Elongatoolithidae polyphyletic. Some oogenera like Undulatoolithus (Wang et al. Reference Wang, Zhao, Wang, Li and Zou2013) and Heishanoolithus (Zhao and Zhao Reference Zhao and Zhao1999) were rarely studied or described, making comparison of other ootaxon with these oogenera impractical.
This study aims to acquire microstructural and crystallographic characteristics that were overlooked in former studies by applying polarizing light microscopy (PLM), SEM, and EBSD to seven representative elongatoolithid oogenera from China. The newly observed features provide more objective prerequisites to discuss the aforementioned parataxonomic problems.
Materials and Methods
Seven elongatoolithid oospecies were selected for study, including Elongatoolithus elongatus (IVPP V 734), Macroolithus yaotunensis (IVPP V 2781) (Zhao Reference Zhao1975), Heishanoolithus changii (IVPP V 11578) (Zhao and Zhao Reference Zhao and Zhao1999), Undulatoolithus pengi (PXM V 0016) (Wang et al. Reference Wang, Zhao, Wang, Li and Zou2013), Paraelongatoolithus reticulatus (IVPP V 16514) (Wang et al. Reference Wang, Wang, Zhao and Jiang2010a), Nanhsiungoolithus chuetienensis (IVPP V 2783) (Zhao Reference Zhao1975), and Macroelongatoolithus xixiaensis (TTM 15) (Wang et al. Reference Wang, Zhao, Wang, Jiang and Zhang2010b). All eggshell samples are housed in IVPP.
The only U. pengi clutch, PXM V 0016, contains five complete eggs and three broken eggs (Supplementary Fig. 13). Six eggs in the clutch were sampled and thin sectioned to study the microstructural variation among eggs in the clutch. The clutch is housed in PXM.
Some of the specimens were thin sectioned in the original studies, and the remaining resin blocks were sufficient for sample preparation for EBSD analysis. As for those without remaining resin blocks, new samples were taken from the specimens for both thin section preparation and EBSD analysis. The eggshells were thin sectioned down to 30 μm following standard procedures (Zhu et al. Reference Zhu, Fang, Wang, Deng, Liu, Wen, Wang and Wang2021) (Fig. 1), and the remaining resin blocks were then used for EBSD analysis. Thin sections were observed and photographed with a Zeiss Axio imager A2m polarized light microscope, and both normal and cross-polarized light were used (Fig. 1).
Figure 1. Thin sections of the eggshells under normal (left) and cross-polarized light (right). A, B, Elongatoolithus elongatus (IVPP V 734); C, D, Macroolithus yaotunensis (IVPP V 2781); E, F, Heishanoolithus changii (IVPP V 11578); G, H, Undulatoolithus pengi (PXM V 0016); I, J, Paraelongatoolithus reticulatus (IVPP V 16514); K, L, Nanhsiungoolithus chuetienensis (IVPP V 2783); M, N, Macroelongatoolithus xixiaensis (TTM 15).
Elongatoolithid eggs usually have linear extended ornamentation (ridge or nodes, especially in the equatorial region; Mikhailov Reference Mikhailov1997b); therefore, analyses were made on sections that are perpendicular to the extension direction of the ornamentation for comparison.
Resin blocks with radial sections exposed after thin sectioning were polished for EBSD analysis. The preparation process was similar to that followed by previous researchers (Moreno-Azanza et al. Reference Moreno-Azanza, Mariani, Bauluz and Canudo2013; Choi et al. Reference Choi, Han and Lee2019). The surface to be observed for each sample was ground with silicon carbide abrasive paper using an EXAKT 400CP variable speed grinder-polisher in order to obtain a flat observation surface; then it was polished with 1 μm (5 min) and 0.25 μm (5 min) diamond suspension to remove scratches before finally being polished with 0.04 μm (10 min) and 0.02 μm (10 min) colloidal silica on a polishing cloth to remove superficial damage. The polished sample was then coated with 4–5 nm carbon to make the sample conductive, which is also thin enough to ensure the quality of the Kikuchi diffraction pattern (Pérez-Huerta and Cusack Reference Pérez-Huerta and Cusack2009). The block was stuck to the sample stage by carbon tape, with its sample surface tilted by 70°. The EBSD analyses were performed with an Oxford Nordlys Nano EBSD detector attached to a field emission SEM (Nova NanoSEM 450, housed in IGGCAS), and the Kikuchi patterns were analyzed with AZtec software. The scanning settings include 20 kV accelerating voltage, 80 nA beam current, and 18 mm working distance. The scanning step size ranged from 2 to 6 μm.
GB maps were obtained under thresholds of 5°, 10°, and 20°, representing the lower limit of low-, medium-, and high-angled GB, respectively. The noise reduction procedure for GB maps included wild spike elimination and one step of zero solution correction (i.e., a pixel is surrounded by eight pixels, and if six of them have the same value, then the center pixel, which might have no data as “zero solution,” is set to that value). Euler and IPF-Y (red pixels represent calcite grains that have the c-axis pointed to the outer surface) maps were obtained with raw data.
The secondary calcite data were removed before the generation of misorientation histograms and pole figures. Misorientation histograms were acquired with the lower limit of 5° and a step width of 2.5°. Misorientation angle distributions of correlated and uncorrelated pairs were obtained and graphed under the same noise reduction procedure used in GB maps. Pole figures were present in both scattered dots and contours (10° half width and 3° cluster size).
SEM observations were first made on fresh fracture surfaces and then on acid-treated (5% acetic acid, 1 min) fracture surfaces. The observations were performed with a Zeiss EVO MA 25 SEM housed in IVPP. The scanning settings include 10 kV accelerating voltage, 5 pA beam current, and 7 to 10 mm working distance.
Anatomical terminology in this paper follows Mikhailov (Reference Mikhailov1997b).
Institutional Abbreviations
IVPP, Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing, China; IGGCAS, Institute of Geology and Geophysics, Chinese Academy of Sciences; PXM, Pingxiang Museum, Pingxiang, Jiangxi Province, China; TTM, Tiantai Museum, Tiantai, Zhejiang Province, China.
