Introduction
Molecular studies have suggested the first appearance of eumetazoans in the Ediacaran Period or earlier (Peterson et al., Reference Peterson, Cotton, Gehling and Pisani2008; Erwin et al., Reference Erwin, Laflamme, Tweedt, Sperling, Pisani and Peterson2011). However, this is not supported by fossil evidence because undisputed animal fossils in Precambrian rocks are extremely rare. The nature of fossil embryos from the Ediacaran Doushantuo Formation of South China (Xiao et al., Reference Xiao, Zhang and Knoll1998) is vague and has remained hotly debated (Xue et al., Reference Xue, Zhou and Tang1999; Bailey et al., Reference Bailey, Joye, Kalanetra, Flood and Corsetti2007; Huldtgren et al., Reference Huldtgren, Cunningham, Yin, Stampanoni, Marone, Donoghue and Bengtson2011; Chen et al., Reference Chen, Xiao, Pang, Zhou and Yuan2014). The material assigned to the earliest bilaterian animal Vernanimalcula guizhouena Chen et al., Reference Chen, Bottjer, Oliveri, Dornbos, Gao, Ruffins, Chi, Li and Davidson2004 represents, most likely, artifacts of diagenesis (Bengtson et al., Reference Bengtson, Cunningham, Yin and Donoghue2012). On the other hand, the fossil evidence demonstrates that all major groups of animals are already present early in the Cambrian Period (Budd, Reference Budd2003a, Reference Budd2013; Chen et al., Reference Chen, Waloszek, Maas, Braun, Huang, Wang and Stein2007; Erwin et al., Reference Erwin, Laflamme, Tweedt, Sperling, Pisani and Peterson2011; Erwin and Valentine, Reference Erwin and Valentine2013; Budd and Jackson, Reference Budd and Jackson2015). This indicates that the early evolution of animals took place well within the end of the Precambrian (Zhu et al., Reference Zhu, Zhuravlev, Wood and Sukhov2017).
Cambrian strata preserve the earliest unambiguous animals and are thus crucial for uncovering not only the origin but also the major early steps and significant developments in metazoan evolution. The Fortunian Stage, the oldest stage of the Cambrian Period (Peng et al., Reference Peng, Babcock and Cooper2012), is well known for the occurrence of skeletal remains known as small shelly fossils (SSFs). SSFs have a worldwide distribution, such as in Siberia (Voronova and Missarzhevsky, Reference Voronova and Missarzhevsky1969; Val’kov, Reference Val’kov1983), Australia (Bengtson et al., Reference Bengtson, Conway Morris, Cooper, Jell and Runnegar1990), and South China (Qian and Bengtson, Reference Qian and Bengtson1989; Steiner et al., Reference Steiner, Li, Qian and Zhu2004a), and they form the foundation of a global biostratigraphic correlation (Steiner et al., Reference Steiner, Li, Qian and Zhu2004a, Reference Steiner, Li, Qian, Zhu and Erdtmann2007). Another important group of exceptionally preserved microfossils include taxa such as Olivooides Qian, Reference Qian1977 (Bengtson and Yue, Reference Bengtson and Yue1997; Donoghue et al., Reference Donoghue, Kouchinsky, Bengtson, Cunningham, Dong, Repetski, Val’kov and Waloszek2006a; Dong et al., Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016), Quadrapyrgites Li et al., Reference Li, Hua, Zhang, Zhang, Jin and Liu2007 (Liu et al., Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a; Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014), Pseudooides Qian, Reference Qian1977 (Steiner et al., Reference Steiner, Zhu, Li, Qian and Erdtmann2004b), a sea anemone (Han et al., Reference Han, Kubota, Uchida, Stanley, Yao, Shu, Li and Yasui2010), the oldest known scalidophoran animals (Liu et al., Reference Liu, Xiao, Shao, Broce and Zhang2014b; Zhang et al., Reference Zhang, Xiao, Liu, Yuan, Wan, Muscente, Shao, Gong and Cao2015; Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016), the oldest known eumetazoan larvae (Zhang and Dong, Reference Zhang and Dong2015), and a plausible deuterostomian animal (Han et al., Reference Han, Conway Morris, Ou, Shu and Huang2017). These exceptionally preserved microfossils have stimulated divergent interpretations and exciting debate on the origin and early evolution of animals.
