Non-technical Summary
Sexual dimorphism—where males and females of a species look different—is common in extant and fossil records of crustaceans. In phyllocarids, the most basal group of malacostracans, sexual dimorphism has only been confirmed in some living leptostracans and a few well-preserved or soft-bodied archaeostracan fossils. After all, it is highly challenging to identify sexual dimorphism in lots of phyllocarid fossils that are preserved as isolated parts. In this study, we took another look at a group of Soomicaris ordosensis fossils from the Upper Ordovician of the North China and Tarim Plates. We used some geometric morphometric techniques (like elliptic Fourier analysis) to take a closer look at the sexual dimorphism in these fossils. The results show that S. ordosensis carapaces have significant sexual dimorphism. The length and shape of the carapace varied between male and female specimens, and in both groups, different parts of the body grew at different rates, which changed their proportions as they got larger. Furthermore, the two forms occur in roughly equal numbers, which suggests a simple male–female system. The geometric morphometrics used in this study are a highly effective way to figure out the subtle differences in the carapace morphology of phyllocarids. This highlights the hidden dimorphism of fossil animals.
Introduction
Sexual dimorphism, defined as the presence of morphological, physiological, or behavioral differences between the sexes (Accioly et al. Reference Accioly, Lima-Filho, Santos, Barbosa, Campos, Souza and Baebosa2014), is a well-documented phenomenon that is prevalent among extant crustaceans (e.g., Gilbert and Williamson Reference Gilbert and Williamson1983; Ledesma et al. Reference Ledesma, Van der Molen and Barón2010; Song et al. Reference Song, Moreira and Min2013; Ozawa Reference Ozawa and Moriyama2013; Nogueira et al. Reference Nogueira, Gois, Pescinelli and Costa2023). In addition to the evident disparities in sexual organs, substantial variations in morphology and size frequently manifest between males and females. Sexual dimorphism is the most commonly known form of dimorphism. However, it is challenging to attribute morphological differences in the fossil record to sex when genitalia or other reproductive structures are not preserved. Nevertheless, the existence of sexual dimorphism has been demonstrated or strongly suggested in several fossil groups, such as crabs and ammonites (e.g., Klug et al. Reference Klug, Zatoń, Parent, Hostettler, Tajika, Klug, Korn, DeBaets, Kruta and Mapes2015; Jones et al. Reference Jones, Schweitzer and Feldmann2022). In certain cases, it is therefore feasible to derive robust inferences regarding the presence of dimorphism in a fossil population, attributing this dimorphism to differences in sex. For instance, certain species of female ostracods develop brood pouches that become visible on their mineralized exoskeletons (Ozawa Reference Ozawa and Moriyama2013). Paleontologists had previously observed dimorphic populations in fossil clam shrimp for several decades before the demonstration of the correlation between brooding eggs and the dimorphism (Astrop et al. Reference Astrop, Park, Brown and Weeks2012 and references therein). Sexual dimorphism, a significant research topic in the field of evolutionary biology, is crucial for comprehending the mechanisms of natural selection and biological adaptation of extinct organisms.
Leptostracans, the most basal extant group of malacostracans, exhibit sexual dimorphism in species of the genera Nebalia Leach, Reference Leach1814, Nebaliella Thiele, Reference Thiele1904, and Sarsinebalia Dahl, Reference Dahl1985. Significant disparities have been documented in the morphology of various anatomic structures, including the antennae, carapace, and abdomen (e.g., Walker-Smith Reference Walker-Smith1998; Moreira et al. Reference Moreira, Cacabelos and Domínguez2003, Reference Moreira, Esquete and Cunha2021; Song et al. Reference Song, Moreira and Min2013 and references therein; Petryashov Reference Petryashov2017; Song and Min Reference Song and Min2017; Hirata et al. Reference Hirata, Fujiwara and Kikuchi2019, Reference Hirata, Rybakova, Simakova, Fujiwara, Moskalenko and Kikuchi2023; Supplementary Fig. 1A,B). For instance, the females of Nebalia pseudotroncosoi Song et al., Reference Song, Moreira and Min2013 develop rounded denticles along the posterior dorsal margin of pleonites 5 to 7, while the males show clearly acute denticles (Song et al. Reference Song, Moreira and Min2013). Leptostracans are regarded as the sole extant order within the class Phyllocarida; the remaining orders—Archaeostraca, Hymenostraca, and Hoplostraca—are extinct (Schram and Koenemann Reference Schram and Koenemann2021). The monophyly of Phyllocarida and Archaeostraca remains contentious (Hegna et al. Reference Hegna, Luque, Wolfe, Poore and Thiel2020), yet for the sake of convenience, they will be discussed herein as monophyletic.