Results
Preservation State
IPF-Y maps were used to characterize the c-axis orientation of calcite grains. The IPF-Y map is useful in identifying diagenetic structures (Cusack et al. Reference Cusack, England, Dalbeck, Tudhope, Fallick and Allison2008b; Moreno-Azanza et al. Reference Moreno-Azanza, Mariani, Bauluz and Canudo2013, Reference Moreno-Azanza, Bauluz, Canudo, Gasca and Torcida Fernández-Baldor2016; Eagle et al. Reference Eagle, Enriquez, Grellet-Tinner, Pérez-Huerta, Hu, Tütken and Montanari2015; Choi and Lee Reference Choi and Lee2019; Choi et al. Reference Choi, Han and Lee2019). All examined specimens show a regularly arranged c-axis of calcite grains, which have more oblique or horizontal grains in the mammillary layer (shows more blue and green in IPF-Y maps) and more grains aligned toward the outer surface in the continuous layer (red in IPF-Y maps) (Fig. 2). The results indicate that the grain orientation of all seven specimens are basically original rather than having undergone severe diagenetic processes. The calcite crystals of all specimens are more dispersed in the lower part and are more aligned in the upper part (Supplementary Figs. 3, 7, 11, 17, 22, 26, 30), which are normal for the growth of most amniotic eggshells (Moreno-Azanza et al. Reference Moreno-Azanza, Mariani, Bauluz and Canudo2013, Reference Moreno-Azanza, Bauluz, Canudo and Mateus2017) except for gecko eggshells (Choi et al. Reference Choi, Han, Kim and Lee2018). The examination of crystallographic orientation adds to the usability of data, allowing reliable interpretation.
Figure 2. IPF-Y maps of the eggshells. A, Elongatoolithus elongatus (IVPP V 734); B, Macroolithus yaotunensis (IVPP V 2781); C, Heishanoolithus changii (IVPP V 11578); D, Undulatoolithus pengi (PXM V 0016); E, Paraelongatoolithus reticulatus (IVPP V 16514); F, Nanhsiungoolithus chuetienensis (IVPP V 2783); G, Macroelongatoolithus xixiaensis (TTM 15). Red pixels indicate c-axis vertical to the outer surface, while green and blue pixels indicate c-axis parallel to the outer surface. CL, continuous layer; ML, mammillary layer; SZ, squamatic zone; EZ, external zone; SC, secondary calcite.
Secondary calcite grains present in the eggshells of Elongatoolithus elongatus (Figs. 3A, 4A, Supplementary Fig. 1A) and Nanhsiungoolithus chuetienensis (Figs. 3F, 4F, Supplementary Fig. 24A) exhibit syntaxial overgrowth.
Figure 3. Euler maps of the eggshells. A, Elongatoolithus elongatus (IVPP V 734); B, Macroolithus yaotunensis (IVPP V 2781); C, Heishanoolithus changii (IVPP V 11578); D, Undulatoolithus pengi (PXM V 0016); E, Paraelongatoolithus reticulatus (IVPP V 16514); F, Nanhsiungoolithus chuetienensis (IVPP V 2783); G, Macroelongatoolithus xixiaensis (TTM 15). CL, continuous layer; ML, mammillary layer; SZ, squamatic zone; EZ, external zone; SC, secondary calcite.
Figure 4. Grain boundary (GB) maps of the eggshells. A, Elongatoolithus elongatus (IVPP V 734); B, Macroolithus yaotunensis (IVPP V 2781); C, Heishanoolithus changii (IVPP V 11578); D, Undulatoolithus pengi (PXM V 0016); E, Paraelongatoolithus reticulatus (IVPP V 16514); F, Nanhsiungoolithus chuetienensis (IVPP V 2783); G, Macroelongatoolithus xixiaensis (TTM 15). Green (5–10°), blue (10–20°), and purple (above 20°) lines stand for low-, medium-, and high-angled GBs, respectively. CL, continuous layer; ML, mammillary layer; SZ, squamatic zone; EZ, external zone; SC, secondary calcite.
Morphology and Arrangement of Calcite Grains
The combined use of IPF-Y, Euler, and GB maps allowed detailed characterization of the orientation, morphology and arrangement of the calcite crystals in the eggshells.
Though the bottom of the mammillary layer was not completely preserved, Elongatoolithus elongatus shows gradual transition between the mammillary and continuous layer under normal light (Supplementary Fig. 1A). The mammillae in the mammillary layer are closely arranged, and each is composed of radiating wedges (Supplementary Fig. 1C). In the continuous layer, the calcite crystallites inlay with each other and are separated by rugged GBs (Figs. 3A, 4A). In the scanned area, most of the GBs in the continuous layer were high-angled, although there are also areas that are rich in low-angled GBs. The squamatic grains show irregular extinction under cross-polarized light (Supplementary Fig. 1B,F), which reflects typical squamatic ultrastructure. The surface is moderately undulating (Fig. 1A) due to the presence of linear ridges (Young Reference Young1954: plate I, fig. 1; Zhao Reference Zhao1975; Zhao et al. Reference Zhao, Wang and Zhang2015: fig. 14).
Macroolithus yaotunensis has a distinct boundary between the mammillary and continuous layer under normal light (Supplementary Fig. 5C), which is more abrupt than that of Elongatoolithus and is the major diagnostic characteristic of Macroolithus (Zhao Reference Zhao1975). The boundary can be well defined by the abrupt change of grain orientation (Fig. 2B) or GBs (Fig. 4B). The mammillary layer of M. yaotunensis is composed of compactly arranged mammillae that consist of radially arranged slender calcite wedges. The continuous layer of M. yaotunensis, however, is not simply composed of calcite grains that have their c-axes aligned with their long axes. Three morphological types of calcite grains exist in the continuous layer of M. yaotunensis (Fig. 4B). Type A (Fig. 4B) consists of small scaly calcite crystallites that are most abundant in the continuous layer, the c-axes of which are nearly perpendicular to the outer surface (Fig. 2B), while the extending directions of the grains are aligned with the undulation of the accretion lines (Figs. 3B, 4B). The squamatic crystallites are separated by rugged high-angled GBs, while rugged low-angled GBs are also developed in each one, forming squamatic subgrains and indicating strong presence of squamatic ultrastructure. Type B (Fig. 4B) exists only below the center of ridges as several large prismatic crystallites that are gathered together. High-angled GBs that are not rugged separate these prisms, and low-angled GBs are nearly absent in the prisms. The prisms show columnar extinction (Supplementary Fig. 5F), and squamatic ultrastructure is also absent and replaced by parallel cleavages under an SEM (Supplementary Fig. 6C). It is also possible to identify this structure under plain light, as the accretion lines are more parallel (Supplementary Fig. 5E). The prismatic columns also cause the constriction of accretion lines beneath the ridges, where the accretion lines seems to be jacked up toward the top of an ornamentation (Supplementary Fig. 5E). However, it is possible that this structure may not show up in some thin sections or under some ridges because the position of the section can occasionally deviate from the center of ridges; therefore, a thin section with proper orientation and field of view width would decrease the probability of misidentification. Type C (Fig. 4B) consists of bunches of radially arranged slender calcite subgrains that usually originated near the top of the prisms of the second type and extend to the tops of ridges. These slender calcite subgrains are separated by linear low-angled GBs, and show sweeping extinction under cross-polarized light (Supplementary Fig. 5F). The undulation near the surface is strong, but can be weak in some places (Fig. 1C), because the sections do not always cut through the peak of a node (Young Reference Young1965: plates I–III; Zhao Reference Zhao1975; Zhao et al. Reference Zhao, Wang and Zhang2015: fig. 18).
The first layer of Heishanoolithus changii is the mammillary layer, and its structure is very similar to that of E. elongatus, whose upper border is also gradual (Supplementary Fig. 1A,C). The second layer is the continuous layer; however, unlike any other specimens in this study, the continuous layer of H. changii consists of two zones (Fig. 2C). The inner zone shows irregular extinction (Supplementary Fig. 9D) that is similar to the continuous layers of E. elongatus and M. yaotunensis, which show squamatic ultrastructure (Supplementary Fig. 10B). Therefore, this zone can be clearly identified as the squamatic zone. The GBs in the squamatic zone are mostly high-angled and very rugged (Fig. 4C). The outer zone consists of a layer of compactly arranged calcite prisms that are separated by relatively straight high-angled GBs (Fig. 4C). Low-angled GBs are nearly absent in the outer zone. The outer zone therefore shows columnar extinction (Supplementary Fig. 9F), which appears to be very similar to the typical external zone in many avian eggs (Mikhailov Reference Mikhailov1997a). The accretion lines are more prominent and isolated in this part (Supplementary Fig. 9E). The surface undulation is similar to Elongatoolithus (Fig. 1E), although its surface ornamentation is composed of dense (but still dispersed) nodes that are approximately oriented along the longitudinal axis (Zhao and Zhao Reference Zhao and Zhao1999: plate I, fig. 2).
In the sample of Undulatoolithus pengi (egg no. 6 in the clutch; Supplementary Fig. 13), an unexpected additional layer is clearly identified near the outer surface in the Euler map (Fig. 3D) and GB map (Fig. 4D). The boundary between the additional layer and the structure below is distinct: radiating GBs (both low- and high-angled) reoccur and extend to the outer surface of the eggshell (Fig. 4D). Most GBs in this layer are not continuous with those below. This additional structure can also be identified under cross-polarized light (Supplementary Fig. 14B), but could easily be overlooked when viewed under normal light, because it could be confused with the dense accretion lines, just as it was in the original literature (Wang et al. Reference Wang, Zhao, Wang, Li and Zou2013). However, other sampled eggs (egg nos. 1–4, 7) do not have this structure (Supplementary Fig. 19B,D,F,H,L). We interpret this additional layer as an abnormal structure, as it is similar to some pathological structures in other fossil eggshells (Jackson et al. Reference Jackson, Garrido, Schmitt, Chiappe, Dingus and Loope2004; Grellet-Tinner et al. Reference Grellet-Tinner, Corsetti and Buscalioni2010); therefore, the rest of the eggshell was used for comparison and is described here. The mammillary layer of U. pengi is similar to that of E. elongatus, which is composed of compactly arranged mammillae that are composed of radiating wedges, and has a gradual boundary with the continuous layer (Supplementary Fig. 14C). The rugged GB in its continuous layer is not as rich as that in E. elongatus (Fig. 4D), and sweeping extinction (Supplementary Fig. 14E) appears just below the ornamentation with radially arranged calcite grains (Supplementary Fig. 15D). The rugged high-angled GBs are more perpendicular to the eggshell surface when presented in the outer part of the eggshell. The surface undulation of U. pengi is very strong (Fig. 1G), even in those samples without an abnormal layer (Supplementary Fig. 19A,C,E,G,K), reflecting its strong linearly arranged ridges and nodes (Wang et al. Reference Wang, Zhao, Wang, Li and Zou2013: fig. 3). Paraelongatoolithus reticulatus has eggshell units that are less fused than those in other oospecies in this study, although they are not simply spherulites like those of the dinosauroid-spherulitic basic type. Each eggshell unit is composed of a fan-shaped mammilla with radially arranged slender calcite grains (Supplementary Fig. 20C) that are slimmer than all oogenera in this study except Nanhsiungoolithus, and large calcite wedges originate from the upper edge of mammillae and reach the outer surface (Fig. 4E). The mammilla of Paraelongatoolithus can be classified as acicular. There is a clear border between the mammillary and continuous layer (Supplementary Fig. 20A). The large wedges form the continuous layer and show columnar to slightly fan-shaped extinction (Supplementary Fig. 20B). Low-angled GBs are rich in the mammillae, but are nearly absent in the continuous layer. The wedges in the continuous layer are separated by rugged high-angled GBs that are extended vertically. The surface curvature is not fully complete due to possible erosion (Fig. 1I), but undulation is visible as a result of the reticulated ornamentation (Wang et al. Reference Wang, Wang, Zhao and Jiang2010a: fig. 2A).
The eggshell of N. chuetienensis has broad mammillae that are looser than those of Macroolithus (Supplementary Fig. 24C), and each mammilla is composed of radially arranged calcite grains (Supplementary Fig. 24C) that are similar to those of Paraelongatoolithus. The continuous layer shows columnar extinction under cross-polarized light (Supplementary Fig. 24B,F) and consists of large but slender prismatic calcite grains that are separated by high-angled GBs that are perpendicular to the surface. Low-angled GBs are rare in the continuous layer (Fig. 4F). The GBs are less rugged than those in E. elongatus and M. yaotunensis. The surface undulation of Nanhsiungoolithus is very weak (Fig. 1K), which is consistent with the description of its smooth or faint ridged surface (Young Reference Young1965: plates XII, XIIIA; Zhao Reference Zhao1975; Zhao et al. Reference Zhao, Wang and Zhang2015: fig. 20).
In Macroelongatoolithus xixiaensis, the mammillary layer is similar to that of M. yaotunensis. However, in this specimen, the upper border with the continuous layer is not very distinct compared with other specimens (Zelenitsky et al. Reference Zelenitsky, Carpenter and Currie2000) or even other eggshells from different positions of TTM 15 (Wang et al. Reference Wang, Zhao, Wang, Jiang and Zhang2010b: fig. 4A). This is due to the stacked wedges on top of normal mammillae and the disordered blocky structure located at the lower part of the continuous layer (Supplementary Fig. 28A,C), which were interpreted as pathological structures by former researchers (Wang et al. Reference Wang, Zhao, Wang, Jiang and Zhang2010b). In the disordered area, the accretion lines are less developed and the GB map shows blocky grains separated by rugged high-angled GBs with few low-angled GBs inside the block (Fig. 4G). The upper part of the continuous layer is normal and has more prismatic calcite grains near the ornamentation (Supplementary Fig. 28E,F). The prismatic grains are separated by high-angled GBs, which are often slightly rugged. Rugged low-angled GBs are rich within the continuous layer, indicating the presence of squamatic ultrastructure. The low-angled GBs became straightly extended and radially arranged near the outer surface or at the location of ridges, forming radially arranged slender subgrains (Supplementary Fig. 28F) similar to those in M. yaotunensis. This specimen was located near the pointed end of the egg, and the surface undulation of this eggshell is very strong but slightly irregular (Fig. 1M), which is caused by the dense nodes on the surface (Wang et al. Reference Wang, Zhao, Wang, Jiang and Zhang2010b: fig. 3B).