Despite the presence of abundant and diverse phyllocarid fossils in Paleozoic strata (e.g., Rolfe Reference Rolfe, Brooks and Moore1969; Rode and Lieberman Reference Rode and Lieberman2002; Collette and Hagadorn Reference Collette and Hagadorn2010; Liu et al. Reference Liu, Fan, Zong and Gong2023a,Reference Liu, Poschmann, Fan and Gongb), the accurate identification of sexual dimorphism has proven challenging in most cases (Racheboeuf Reference Racheboeuf1998). Dimorphism has been documented in a limited number of well-preserved or soft-bodied archaeostracan specimens. These specimens typically possess hooked antennae, similar to those used for clasping in extant male Nebaliopsis Sars, Reference Sars1887 (Jones et al. Reference Jones, Feldmann and Mikulic2015; Poschmann et al. Reference Poschmann, Bergmann and Kühl2018). In general, the majority of phyllocarids are represented by isolated carapaces in Paleozoic strata (e.g., Jones and Woodward Reference Jones and Woodward1888, Reference Jones and Woodward1892; Collette and Hagadorn Reference Collette and Hagadorn2010; Racheboeuf and Crasquin Reference Racheboeuf and Crasquin2010). Consequently, the outline characteristics and the length-to-height ratio of the carapace serve as pivotal reference criteria for distinguishing between different species when the abdominal and tailpiece structures are not preserved. Rolfe (Reference Rolfe, Brooks and Moore1969) conducted statistical work on the Silurian Ceratiocaris M’Coy, Reference M’Coy1849, from Scotland, but no significant differences in carapace shape were found between forms regarded as sexual dimorphism and distinct species. This pattern is indicative of two factors. First, it reflects the inherent limitations of traditional morphometric methods. The length, width, the length-to-width ratio, and area of the carapace are incapable of accurately capturing variations in carapace contours. Second, the compression of the sediment further distorts the original outline of the carapace and, consequently, any measurements derived from it. Moreover, the paucity of fossil specimens represents a pivotal limiting factor that hinders the study of sexual dimorphism in phyllocarids.
Geometric morphometric techniques, such as landmarks, semi-landmarks, and Fourier analysis, have been widely applied to quantify and understand changes in shape and morphological variations in organisms. These techniques are successful because they can accurately capture morphological contours and discern subtle changes in curvature (Kuhl and Giardina Reference Kuhl and Giardina1982; Bookstein Reference Bookstein1997). They have been used to study sexual dimorphism in extant and fossil species (Astrop et al. Reference Astrop, Park, Brown and Weeks2012; Accioly et al. Reference Accioly, Lima-Filho, Santos, Barbosa, Campos, Souza and Baebosa2014; Martins et al. Reference Martins, Hunt, Thompson, Lockwood, Swaddle and Puckett2020), phylogeny (Astrop Reference Astrop2011), ontogeny (Webster Reference Webster2007), interspecific differences (Barría et al. Reference Barría, Sepúlveda and Jara2011; Ma et al. Reference Ma, Pates, Wu, Lin, Liu, Wu, Zhang and Fu2023), and intraspecific variation (Kapiris and Thessalou-Legaki Reference Kapiris and Thessalou-Legaki2001) in various fields. However, before applying these techniques to fossil material, one must evaluate taphonomic bias. In this study, carapaces that experienced only lateral compression and exhibited no discernible marginal distortion were selected to minimize the influence of sedimentary compression on outline morphology. Geometric morphometric analysis of carapace shape and size, particularly elliptic Fourier analysis (EFA), provides the best current method for detecting sexual dimorphism in fossil phyllocarids in which distinct sexual characteristics are absent.
Caryocaridids constituted a widely distributed group of planktonic arthropods during the Ordovician period, frequently found in association with graptolites in fine-grained sedimentary rocks, such as mudstones and shales (Vannier et al. Reference Vannier, Racheboeuf, Brussa, Williams, Rushton, Servais and Siveter2003; Liu et al. Reference Liu, Fan, Zong and Gong2022). In most cases, the carapaces are generally laterally compressed, superbly preserved, and abundant in quantity within the strata containing caryocaridid fossils. Therefore, caryocaridid carapaces are a useful sample set for studying sexual dimorphism in phyllocarids. This study focuses on a large sample of Soomicaris ordosensis Liu et al., Reference Liu, Fan, Zong and Gong2023a specimens from the Upper Ordovician in the North China and Tarim Plates. Through EFA of caryocaridid carapace morphology, we present the first quantitative evidence of sexual dimorphism in phyllocarids. This research offers novel concepts and methodologies for future studies on sexual dimorphism in phyllocarids. Additionally, it contributes to refining phyllocarid systematics and understanding reproductive biology.
Materials and Methods
Materials
A total of 153 carapaces of the caryocaridid S. ordosensis from Xinjiang (14 specimens) and Inner Mongolia (139 specimens) were utilized in this study (Supplementary Fig. 2). Subsequent data processing and analysis were exclusively conducted on the specimens from Inner Mongolia (Fig. 1), owing to the limited number of specimens from Xinjiang. Of the 139 specimens, nearly 10% exhibit obvious distortion, and 33 specimens have lost more than 20% of the carapace to breakage. Consequently, only 92 specimens (about 70%) retained intact or only slightly damaged outlines without obvious tectonic distortion. These were selected for morphometric analysis. Despite the two-dimensional compression of the specimens, their marginal contours are adequately preserved, enabling the acquisition of reliable outline data. The specimens from Inner Mongolia were collected from the upper part of the sixth bed (a thickness of approximately 1 m) of the lower part of the Lashizhong Formation in the Hatukegou section, Wuhai area, North China Plate (Liu et al. Reference Liu, Fan, Zong and Gong2023a; Supplementary Fig. 2D,F). These carapaces were particularly abundant in a 20-cm-thick layer of mudstones within the sixth bed. In addition to the phyllocarids, the same horizon also yielded abundant graptolites, thin-shelled brachiopods, and a small number of trilobites and cephalopods. The most recent biostratigraphic research indicates the distribution of S. ordosensis of Wuhai in the Wulalike Formation (late Darriwilian to early Sandbian, the upper part of the Didymograptus murchisoni and the Nemagraptus gracilis graptolite zones; Wang Reference Wang2024) and lower part of the Lashizhong Formation (late Sandbian, lower part of the Climacograptus bicornis graptolite zone; Wang Reference Wang2024). The specimens from Xinjiang were collected from the upper part of the Queerqueke Formation (late Sandbian, Climacograptus bicornis graptolite zone) in the Kuruktagh region, Tarim Plate (BGXUAR 1982; Zhong and Hao Reference Zhong, Hao, Zhong and Hao1990; Supplementary Fig. 2C,E). The fossils from the Wuhai area in Inner Mongolia are currently housed in the collections of the State Key Laboratory of Geomicrobiology and Environmental Changes (GMEC) at China University of Geosciences, Wuhan (CUG). This laboratory is the successor to the State Key Laboratory of Biogeology and Environmental Geology (BGEG) at CUG. The fossils from the Kuruktag region in Xinjiang are housed in the collections of the No.1 Regional Geological Survey Team, which is affiliated with the Xinjiang Bureau of Geology and Mineral Exploration and Development (XBGME).