Misorientation Angle Distribution
In general, all specimens studied show low-angled dominant correlated misorientation (Fig. 5), which is similar to the results of former research on elongatoolithid eggs (Choi et al. Reference Choi, Han and Lee2019). However, the distribution of high-angled correlated misorientation varies among these oospecies. High-angled correlated misorientation is lowest in Macroolithus yaotunensis (Fig. 5B), with only 39.88% above 20°, while it is higher in Elongatoolithus elongatus (64.27%; Fig. 5A), Heishanoolithus changii (65.02%; Fig. 5C), Undulatoolithus pengi (68.32%; Fig. 5D), Paraelongatoolithus reticulatus (59.81%; Fig. 5E), Nanhsiungoolithus chuetienensis (65.02%; Fig. 5F), and Macroelongatoolithus xixiaensis (62.16%; Fig. 5G). The significantly low proportion of high-angled correlated misorientation of M. yaotunensis is most likely to be the result of rich low-angled GBs in its eggshell (Fig. 4B).
Figure 5. Misorientation histograms of the eggshells. A, Elongatoolithus elongatus (IVPP V 734); B, Macroolithus yaotunensis (IVPP V 2781); C, Heishanoolithus changii (IVPP V 11578); D, Undulatoolithus pengi (PXM V 0016); E, Paraelongatoolithus reticulatus (IVPP V 16514); F, Nanhsiungoolithus chuetienensis (IVPP V 2783); G, Macroelongatoolithus xixiaensis (TTM 15). Data from areas of matrix or secondary calcite were removed before analysis.
In the meantime, it is inappropriate to compare misorientation between different eggshells, because the misorientation distribution of different structural layers can vary significantly in an eggshell; therefore misorientation distribution must be considered before being used for character coding. Taking misorientation histograms of different structural layers in an eggshell into account, it is clearly shown that the low-angled misorientation distribution in E. elongatus (Supplementary Fig. 4), M. yaotunensis (Supplementary Fig. 8), and M. xixiaensis (Supplementary Fig. 31) is contributed to by the continuous layer, which is rich in low-angled GBs. In contrast, the mammillary layer contributed most of the low-angled misorientation in H. changii (Supplementary Fig. 12), P. reticulatus (Supplementary Fig. 23), and N. chuetienensis (Supplementary Fig. 27).
Discussion
Criteria for Two Squamatic Ultrastructure Types
Squamatic ultrastructure, characterized by a particular squamatic (scaly) pattern (Mikhailov Reference Mikhailov1997b), is regarded as an important structure that makes ornithoid-type eggshell distinct from any other type of eggshell (Mikhailov Reference Mikhailov2014), although it does not appear to be clearly present in all fossil ornithoid-type eggshells. This structure is often observed in the continuous layer of some avian and non-avian dinosaur eggshells; therefore, the continuous layer is also commonly referred to as the squamatic zone in terms of ultrastructural zone, if the squamatic ultrastructure is shown. Usually, the identification of squamatic ultrastructure requires a section under an SEM, but it can also be identified by the extinction pattern through cross-polarized light through observation of a thin section under a polarized microscope (Mikhailov Reference Mikhailov1991), as an area with strong squamatic ultrastructure would show irregular extinction rather than sweeping or columnar extinction. Choi et al. (Reference Choi, Han and Lee2019) offered an alternative way to visualize the distribution of squamatic ultrastructure by GB mapping through EBSD analysis, in which they suggested that the squamatic ultrastructure could be represented by the presence of rugged GBs. This definition is a good start to visually identify squamatic ultrastructure. Furthermore, on that basis, we would like to differentiate two types of squamatic ultrastructure with EBSD by the general morphology and proportion of low-angled GBs in those rugged GBs.
Type I squamatic ultrastructure here is defined as areas of rugged high-angled GBs (Fig. 6A). This is basically the same as the definition of Choi et al. (Reference Choi, Han and Lee2019), except for the absence of low-angled GBs. In extant Neognathae, the whole squamatic zone in the eggshell of domestic duck (Anas platyrhynchos domesticus), domestic chicken (Gallus gallus domesticus), Japanese tit (Parus minor), and Korean magpie (Pica sericea) (Choi et al. Reference Choi, Han and Lee2019) clearly shows this type. The squamatic zone in the eggshell of Australian brush turkey (Alectura lathami), malleefowl (Leipoa ocellata) (Grellet-Tinner et al. Reference Grellet-Tinner, Lindsay and Thompson2017), and American flamingo (Phoenicopterus ruber) (Grellet-Tinner et al. Reference Grellet-Tinner, Murelaga, Larrasoaña, Silveira, Olivares, Ortega, Trimby and Pascual2012) probably exhibits type I squamatic ultrastructure. A fossil neognath like a suggested Phoenicopteridae (Grellet-Tinner et al. Reference Grellet-Tinner, Murelaga, Larrasoaña, Silveira, Olivares, Ortega, Trimby and Pascual2012) also has a squamatic ultrastructure similar to that of P. ruber eggshell. In extant Palaeognathae, the whole squamatic zone of the eggshells of ostrich (Struthio camelus), emu (Dromaius novaehollandiae), cassowary (Casuarius casuarius), kiwi (Apteryx mantelli), tinamou (Eudromia elegans and Nothoprocta perdicaria), and the “thick moa eggshells” (Dinornis), as well as the inner part of the squamatic zone of the eggshells of rhea (Rhea sp.), elephant bird (Aepyornithidae), and the “thin and middle thickness moa eggshells” (Choi et al. Reference Choi, Hauber, Legendre, Kim, Lee and Varricchio2023) show type I squamatic ultrastructure. A fossil palaeognath species Lithornis vulturinus (Grellet-Tinner and Dyke Reference Grellet-Tinner and Dyke2005; Choi et al. Reference Choi, Hauber, Legendre, Kim, Lee and Varricchio2023) also shows type I squamatic ultrastructure in its eggshell, but note that there are possible diagenetically induced cleavages in its calcite grains (Choi et al. Reference Choi, Hauber, Legendre, Kim, Lee and Varricchio2023: fig. 10C), which might have destroyed low-angled GBs. Fossil non-avian dinosaur eggshells sometimes show this type of squamatic ultrastructure, such as Prismatoolithus levis (egg of Troodon) (Choi et al. Reference Choi, Han and Lee2019), Trigonoolithus amoae (Moreno-Azanza et al. Reference Moreno-Azanza, Mariani, Bauluz and Canudo2013, Reference Moreno-Azanza, Canudo and Gasca2014), Protoceratopsidovum sincerum, Protoceratopsidovum fluxuosum (Choi et al. Reference Choi, Barta, Moreno-Azanza, Kim, Shaw and Varricchio2022a) and the inner sublayer of Reticuloolithus acicularis (Choi and Lee Reference Choi and Lee2019). Gobioolithus minor shows type I squamatic ultrastructure in its squamatic zone (Choi et al. Reference Choi, Han and Lee2019). However, this type of squamatic ultrastructure is qualitative, and could easily be interpreted differently, because the ruggedness of high-angled GBs is its major diagnosis in the GB map, which is somehow subjective (but see an attempt to quantify ruggedness in Choi et al. [2019: fig. S3]). More work needs to be done to refine this criterion for this type.
Figure 6. Diagrammatic sketches of two types of squamatic ultrastructure. A, Type I, rugged grain boundaries (GBs) are mostly high-angled; B, type II, rugged GBs are both high-angled and low-angled.
Identification of type II squamatic ultrastructure is less subjective in the GB map. We define it as areas with both rugged low- and high-angled GBs, in which the low-angled GBs are far more developed, intertwine, and form multiple, small scaly subgrains within the large grains formed by high-angled GBs (Fig. 6B). Good examples of this type in recent avian eggshell include the outer part of the squamatic zone of rhea (Rhea sp.), elephant bird (Aepyornithidae) eggshells, and the “thin and middle thickness moa eggshells” (Choi et al. Reference Choi, Hauber, Legendre, Kim, Lee and Varricchio2023). Some fossil non-avian dinosaur eggshells also exhibit type II squamatic ultrastructure, including the whole continuous layer of Elongatoolithus sp. and most of the continuous layer of Elongatoolithus subtitectorius (Choi et al. Reference Choi, Barta, Moreno-Azanza, Kim, Shaw and Varricchio2022a) and Macroelongatoolithus xixiaensis (Choi et al. Reference Choi, Han and Lee2019). It is also the same for the outer sublayer of R. acicularis (Choi and Lee Reference Choi and Lee2019), as long as it is not diagenetically altered too much.
The continuous layer of elongatoolithid eggshell is usually known to have type II squamatic ultrastructure. Our results show that the existence and distribution of type II squamatic ultrastructure are not the same in all examined oospecies, and type I squamatic ultrastructure may also occur in some. Elongatoolithus elongatus has type II squamatic ultrastructure throughout its continuous layer, while Macroolithus yaotunensis and M. xixiaensis also have type II squamatic ultrastructure in most parts of the continuous layer other than those places with prismatic grains, and these oospecies are also the most widely known elongatoolithid oospecies. Type II squamatic ultrastructure is far less developed in the continuous layer of the other four oogenera, which is reflected by the decrease of rugged low-angled GBs (Fig. 4C–F) and the rise of high-angled misorientations in the continuous layer (Supplementary Figs. 12C, 18C, 23B, 27B). The rugged low-angled GBs are scattered in the continuous layer of these eggshells rather than forming a network, so we interpret the continuous layer as having only type I squamatic ultrastructure.
It is interesting that type II squamatic ultrastructure seems to be absent in all Neognathae eggshells in literature, while it is developed in a small number of Mesozoic non-avian theropod eggshells and present in certain parts of the squamatic zones in some palaeognath eggshells (Table 1). Typical elongatoolithid eggshells such as Elongatoolithus, Macroolithus, and Macroelongatoolithus tend to have type II squamatic ultrastructure in almost the whole squamatic zone.
Table 1. List of eggshells in the literature that can be used to identify types of squamatic ultrastructure
a SZ, squamatic zone.
Ultrastructural Variation among Elongatoolithid Eggs
Oogenus- and oospecies-level classification of elongatoolithid eggshells has been focused on quantitative characters such as eggshell thickness and the ratio of layers or some non-quantitative characters like ornamentation that can be categorized (Mikhailov Reference Mikhailov1997b; Tanaka et al. Reference Tanaka, Lv, Kobayashi, Zelenitsky, Xu, Jia, Qing and Tang2011). The microstructural variation in elongatoolithids has been relatively unexplored.
Although all could be classified as calcite radial ultrastructure (Mikhailov Reference Mikhailov1997b), the mammillary structures of the oospecies studied in this work are not the same. The mammillae of E. elongatus, M. yaotunensis, Heishanoolithus changii, Undulatoolithus pengi, and M. xixiaensis (Fig. 7A–D,G) are compactly arranged within the mammillary layer and can hardly be separated from each other, while each mammilla is composed of radiating wedges. On the contrary, the mammillae of P. reticulatus and Nanhsiungoolithus chuetienensis (Fig. 7E,F) are more loosely arranged and the calcite wedges that constitute a mammilla are more slender than those in mammillae of former oospecies, which in morphology are similar to the acicular structures described by other researchers (Grellet-Tinner and Makovicky Reference Grellet-Tinner and Makovicky2006; Choi and Lee Reference Choi and Lee2019; Choi et al. Reference Choi, Han and Lee2019). The difference in mammillary morphology has already been mentioned by former researchers and could also be used to describe the mammillary morphology of some other theropod eggshells (Tanaka et al. Reference Tanaka, Zelenitsky, Saegusa, Ikeda, DeBuhr and Therrien2016). The slender calcite wedges in the mammilla of N. chuetienensis and P. reticulatus are not constant in width (e.g., in P. reticulatus, the width of the wedges in the mammillae ranges from 5–8 μm near the inner surface to about 25 μm near the continuous layer; Fig. 4E). The difference between acicular and wedged mammillae might simply result from different areal density of organic cores on the shell membrane.
Figure 7. Mammillary structures of the eggshells. A, Elongatoolithus elongatus (IVPP V 734); B, Macroolithus yaotunensis (IVPP V 2781); C, Heishanoolithus changii (IVPP V 11578); D, Undulatoolithus pengi (PXM V 0016); E, Paraelongatoolithus reticulatus (IVPP V 16514); F, Nanhsiungoolithus chuetienensis (IVPP V 2783); G, Macroelongatoolithus xixiaensis (TTM 15). CL, continuous layer; ML, mammillary layer.