Figure 1. Fossils and contour maps of different types of the Upper Ordovician Soomicaris ordosensis carapace. A, B, Type 1, BGEG–WHJ–25, XBGME–XPH–15. C, D, Type 2, BGEG–WHJ–54, XBGME–XPH–03. E, F, Type 3, BGEG–WHJ–14, XBGME–XPH–16. G, The outline diagrams of three types of caryocaridid carapaces. The arrow indicates anterior direction. Yellow, blue and green each represent one of the three carapace types.
Imaging and Data Preparation
All carapace specimens imaged were photographed using a Nikon D5100 camera with a 55 mm F3.5 micro-Nikkor lens mounted on a copy stand. The contours of the carapace were meticulously delineated by employing the Bezier curve functionality in CorelDRAW 2022. Subsequently, these contours were filled in to create a black silhouette against a white background. In life, the carapace of S. ordosensis exhibited mirrored left and right sides, united at a dorsal fold. In death, these two sides frequently become overprinted by one another. In such cases, the details of the other side were occasionally obscured by one of the two slightly displaced lateral compressions. The specimens were treated as a single specimen, and the outline was delineated by referring to the morphology of both carapaces. All specimens were imaged in lateral view, with the anterior oriented to the left. Specimens oriented to the right were reflected to ensure proper alignment. As an experiment, the analysis was run with one carapace flipped left to right. It could easily be distinguished from the correctly oriented cohort. Consequently, it can be concluded that the integrity of our results was not compromised by misorientation. The carapace length (L) and height (Hm) were measured with a vernier caliper under an Olympus SZ2-ILST stereomicroscope. Concurrently, the anterior angle of the carapace (α) was measured using CorelDRAW 2022, and the length and height data obtained using the vernier caliper were meticulously examined. Carapace perimeter and areas were measured using ImageJ software (Supplementary Table 1).
Data Analysis
After initial visual inspection, these specimens were classified into three morphotypes based on the contour of the posterior margin of the carapace (Fig. 1A–F): Type 1 (46 specimens, with a posterodorsal spine; Fig. 1A,B), Type 2 (6 specimens, with a transitional and faint spine; Fig. 1C,D), and Type 3 (40 specimens, with a spineless, convex posterior margin; Fig. 1E,F). To test the statistical validity of the three visually defined morphotypes, we performed EFA, followed by principal component analysis (PCA) and hierarchical clustering (Ward’s method) on the outline data.
The SHAPE 1.3 software package (National Agricultural Research Organization of Japan) was used to perform the EFA and PCA. This package consists of four programs: ChainCoder, CHC2NEF, PrinComp, and PrinPrint (Iwata and Ukai Reference Iwata and Ukai2002). First, ChainCoder was used to convert the outlines into vectorized objects, also referred to as chain codes. Then, the chain codes were transformed into elliptic Fourier descriptors (EFDs) and normalized using the CHC2NEF program (Iwata and Ukai Reference Iwata and Ukai2002; Braig et al. Reference Braig, Haug, Schädel and Haug2019). For the EFA, an appropriate maximum harmonic number (N = 20) was selected to fit the carapace outlines and conduct the morphological analysis (Fig. 2A–F). Finally, the PrinComp software was used to perform PCA on the variance–covariance matrix of the standardized coefficients. PrinPrint software produced a graphical output of the mean shape with standard deviations. To more intuitively display variations in carapace contour shape reflected by each principal component, EFDs of the carapace contours were calculated for the mean and ±2 standard deviations of each principal component. Then, the contour curve shapes were fit accordingly (Braig et al. Reference Braig, Haug, Schädel and Haug2019; Supplementary Fig. 3).
Figure 2. The contour-fitting effect of the caryocaridid carapace under different maximum harmonic numbers (A–F) and explanation diagram of the principle of change between the measured height (Hm) and original height (Ho) of the carapace (G, H).
Utilizing the approach established by Rolfe and Burnaby (Reference Rolfe and Burnaby1961), who defined the “size factor, Log (L*H)” and the “shape factor, Log (L/H)”, a comparative analysis was conducted among the three morphotypes of S. ordosensis carapaces. However, this method proved ineffective in adequately distinguishing them (Supplementary Fig. 4), suggesting that the carapace length-to-height ratio possesses minimal discriminatory capacity for these three carapace outlines. Subsequent to the PCA, it was ascertained that PC 1 predominantly captured variation in two metrics: overall carapace height and the length-to-height ratio. Subsequent to the analysis, it was determined using Rolfe and Burnaby’s (Reference Rolfe and Burnaby1961) approach that these aspects did not contribute meaningfully to group separation. PC 1 was therefore excluded from analyses concerning the three morphotypes. It is imperative to acknowledge that PC 1 is not “useless” in this context; rather, it captures the predominant axis of overall size. However, in the context of the dataset under consideration, it fails to effectively differentiate the morphological groups of interest. The between-group signal is instead carried on PC 2 and higher PCs. The retention of PC 1 for the purpose of hierarchical clustering analyses resulted in an augmentation of the Euclidean distance; however, it failed to enhance discriminatory power. Consequently, a hierarchical clustering analysis (Ward method) was conducted on PC 2–PC 6 using the linkage function (scipy.cluster package), which more effectively captured subtle morphological variations among the morphotypes.