So far, we have observed that, aside from scaly calcite subgrains consisting of type II squamatic ultrastructure, there are at least two additional morphological types of calcite grains in the continuous layer of elongatoolithid eggshells. One type is usually columnar, prismatic, or wedge-shaped and can exist in multiple places in the continuous layer. It could appear as several columns at the central base of the ridges of M. yaotunensis (type B; Fig. 4B) and columns or wedges in the continuous layer of P. reticulatus (Fig. 4E) and N. chuetienensis (Fig. 4F). These grains are relatively large and separated from each other by high-angled GBs and lack low-angled GBs within them. The other type is characterized by radially arranged slender calcite subgrains that are separated by straight, extending low-angled GBs, which are similar to the radiating subgrains in eggshells possessing spherulitic eggshell units (Grellet-Tinner et al. Reference Grellet-Tinner, Sim, Kim, Trimby, Higa, An, Oh, Kim and Kardjilov2011; Trimby and Grellet-Tinner Reference Trimby and Grellet-Tinner2011; Moreno-Azanza et al. Reference Moreno-Azanza, Mariani, Bauluz and Canudo2013, Reference Moreno-Azanza, Bauluz, Canudo, Gasca and Torcida Fernández-Baldor2016, Reference Moreno-Azanza, Bauluz, Canudo and Mateus2017; Eagle et al. Reference Eagle, Enriquez, Grellet-Tinner, Pérez-Huerta, Hu, Tütken and Montanari2015), therefore lacking squamatic ultrastructure. This type often appears right beneath the ornamentations, and it seems possible that the oospecies with relative prominent ridges tend to show this type, as in M. yaotunensis, U. pengi (Fig. 4B,D), and M. xixiaensis (Choi et al. Reference Choi, Han and Lee2019: fig. 4B), which explains the decrease of squamatic ultrastructure in calcite grains that constitute the ornamentation (Choi et al. Reference Choi, Han and Lee2019).
In M. yaotuneinsis, all three morphological types of calcite grains can be observed within a proper radial thin section that crosses the middle part of a prominent ridge (Supplementary Fig. 5). On the other hand, it could be possible that only scaly subgrains indicating type II squamatic ultrastructure can be observed because the thin sections cross the region between the ridges or eggshells with low ridges (usually those from the very end of the polar region). It is also possible that the calcite grains of the continuous layer might exhibit a different way in another part of the egg or in a section with different angle crossing the eggshell, neither of which occurs in this sample; this would require thorough study in the future.
Historically, the microstructure of M. yaotunensis was not highly appreciated; the variation of calcite grains in its continuous layer was unknown because of the lack of EBSD study. However, structures reflecting different calcite grain morphology in Macroolithus can be seen in some old work. For example, the columns beneath ornamentations can be seen as structures that jack up accretion lines in many published photographs of M. yaotunensis (Cheng et al. Reference Cheng, Ji, Wu and Shan2008; Wang et al. Reference Wang, Zhang, Sullivan and Xu2016). Some early published sketches of elongatoolithid eggshells also reflected this structure by illustrating the presence of jacked-up accretion lines near the ridges (Sochava Reference Sochava1969: fig. 2A; Sabath Reference Sabath1991; Mikhailov Reference Mikhailov1994: figs. 7.2I, 7.13C), but provided little description. A continuous layer with similar structure can also be seen in another oospecies of Macroolithus, M. rugustus, which is diagnostically different from M. yaotunensis mainly in the non-undulating boundary between the mammillary and continuous layers (Zhao Reference Zhao1975). The third oospecies of Macroolithus, M. mutabilis, is known only from pictures of fragments (Mikhailov Reference Mikhailov1994); therefore, we cannot be sure that the continuous layer of its eggshell is also similar. Should all Macroolithus oospecies have the same combination of calcite morphology in the continuous layer, this would be a good auxiliary diagnostic characteristic for this oogenus, as the major microstructural diagnosis of Macroolithus, the distinct boundary between the mammillary and continuous layers, can be easily modified by diagenesis (Wang et al. Reference Wang, Zhang, Sullivan and Xu2016: fig. 4B; Bi et al. Reference Bi, Amiot, Peyre de Fabrègues, Pittman, Lamanna, Yu and Yu2021: fig. 2B), but the combination of accretion lines jacked up by calcite columns below the ridges and the radiating calcite subgrains that show sweeping extinction could be more preservable and recognizable in the same specimen (Wang et al. Reference Wang, Zhang, Sullivan and Xu2016: fig. 4B).
Same combination of the three grain types can also be observed in Macroelongatoolithus. Choi et al. (Reference Choi, Han and Lee2019) reported that the squamatic ultrastructure decreases in the calcite grains that constitute the ornamentation of Macroelongatoolithus, which in our view is similar to what happens in Macroolithus, as those grains became more prismatic under ornamentation and those radially arranged subgrains near the outer surface (Choi et al. Reference Choi, Han and Lee2019: fig. 4B) are remarkably similar to the second and third types of calcite grains in Macroolithus in shape, GB characters, and distributions. Jin et al. (Reference Jin, Azuma, Jackson and Varricchio2007) observed the prismatic columns in the continuous layer of Macroelongatoolithus, which it is primarily or completely concealed by squamatic ultrastructure, and described this as “cryptoprismatic.” We find these cryptoprismatic columns similar to the type B calcite grains in Macroolithus, whose grain boundaries also seems linear (Jin et al. Reference Jin, Azuma, Jackson and Varricchio2007: fig. 2E).