Data Processing
The two carapace valves of extant phyllocarids exhibit a distinct convexity and manifest an overall oval-elliptical configuration in their longitudinal section. However, most phyllocarid carapaces preserved in mudstones exhibit a flattened morphology, oriented in a dorsal, ventral, or lateral direction, a consequence of the compaction process. Consequently, Rolfe (Reference Rolfe1962) proposed an adjustment to the measured height of the flattened carapace (Hm) to estimate the original height of the convex carapace (Ho) in life (illustrated in the reconstruction; Fig.2G). The utilization of the height/length ratios of multiple three-dimensionally preserved fossil phyllocarids enables the conversion of Hm to an approximate Ho. In Novák’s Reference Novák1885 study, he depicted the carapace of the Devonian Aristozoe Barrande, Reference Barrande1872 as having an overall oval longitudinal cross-section, with a short-to-long axis ratio (a/b) of 0.69 for its oval shape (Novák Reference Novák1885). The carapace of the three-dimensionally preserved Cinerocaris magnifica Briggs et al., Reference Briggs, Sutton, Siveter and Siveter2004, from the mid-Silurian (Wenlock) Herefordshire biota in the United Kingdom, has an a/b ratio of approximately 0.34 (Briggs et al. Reference Briggs, Sutton, Siveter and Siveter2004). The abdominal segment of the Silurian Ceratiocaris papilio Salter, Reference Salter and Murchison1859 from Scotland exhibits a longitudinal section with an a/b ratio of 0.76 (Rolfe Reference Rolfe1962). The abdominal segment of the Devonian Aristozoe from the Czech Republic exhibits a longitudinal section contour that varies from circular to elliptical, with an overall a/b range of 0.73–1.00 and an average of 0.92 (Barrande Reference Barrande1872; Fig. 2H). However, it should be noted that the carapace thickness of Aristozoe differs significantly from that of caryocaridids (Liu et al. Reference Liu, Fan, Du, Ma, Yin, Zong and Gong2024b). Research has demonstrated that thicker carapaces tend to exhibit greater compressive strength and undergo less deformation. Additionally, the curvature of the abdominal segment does not necessarily reflect the curvature of the bivalved carapace. The carapace thickness and lifestyle of C. magnifica have been found to be more similar to those of caryocaridids, and its three-dimensional preservation likely makes its carapace curvature more accurate. Consequently, the present study has sought to estimate the original height of the caryocaridid carapaces using the formula for the circumference of an ellipse and an a/b ratio of 0.34 (Supplementary Table 1). In addition, Origin 2019 software was utilized to statistically generate size distribution and violin plots based on carapace length and height.
Results
The first three principal components (PC 1, PC 2, and PC 3) in the PCA reflect more than 90% of the morphological variation in carapace outline. PC 1 reflects the trend of changes in carapace height and length: height ratio (see earlier discussion). As the value of PC 1 increases, the carapace exhibits an increased height relative to its length, and the convexity of the dorsal and ventral margins of the carapace becomes more pronounced. PC 2 and PC 3 are principally based on variations in the anterior and posterior margins of the carapace, given the relatively consistent height of the carapace. PC 4 primarily reflects variations in the posterior margin of the carapace, manifesting as convex, linear, and sigmoidal recessing to the right with a concave margin. Despite the presence of some overlap in the distribution areas of the three types in the PCA space, clear distinctions remain evident (see Fig. 3A,B). The distribution of Type 1 and 3 is characterized by PC 1 values. The PC 2 values of Type 1 are generally lower than those of Type 3, while the PC 3 values of Type 1 are overall higher than those of Type 3. In the PCA space, Type 2 is largely separated from Type 3 in the PC 1–PC 2 plane (Fig. 3A), while Type 2 and Type 1 overlap almost completely across both PC 1–PC 2 and PC 2–PC 3 (Fig. 3A,B). It is noteworthy that the first three principal components are responsible for the primary morphological variation of the carapace outline. Furthermore, Type 2 is analogous to Type 1 in terms of carapace outline, particularly in instances where the posterodorsal spine is not prominent, in which the posterior margin of the carapace may also appear linear (Fig. 1D). Overall, the number of Type 2 specimens is low compared with Types 1 and 3 in the entire group. It appears that a more appropriate interpretation of Type 2 is as a partial variation of Type 1. These results are further substantiated by the findings of hierarchical clustering analysis (Fig. 3C). Despite the fact that this analysis did not differentiate between all specimens of Types 1 and 3, they can be distinctly classified into two morphological groups. All Type 2 specimens are also recovered into a group that is closely related to Type 1. Therefore, based on the evident posterior margin configuration of the carapace and the findings of multivariate analysis, the carapace morphology of S. ordosensis can be categorized into two robust forms (Types 1+2 and Type 3); Type 2 is consistently nested within Type 1. Consequently, in the subsequent combined analyses of carapace allometry and population‐level statistics, all data of Type 2 should be merged into those of Type 1.
Figure 3. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) of the Upper Ordovician Soomicaris ordosensis carapace. A, B, PCA shows the visualization results of elliptic Fourier analysis (EFA) of the morphology of three different caryocaridid carapaces. Colors are the same as in Fig. 1G. C, Hierarchical clustering results on elliptic Fourier analysis of S. ordosensis carapace shapes. The colors for the carapace types are the same as in Fig. 1G.
The scatter plot analysis of the length and height data of caryocaridid carapaces shows that the length and height of S. ordosensis carapaces are strongly linearly correlated. The fitted slopes of the length-to-height ratio for Types 1+2 and Type 3 are 3.1487 and 2.8739, respectively, with high degrees of fit (R² = 0.7677 and 0.7175). This indicates a clear difference in the length-to-height ratio of the carapaces between the two types. A regression analysis was performed using the allometric growth equation Y = aXᵇ as a template. Taking the length (L) of the laterally compressed carapace as the independent variable (X) and the carapace height (Ho) as the dependent variable (Y), the equation was linearized to lnY = ln(a) + bln(X). The regression analysis shows that the allometric growth coefficients for the carapaces of the two forms of S. ordosensis are 0.9313 and 0.976, respectively, indicating that the height and length of the carapace generally follow an allometric growth pattern during development (Fig. 4A). Like all euarthropods, S. ordosensis likely grew through a series of molting stages (instars), so the measured carapaces therefore represent a mixture of exuviae and subadult or adult individuals. The violin plots show that the carapace length of Types 1+2 ranges from 6.55 to 17.79 mm (average 11.6 mm) and the carapace height ranges from 1.73 to 4.75 mm (Fig. 4B). The total length of Type 3 ranges from 6.11 to 17.86 mm (average 11.57 mm) and the carapace height ranges from 1.33 to 5.27 mm. There is a small difference in length between Types 1+2 and 3, with the most abundant group measuring around 10 mm long. However, the most abundant width in Types 1+2 is significantly lower than in Type 3 (Fig. 4B). Both carapace morphotypes (Types 1+2 and Type 3) are present across all size classes in the entire population, including small individuals with carapace lengths of 6–8 mm. This suggests that sexually dimorphic carapace traits likely became established early in ontogeny. The absence of specimens with a carapace length less than 5 mm may reflect either poor preservation potential for early juveniles or that these individuals occupied a distinct ecological niche.
Figure 4. Data analysis chart of the two different forms (Types 1+2, Type 3) of Soomicaris ordosensis carapace. A, The individual development relationships between the measured length (L) and original height (Ho) of the carapace. B, Violin diagrams depicting the measured length (L) and original height (Ho) of the carapace.
Discussion
The recognition of two distinct morphological groups within a population of fossil arthropods immediately suggests the possibility of sexual dimorphism as the underlying cause. In gonochoric species, the population is divided into two distinct categories: females and males. The members of the different sexes can often be recognized by traits other than their genitalia, including size and color. Anderson (Reference Anderson1994) proposed that disparities in size, shape, ornamentation, and other external morphological characteristics of the same species within the same horizon of fossils can substantiate the interpretation of sexual dimorphism for this organism. In numerous extant arthropods, sexual dimorphism manifests only after sexual maturity, frequently appearing abruptly during the final molt. However, mounting evidence from fossil and modern taxa indicates that certain dimorphic traits may emerge at an earlier stage (Kamiya Reference Kamiya1992; Ozawa Reference Ozawa and Moriyama2013; Fu et al. Reference Fu, Zhang, Budd, Liu and Pan2014; Marochi et al. Reference Marochi, Costa, Leite, Da Cruz and Masunari2019). For instance, in the Cambrian arthropod Isoxys Walcott, Reference Walcott1890, dimorphic carapace morphologies are discernible even in late juvenile stages (length of carapace >10 mm) (Fu et al. Reference Fu, Zhang, Budd, Liu and Pan2014). Similarly, in certain extant brachyurans, sexual dimorphism in carapace size can be identified during the juvenile phase (Marochi et al. Reference Marochi, Costa, Leite, Da Cruz and Masunari2019). The presence of sexual dimorphism in small-sized specimens suggests that this divergence is not merely a secondary sexual characteristic tied to reproductive maturity, but rather a stable trait established early in development. Consequently, even fossil assemblages dominated by juvenile or preadult individuals or exuviae—as is common in phyllocarids—may still preserve detectable evidence of sexual dimorphism. Nevertheless, in the absence of preserved genitalia or reproductive structures, establishing a definitive correlation between these auxiliary traits and sex remains a formidable challenge. Consequently, a considerable proportion of the dimorphism observed in the fossil record cannot be unequivocally attributed to sexual dimorphism.