Suggested Parataxonomic Revision of Paraelongatoolithus and Nanhsiungoolithus
Zelenitsky and Sloboda (Reference Zelenitsky, Sloboda, Currie and Koppelhus2005) erected Reticuloolithus hirschi based on eggshell fragments from southern Alberta and Montana, which have reticulated ornamentation and two-layered microstructure with acicular mammillae. The second oospecies of Reticuloolithus, R. acicularis, was named by Choi and Lee (Reference Choi and Lee2019) based on eggshell having a similar structure from the Upper Cretaceous of Wi Island, South Korea. These two ootaxa have been proposed to be produced by dromaeosaurs, because a specimen preserving a partial egg in contact with articulated gastralia of Deinonychus antirrhopus was described by Grellet-Tinner and Makovicky (Reference Grellet-Tinner and Makovicky2006), and its eggshell has macro- and microstructures similar to those of R. hirschi. In the meantime, the eggshell related to D. antirrhopus was assigned to P. reticulatus by Wang et al. (Reference Wang, Wang, Zhao and Jiang2010a), who erected this oospecies based on a partial egg that was excavated from the Upper Cretaceous Chichengshan Formation of the Tiantai Basin, Zhejiang Province, China. Choi and Lee (Reference Choi and Lee2019) suggested that Paraelongatoolithus could be a synonym of Reticuloolithus if the mammilla of Paraelongatoolithus is also acicular. Although our EBSD (Fig. 4E) and SEM (Fig. 7E) results show one well-preserved mammilla that has the same acicular structure as Reticuloolithus in P. reticulatus, which would support the close relationship between Paraelongatoolithus and Reticuloolithus, the structure in the continuous layer of P. reticulatus is different from what is described in R. acicularis and the eggshell related to Deinonychus. The Deinonychus eggshell was described as having slightly different structure between the lower and the upper part of its continuous layer (Grellet-Tinner and Makovicky Reference Grellet-Tinner and Makovicky2006). Choi and Lee (Reference Choi and Lee2019) also identified two sublayers in the continuous layer of the R. acicularis eggshell and suggested the possible homology between the R. acicularis and Deinonychus eggshells. They also claimed that squamatic ultrastructure exists in the continuous layer of R. acicularis, which we classified as type I squamatic ultrastructure (see the first section in the “Discussion”). On the other hand, the basically non-diagenetic continuous layer of P. reticulatus eggshell is composed of a single layer of wedges that are separated by nearly vertically extended high-angled GBs with a few low-angled ones. This may also represent the presence of type I squamatic ultrastructure and is similar to the inner sublayer of R. acicularis, although there is no trace of an outer sublayer that has type II squamatic ultrastructure. One explanation is that a possible outer sublayer in the continuous layer of P. reticulatus specimen is eroded and not preserved, which is supported by its reduced thickness and discontinuity of accretion lines near the outer surface in the radial section (Supplementary Fig. 20A). A specimen identified as P. sincerum (PIN 3143/121; Choi et al. Reference Choi, Barta, Moreno-Azanza, Kim, Shaw and Varricchio2022a: figs. 3, 9) is very similar to Paraelongatoolithus in its acicular mammillae (although they seem to be more compact than those in either Paraelongatoolithus or Reticuloolithus), abrupt boundary between structural layers, and columnar continuous layer, although its outer surface is severely eroded and hinders observation of its ornamentation. Therefore, before future synonymization of these oogenera, the existence of an outer sublayer would have to be solved first by examination on more well-preserved specimens. Currently, based on the reticulated ornamentation, a continuous layer that shows columnar extinction, and acicular mammillae, this study suggests this oogenera ought to be placed outside the oofamily Elongatoolithidae, as these characters differ from those of the typical elongatoolithid oogenera like Elongatoolithus and are not consistent with the diagnosis of Elongatoolithidae.
Similar to Paraelongatoolithus, Nanhsiungoolithus has a continuous layer that shows columnar extinction (Supplementary Fig. 24B) and acicular mammillae (Supplementary Fig. 24C), but its surface ornamentation is not very prominent. Zhao (Reference Zhao1975) described its outer surface as smooth or having faint ridges, which is also our observation. The oofamily Montanoolithidae shares columnar extinction with Nanhsiungoolithus, but has reticulated or anastomosing ornamentation (Zelenitsky and Therrien Reference Zelenitsky and Therrien2008: fig. 3) and a much more compact mammillary layer (Zelenitsky and Therrien Reference Zelenitsky and Therrien2008: fig. 4A; Vila et al. Reference Vila, Sellés and Beetschen2017: fig. 2B) similar to that in Macroolithus. If a better preserved specimen of Nanhsiungoolithus shows reticulated or anastomosing ornamentation, it would be a good reason to move this oogenus out of Elongatoolithidae, just like Paraelongatoolithus.
Abnormal Eggshell Structure in Undulatoolithus
Biological abnormality in the microstructure of eggshells is not rare in eggs of both recent and fossil species. Female birds that ingest synthetic pesticides through the food chain could produce significantly thinner or even soft eggshells (Hickey and Anderson Reference Hickey and Anderson1968; Oestreicher et al. Reference Oestreicher, Shuman and Wurster1971; Lincer Reference Lincer1975; Pruett-Jones et al. Reference Pruett-Jones, White and Emison1981; Lundholm Reference Lundholm1997; Holm et al. Reference Holm, Blomqvist, Brandt, Brunström, Ridderstråle and Berg2006; Bouwman et al. Reference Bouwman, Yohannes, Nakayama, Motohira, Ishizuka, Humphries, van der Schyff, du Preez, Dinkelmann and Ikenaka2019). Delayed oviposition derived by environmental stresses can cause the formation of multilayered eggshells in birds, non-avian dinosaurs, and turtles (Romanoff and Romanoff Reference Romanoff and Romanoff1949; Ewert et al. Reference Ewert, Firth and Nelson1984; Jackson and Varricchio Reference Jackson and Varricchio2003; Jackson et al. Reference Jackson, Garrido, Schmitt, Chiappe, Dingus and Loope2004; Sellés et al. Reference Sellés, Vila and Galobart2017; Bailleul et al. Reference Bailleul, O'Connor, Zhang, Li, Wang, Lamanna, Zhu and Zhou2019). Other disorders of the avian female reproductive system related to abnormal eggshell and rarely seen microstructural abnormalities in non-avian dinosaur eggshells have also been reported, most of which are related to aging or infection of the female birds (Romanoff and Romanoff Reference Romanoff and Romanoff1949; Keymer Reference Keymer1980; Grellet-Tinner et al. Reference Grellet-Tinner, Corsetti and Buscalioni2010).
In the eggshell of U. pengi, this study identified an abnormal layer additional to the outside of the normal eggshell. At some places, this layer seems to grow from a series of nucleation centers (Supplementary Fig. 14F), while at most places, there are no nucleation centers, and the grains maintain the orientation of the calcite grains beneath this layer (Fig. 3D, Supplementary Fig. 14E). The two conditions of the additional layer of U. pengi are similar to type I and II abnormality (sensu Jackson et al. Reference Jackson, Garrido, Schmitt, Chiappe, Dingus and Loope2004) reported in titanosaur eggshells, respectively (Jackson et al. Reference Jackson, Garrido, Schmitt, Chiappe, Dingus and Loope2004; Jackson and Schmitt Reference Jackson and Schmitt2008). Delayed oviposition as a stress response to external stimuli would be the best explanation for this abnormal structure.
Undulatoolithus pengi is known from the only clutch (PXMV-0016) consisting of four pairs of eggs (Wang et al. Reference Wang, Zhao, Wang, Li and Zou2013). While the abnormal layer can be clearly seen in multiple samples from egg no. 6 from the clutch, it is absent in the other five eggs sampled (Supplementary Fig. 19), suggesting the abnormal nature of this layer. Further research may be carried out on its significance for dinosaur breeding behavior.