The key to identifying sexual dimorphism is the presence of a specific morphological trait that systematically separates the population into two distinct groups: males and females. In the fossil record, the fossilization of genitalia is an infrequent occurrence (e.g., Størmer and Kjellesvig-Waering Reference Størmer, Kjellesvig-Waering and Westermann1969; Trusova Reference Trusova1971; Siveter et al. Reference Siveter, Sutton, Briggs and Siveter2003; Orr et al. Reference Orr, Briggs and Kearns2008; Shen and Schram Reference Shen and Schram2014; Schweitzer et al. Reference Schweitzer, Feldmann, Audo, Charbonnier, Fraaije, Franţescu and Hyžný2024). Consequently, reliable identification of sex is rendered extremely difficult. The presence of preserved evidence of brooding or egg-laying can serve as indirect indicators of sex. However, the presence of preserved arthropod eggs is also relatively rare, with reports limited to certain arthropods (e.g., Siveter et al. Reference Siveter, Siveter, Sutton and Briggs2007, Reference Siveter, Tanaka, Farrell, Martin, Siveter and Briggs2014; Hegna et al. Reference Hegna, Martin and Darroch2017; Jauvion et al. Reference Jauvion, Audo, Bernard, Vannier, Daley and Charbonnier2020; Ou et al. Reference Ou, Vannier, Yang, Chen, Mai, Shu, Han, Fu, Wang and Mayer2020; Wang et al. Reference Wang, Matzke-Karasz, Horne, Zhao, Cao, Zhang and Wang2020; Fu et al. Reference Fu, Cai, Chen and Huang2022; Moon et al. Reference Moon, Caron and Gaines2022; Van Houte et al. Reference Van Houte, Hegna and Butler2022; Charbonnier et al. Reference Charbonnier, Gilardet, Garassino and Odin2023; Ma et al. Reference Ma, Pates, Wu, Lin, Liu, Wu, Zhang and Fu2023; Dias et al. Reference Dias, Carvalho, Souza-Dias, Zefa, Barros, Prado and Osés2025; Lamsdell et al. Reference Lamsdell, Falk, Hegna and Meyer2025; Laville and Hegna Reference Laville, Hegna, Charbonnier and Forel2025). Among the numerous bivalved arthropods, the females typically brood their young beneath the carapace. Studies have found that carrying eggs or embryos has a distinct impact on the carapace morphology of female specimens in comparison to male specimens in the clam shrimps (Astrop et al. Reference Astrop, Park, Brown and Weeks2012) and ostracod species (Ozawa Reference Ozawa and Moriyama2013). In certain ostracods, the presence of an extended male copulatory apparatus can also result in carapace elongation (Horne et al. Reference Horne, Danielopol, Martens and Martens1998; Hunt et al. Reference Hunt, Martins, Puckett, Lockwood, Swaddle, Hall and Stedman2017). While the distinction between male and female morphs may not always be discernible, sex has been identified as a primary factor contributing to carapace dimorphism.
This insight has prompted the tentative inference of sexual dimorphism in various fossil bivalved arthropods, including bradoriids (Zhang Reference Zhang1987), Isoxys (Fu et al. Reference Fu, Zhang, Budd, Liu and Pan2014; Nielsen et al. Reference Nielsen, Rasmussen and Harper2017), Branchiocaris Briggs, Reference Briggs1976 (Wu et al. Reference Wu, Fu, Zhang, Daley and Shu2016), ostracods (Alexander Reference Alexander1932; Hunt et al. Reference Hunt, Martins, Puckett, Lockwood, Swaddle, Hall and Stedman2017), and clam shrimp (Astrop et al. Reference Astrop, Park, Brown and Weeks2012; Monferran et al. Reference Monferran, Gallego, Astrop and Cabaleri2013; Stigall et al. Reference Stigall, Hembree, Gierlowski-Kordesch and Weismiller2014; Hethke et al. Reference Hethke, Fürsich, Schneider and Jiang2017). However, a rigorous testing of some of these analyses based on simple morphometric data (length, height, L/H ratio) is still necessary. It is evident that the process of determining the sex of fossils would be greatly facilitated if eggs were more frequently preserved. However, there may be behavioral reasons why eggs are not more frequently preserved within fossil carapaces. For instance, female clam shrimp that are either dying or in a habitat that has nearly dried out will rapidly shed their eggs by flexing their tails (T.A.H., personal observation).
The gonopores, which are located near the bases of the sixth or eighth thoracomeres, have been identified as a significant taxonomic characteristic for distinguishing between male and female extant leptostracans (Schram and Koenemann Reference Schram and Koenemann2021). However, this feature is often challenging to observe in modern animals (Haney and Martin Reference Haney and Martin2005) and is not feasible in fossil remains. The most prominent sexual dimorphism exhibited by leptostracans is primarily manifested in external morphology, including the antennules, antennae, furca, pleopods, and carapace (Supplementary Fig. 1A,B). However, at present, sexual dimorphism has only been well-documented in Nebalia (Wägele Reference Wägele1983; Song et al. Reference Song, Moreira and Min2013 and references therein; Song and Min Reference Song and Min2017; Moreira et al. Reference Moreira, Esquete and Cunha2021; Hirata et al. Reference Hirata, Rybakova, Simakova, Fujiwara, Moskalenko and Kikuchi2023), Nebaliella (Rolfe Reference Rolfe, Brooks and Moore1969; Walker-Smith Reference Walker-Smith1998; Petryashov Reference Petryashov2017; Hirata et al. Reference Hirata, Fujiwara and Kikuchi2019), and Sarsinebalia (Moreira et al. Reference Moreira, Esquete and Cunha2021).
A comparison of the morphological evidence of sexual dimorphism in extant leptostracans and phyllocarid fossils reveals that sexual dimorphism is often not easily identifiable in the latter. Notable exceptions to this pattern include certain Silurian and Devonian species that preserve a large amount of appendage information (Jones et al. Reference Jones, Feldmann and Mikulic2015; Poschmann et al. Reference Poschmann, Bergmann and Kühl2018). Hannibal (Reference Hannibal1990) noted that multiple species of Echinocaris Whitfield, Reference Whitfield1880 had been documented in numerous Upper Devonian deposits within the United States. This observed variation could be attributed to sexual dimorphism. However, further confirmation is challenging due to the limited number of specimens available. Schram and Koenemann’s (Reference Schram and Koenemann2021) study indicated that Echinocaris exhibits remarkably subtle morphological variations in the Upper Devonian Chagrin Shale in the United States. These variations may be indicative of sexual dimorphism. Racheboeuf (Reference Racheboeuf1998) found that the anterior region of the furcal rami in most specimens of Dithyrocaris oculeus Racheboeuf, Reference Racheboeuf1998 is decorated with oblique lines on both dorsal and ventral sides. However, some specimens lack the ornamentation on the anteroventral side of the furcal rami and have a slightly outward bend in their distal region. It was hypothesized that these characteristics might be indicative of sexual dimorphism rather than the presence of a distinct species. This difference has been also observed in Nebalia bipes (Fabricius, Reference Fabricius1780) (Vannier et al. Reference Vannier, Boissy and Racheboeuf1997). Furthermore, Collette and Hagadorn (Reference Collette and Hagadorn2010) proposed that Ceratiocaris maccoyanus Hall, Reference Hall1859 and C. acuminata Hall, Reference Hall1859 exhibited numerous similarities, suggesting that the former may be a morphological variation or sexual dimorph of the latter.
Our geometric morphometric analysis of the caryocaridids S. ordosensis from the Upper Ordovician Lashizhong Formation of North China primarily identifies two distinct forms. In terms of the overall configuration, the posterior margin of the carapace of Type 3 exhibits notable similarities with that of the genus Rhinopterocaris Chapman, Reference Chapman1903 (Liu et al. Reference Liu, Bicknell, Smith, Fan, Richards, Terezow, Zong and Gong2024a). However, apart from the evident disparities in the posterior margin of the carapace, Types 1 and 3 exbibit a largely congruent anterior margin of the carapace, and the angle range of the anterior horn of the carapace is predominantly between 75° and 85° (Supplementary Table 1). Therefore, the classification of Type 3 as a species of Rhinopterocaris based solely on the characteristics of the posterior margin of the carapace appears to be an inaccurate and potentially misleading attribution. Furthermore, caryocaridids demonstrated a planktonic lifestyle and exhibited the capacity for vertical migration in open waters during the Ordovician (Vannier et al. Reference Vannier, Racheboeuf, Brussa, Williams, Rushton, Servais and Siveter2003; Liu et al. Reference Liu, Fan, Zong and Gong2023a). Sedimentological evidence related to fluctuating environmental and ecological conditions of the Lashizhong Formation in the Wuhai area was not found. Moreover, in addition to S. ordosensis, Saltericaris subula (Chlupáč, Reference Chlupáč1970), which is characterized by a distinctly developed anterior horn of the carapace, was also identified in the same stratigraphic horizon. Consequently, the likelihood of ecological phenotypes being the causative agent for the two distinct types in S. ordosensis appears improbable. In numerous extant crustaceans, substantial disparities in the dimensions and morphology of the carapace are observed between male and female individuals (Smith and Hiruta Reference Smith and Hiruta2004; Astrop et al. Reference Astrop, Park, Brown and Weeks2012; Ozawa Reference Ozawa and Moriyama2013; Petryashov Reference Petryashov2017; Song et al. Reference Song, Moreira and Min2013). In general, females tend to grow larger than males and attain larger sizes than their male counterparts (Gopal et al. Reference Gopal, Gopikrishna, Krishna, Jahageerdar, Rye, Hayes and Paulpandi2010). For example, in the phyllocarids Nebalia and Nebaliella, the male carapace is comparatively diminutive and less elliptical in shape compared with the female one. Furthermore, the ventral margin of the female carapace is typically enlarged and more rounded (e.g., Moreira et al. Reference Moreira, Cacabelos and Domínguez2003, Reference Moreira, Esquete and Cunha2021; Haney and Martin Reference Haney and Martin2000, Reference Haney and Martin2005; Song et al. Reference Song, Moreira and Min2013; Supplementary Fig. 1). Consequently, despite the lack of evidence of soft tissue structure in both types of S. ordosensis, we hypothesize that these two types of different morphological carapace specimens are more likely to reflect sexual dimorphism rather than the existence of two different species.
In extant phyllocarids, eggs or embryos are not, as is the case in the myodocopids and notostracans, retained in the posterior or ventral region of the carapace. Instead, they are found within a basket-like structure created by the intersecting network of the endopodial setae of the thoracic appendages (Shu et al. Reference Shu, Vannier, Luo, Chen, Zhang and Hu1999: fig. 8; Brusca et al. Reference Brusca, Moore and Shuster2016: fig. 21.5D; Hirata et al. Reference Hirata, Fujiwara and Kikuchi2019: fig. 7C). However, these eggs are also found to be enveloped within the carapace, akin to the thoracic appendages. The large and rounded ventral margin of the female phyllocarid carapace appears to be attributable to the brooding of eggs beneath the carapace, although further morphometric analysis is necessary to substantiate this hypothesis. Consequently, we propose that the dimorphism of the carapace outline of S. ordosensis may be associated with the female brooding eggs beneath the carapace, analogous to the behavior of modern leptostracans. This explanation not only aligns with the expectations of morphological and functional studies, but also provides the first indirect evidence for the reproductive biology of archaeostracans. In S. ordosensis, the combination of a specific behavior (egg-brooding) with a distinct dimorphic trait (carapace shape) serves to reinforce the conclusion of sexual dimorphism. It important to note that comparisons in terms of length, height, perimeter, and area between these two types of carapace do not distinguish which morphology has a larger carapace. The presence of dimorphism is distinguished by the observation of distinct morphologies rather than by the measurement of size.