Identification of an External Zone in Heishanoolithus
The external zone is commonly seen in the eggshells of modern birds and is regarded as an ultrastructural zone in the outermost part of the continuous layer, which under chemical treatment would exhibit clear vertical boundaries of columnar or prismatic calcite grains (Mikhailov Reference Mikhailov1997b). Tyler (Reference Tyler1965) cited this structure as the surface crystal layer and treated it as an independent third layer outside the continuous layer. The external zone has been reported in troodontid eggshells (Prismatoolithidae) like P. levis (eggs of Troodon formosus; Varricchio and Jackson Reference Varricchio and Jackson2004), as the GBs in its continuous layer become more linear near the outer surface (Choi et al. Reference Choi, Han and Lee2019). Moreno-Azanza et al. (Reference Moreno-Azanza, Canudo and Gasca2014) reported an external layer in the eggshell of T. amoae (Prismatoolithidae), which appears as a layer of coarse prismatic crystal. However, the prisms are nearly euhedral or subhedral based on their triangular “nodes” on the outer surface (Canudo et al. Reference Canudo, Gasca, Aurell, Badiola, Blain, Cruzado-Caballero and Gomez-Fernández2010: fig. 6-4; Moreno-Azanza et al. Reference Moreno-Azanza, Canudo and Gasca2014: fig. 2A), suggesting that it might be formed by recrystallization. Arriagadoolithus patagoniensis (eggs preserved with an alvarezsaurid theropod) is also reported to have an external layer (Agnolin et al. Reference Agnolin, Powell, Novas and Kundrát2012), although we believe the silicon-rich secondary matrix between its “external layer” and the prismatic layer (Agnolin et al. Reference Agnolin, Powell, Novas and Kundrát2012: fig. 10B,G) suggest that this external layer is probably a layer of secondary calcite. The eggshell of Triprismatoolithus stephensi is reported to have an external layer (Jackson and Varricchio Reference Jackson and Varricchio2010).
In elongatoolithid eggs, the existence of an external zone has never been confirmed. Our EBSD result shows the eggshell of H. changii exhibits a continuous layer that has a typical squamatic ultrastructure in its inner part as the squamatic zone and an overlying zone of prismatic calcite grains (Fig. 4C). The prisms in this zone are separated by vertical GBs that extend linearly, which in morphology suits the definition of an external zone. It is important that the squamatic zone and this zone can even be easily differentiated through a radial thin section with cross-polarized light by the contrast of irregular extinction below and prismatic extinction near the surface (Supplementary Fig. 9B). The accretion lines are more prominent in the external zone of H. changii, which testifies to the biogenic origin of this structure. Parataxonomically, the presence of an external zone in H. changii does not suit the diagnosis of Elongatoolithidae, which lacks an external zone (Zhao Reference Zhao1975; Mikhailov Reference Mikhailov1997b). We hereby suggest removing Heishanoolithus from the oofamily Elongatoolithidae.
A North American oospecies, Continoolithus canadensis (Zelenitsky et al. Reference Zelenitsky, Hills and Currie1996) caught our attention when we noted that the same combination of the irregular extinction in its inner part and the prismatic extinction near the outer surface is shown in a thin-section micro-photograph showing its extinction pattern (Jackson and Varricchio Reference Jackson and Varricchio2010: fig. 8). The accretion lines also became more prominent near the outer one-fifth of the eggshell. Therefore, we find it possible that the eggshell of C. canadensis, which was described as having two structural layers, also has an external zone that is similar to that of H. changii. Also, the thickness, layer ratio, and mammillary shapes of the two oospecies are similar, and the ornamentation types are both dispersituberculate, which suggests a close relationship between them.
Taxonomic Affinity of Elongatoolithid Eggs
The relationship between elongatoolithid eggs and oviraptorosaurs have been widely known due to the discovery of skeleton–egg associations (Norell et al. Reference Norell, Clark, Chiappe and Dashzeveg1995; Dong and Currie Reference Dong and Currie1996; Clark et al. Reference Clark, Norell and Chiappe1999; Sato et al. Reference Sato, Cheng, Wu, Zelenitsky and Hsiao2005; Fanti et al. Reference Fanti, Currie and Badamgarav2012; Pu et al. Reference Pu, Zelenitsky, Lü, Currie, Carpenter, Xu and Koppelhus2017; Jin et al. Reference Jin, Varricchio, Poust and He2020; Bi et al. Reference Bi, Amiot, Peyre de Fabrègues, Pittman, Lamanna, Yu and Yu2021) and eggs containing embryos (Norell et al. Reference Norell, Clark, Demberelyin, Rhinchen, Chiappe, Davidson, McKenna, Altangerel and Novacek1994; Cheng et al. Reference Cheng, Ji, Wu and Shan2008; Weishampel et al. Reference Weishampel, Fastovsky, Watabe, Varricchio, Jackson, Tsogtbaatar and Barsbold2008; Shao et al. Reference Shao, Fan, Jia, Tanaka and Lü2014; Wang et al. Reference Wang, Zhang, Sullivan and Xu2016; Xing et al. Reference Xing, Niu, Ma, Zelenitsky, Yang and Brusatte2022). These cases involve only three oogenera, Elongatoolithus, Macroolithus, and Macroelongatoolithus. However, this study shows Heishanoolithus and Paraelongatoolithus should no longer be regarded as members of Elongatoolithidae, and Nanhsiungoolithus may also be a oogenus outside Elongatoolithidae, and these ootaxa are likely to be produced by other theropod dinosaurs such as dromaeosaurs. Although possible, it is uncertain that other oogenera are also produced by oviraptorosaurs. Therefore, the relationship between elongatoolithid eggs and oviraptorosaurs should not be exclusively extended to all members of this group.
Conclusion
The EBSD technique provides a method to objectively distinguish microstructural differences in eggshells. Our results have revealed more structural information in elongatoolithid eggshells. We have shown that the structure of elongatoolithid eggshells is more diverse than traditionally thought, even in oospecies that are widely known like Macroolithus. The newly identified external zone makes Heishanoolithus no longer suitable as a member of the oofamily Elongatoolithidae. The combination of reticulated ornamentation, a continuous layer that shows columnar extinction, and acicular mammillae suggests that Paraelongatoolithus is not a member of Elongatoolithidae. We also discussed the variation of squamatic ultrastructure in fossil and recent eggshells and proposed practical criteria to distinguish them into two types. These findings may prove to be useful in the future classification of fossil eggshells and discussion about their producers and will help us in understanding the evolution and adaption of fossil and recent eggshells.
Acknowledgments
The authors would like to thank J. Yuan and X. Tang from IGGCAS for useful instruction and discussion on EBSD analysis. X. Jin from IVPP helped with SEM photography. This work was financially supported by National Natural Science Foundation of China (42288201, 41672012), Strategic Priority Research Program of the Chinese Academy of Sciences (XDB26000000), and the Beijing Government.
Competing Interests
The authors declare no competing interests.
Data Availability Statement
All raw data and images are available from the Science Data Bank: https://doi.org/10.57760/sciencedb.15572.