It is important to emphasize that the sexual dimorphism observed in S. ordosensis is not confined to the Wuhai area of the North China Plate. A variety of caryocaridid specimens have been retrieved from the Upper Ordovician strata of the Kuruktagh region (Queerqueke, Yuanbao, and Ya’erdang Mountains) in the northeastern margin of the Tarim Plate (BGXUAR 1982; Zhong and Hao Reference Zhong, Hao, Zhong and Hao1990). A reexamination of previously unstudied legacy collections reveals three distinct carapace morphotypes of S. ordosensis within the upper part of the Queerqueke Formation (Fig. 1B,D,F). Despite the initial intention to expand the scope of this study through additional sampling, access to the area has been restricted due to its designation as a protected zone. However, the synchronous co-occurrence of both morphotypes in coeval strata of the Tarim and North China Plates (Supplementary Fig. 2A) demonstrates that S. ordosensis exhibits pronounced sexual dimorphism and that this pattern is not a local or fortuitous phenomenon.
A significant aspect of research in the domain of evolutionary biology entails the analysis of the morphological characteristics that distinguish between male and female individuals. A further focal point of this field of study is the investigation of the sex ratio within a given population. By examining the ratio or percentage of males to females, it is possible to gain insight into the mating system of the population or species in question (Weeks et al. Reference Weeks, Sanderson, Zofkova and Knott2008). In S. ordosensis, Types 1+2 account for approximately 56.5% of the population. Excluding Type 2, Type 1 accounts for 53.5% of the population. These values are proximate to the 1:1 ratio observed in numerous gonochoric species. A preliminary census of a substantial number of specimens of various Echinocaris species from the Upper Devonian Chagrin Shale of the United States does not reveal any consistent ratio between any two common species that might indicate equal numbers of males and females (Hannibal Reference Hannibal1990). However, numerous extant crustaceans deviate from the anticipated 1:1 sex ratio (e.g., Lewbel Reference Lewbel1978; Vonk and Nijman Reference Vonk and Nijman2006; Gusmão and McKinnon Reference Gusmão and McKinnon2009), and this unequal sex ratio appears to be closely related to different ecological niches (Ewers-Saucedo Reference Ewers-Saucedo2019) and/or mating systems (Weeks et al. Reference Weeks, Sanderson, Zofkova and Knott2008). In the case of clam shrimp, the ratio of the two carapace morphs is indicative of the mating system. The 50:50 ratio indicates a population of males and females in equal proportions (dioecy), a 25:75 ratio suggests a male–hermaphrodite population (androdioecy), and the existence of only one carapace morph indicates the occurrence of parthenogenesis (Astrop et al. Reference Astrop, Park, Brown and Weeks2012; Weeks et al. Reference Weeks, Sanderson, Zofkova and Knott2008). At the same time, fossil groups exhibit certain sex biases. For instance, in the Cretaceous lacustrine strata of Poland, the sex ratio of females to males in certain ostracod species can reach 3:1 or 2:1 (Namiotko and Martins Reference Namiotko and Martins2008). In contrast, the Late Cretaceous crab Dakoticancer overanus Rathbun, Reference Rathbun1917, from the United States, exhibited a male-biased sex ratio (79% males vs. 21% females) (Jones et al. Reference Jones, Schweitzer and Feldmann2022), the reverse of the female-biased ratios reported above. Additionally, the conspicuous absence of males in a population of fairy shrimp from the Cretaceous of Australia has been interpreted as evidence of parthenogenesis (Van Houte et al. Reference Van Houte, Hegna and Butler2022).
In the extant species of leptostracans, Gerken (Reference Gerken1995) found that males constituted less than 15% of the collected specimens in a sample of Nebalia pugettensis (Clark, Reference Clark1932). However, she also mentioned that the sex of many individuals in each sample was uncertain, suggesting the potential for numerous unidentified specimens to be males. Subsequent research by Vetter (Reference Vetter1996) revealed that female specimens of Nebalia sp. in California, USA, constituted approximately 75% of the adult population. Concurrently, Haney and Martin (Reference Haney and Martin2005) observed that in the collected N. gerkenae Haney and Martin, Reference Haney and Martin2000, males constituted 30% of the individuals with determinable sex. In populations of Nebalia, the ratio of female to male individuals exhibits a pronounced skew toward females (Gerken Reference Gerken1995; Vetter Reference Vetter1996; Haney and Martin Reference Haney and Martin2005). The ratio of carapace morphs in the Ordovician caryocaridid S. ordosensis tentatively supports an interpretation of a dioecious mating system.
In summary, geometric morphometrics reveal sexual dimorphism in Ordovician S. ordosensis carapaces, with two shape-based morphotypes consistent across the North China and Tarim Plates. This provides the first quantitative evidence for the reproductive biology of archaeostracans preserved only as isolated carapaces, thereby emphasizing the utility of outline analysis for detecting cryptic dimorphism in arthropod carapaces.
Acknowledgments
We extend our gratitude to Z. Wang (Henan Polytechnic University), D. Tu, Q.Y. Xu, and X.Q. Du (China University of Geosciences, Wuhan) for their invaluable assistance during the fieldwork. We are equally grateful to Q.Q. Song (No.1 Regional Geological Survey Team, Xinjiang Bureau of Geology and Mineral Exploration and Development) for facilitating our search for the phyllocarid specimens of Xinjiang. And we extend our appreciation to J.-H. Song (Nakdonggang National Institute of Biological Resources) for providing images of extant leptostracans. Finally, we greatly appreciate the critical comments and detailed suggestions provided by the reviewers C. E. Schweitzer and J. Vannier, as well as editor A. Klompmaker, which have significantly improved the article. This work was supported by the National Natural Science Foundation of China (nos. 42072041, 41902002, 42272014, 41872034).
Competing Interests
The authors declare no competing interests.
Data Availability Statement
Data available including figures and table from the Dryad Digital Repository: https://doi.org/10.5061/dryad.7h44j1087.
Open access