In the early Cambrian, the present-day southern Shaanxi and northern Sichuan Provinces were located at the northern edge of the Yangtze Platform (Steiner et al., Reference Steiner, Li, Qian and Zhu2004a). On the Yangtze Platform, sediments deposited during the Fortunian Stage contain abundant SSFs and non-mineralized microfossils, which were exceptionally preserved by three-dimensional phosphatization. Representatives of such localities are the Meishucun section in Yunnan Province (Qian and Bengtson, Reference Qian and Bengtson1989; Yang et al., Reference Yang, Steiner, Li and Keupp2014) and the Shizhonggou section in Shaanxi Province (Steiner et al., Reference Steiner, Li, Qian and Zhu2004a).
This paper presents the latest discoveries from another fossil locality, the Zhangjiagou section in southern Shaanxi Province, China. The Zhangjiagou section has yielded, besides abundant SSFs, representatives of the taxa Olivooides, Pseudooides, and Quadrapyrgites, and a diverse suite of cycloneuralians. The current study focuses on the early diversification of the cnidarians (Olivooides, Pseudooides, Quadrapyrgites, and Hexaconularia) and cycloneuralians from this Fortunian Konservat-Lagerstätte.
Geological background and age constraint
The Zhangjiagou Lagerstätte (Fig. 1.1) is located in Dahe Village, Xixiang County, Hanzhong City, Shaanxi Province, China (Fig. 2). The fossil locality was first measured and described as Zhangjiagou section by Li (Reference Li1984), and later as Hexi section by Steiner et al. (Reference Steiner, Qian, Li, Hagadorn and Zhu2014) and as Xixiang section by Liu et al. (Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a, Reference Liu, Xiao, Shao, Broce and Zhangb). The suite of limestone-dominated deposits was identified as Kuanchuanpu Formation (Li, Reference Li1984). The Kuanchuanpu Formation at Zhangjiagou section is ~21 m thick and consists of four members in ascending order (Fig. 2): the 1st member is a light-gray microsparitic limestone (0.8 m); the 2nd member a phosphatic limestone (2.2 m); the 3rd member consists of thick-bedded, dark microsparitic limestone (17.4 m); and the 4th member is a thin-bedded dolomitic limestone (0.6 m). The Kuanchuanpu Formation is underlain conformably by siliceous dolostone of the Ediacaran Dengying Formation, and overlain disconformably by black shale of the Guojiaba Formation. The fossil collection reported here is exclusively from the lower part of the 2nd member (Figs. 1.2, 2). The phosphatic limestone from the 2nd member contains abundant apatite (Fig. 1.3), implying that the Fortunian Yangtze Sea contained a high amount of phosphate, which increased preservation potential of the microfossils.
The Zhangjiagou section in southern Shannxi Province, South China. (1) Overall view; the boundary between the Ediacaran Dengying Formation and the Cambrian Kuanchuanpu Formation, and the subdivisions (the 1st, 2nd, and part of 3rd members) within the lower part of the Kuanchuanpu Formation, are marked; (2) close-up of the boundary between the 1st and 2nd members of Kuanchuanpu Formation; (3) hand specimens of phosphatic limestone from the lower part of the 2nd member.
Location map and stratigraphic column of the Zhangjiagou section in southern Shaanxi Province, South China. Arrow indicates the key horizon yielding the present materials. Revised from Liu et al. (Reference Liu, Xiao, Shao, Broce and Zhang2014b).
The lower part of the 2nd member of the Kuanchuanpu Formation at Zhangjiagou section has yielded abundant SSFs, including the taxa Anabarites Missarzhevsky in Voronova and Missarzhevsky, Reference Voronova and Missarzhevsky1969 (Fig. 3.1–3.6), Protohertzina Missarzhevsky, Reference Missarzhevsky1973 (Fig. 3.7–3.9), long and internally hollow tubes (Fig. 3.10, 3.11), Acanthocassis He and Xie, Reference He and Xie1989 (Fig. 3.12), multi-branched fossils (Fig. 3.13), maikhanellid shells (Fig. 3.14), and siphogonuchitid spicule bundles (Fig. 3.15, 3.16). The occurrence of Anabarites trisulcatus Missarzhevsky in Voronova and Missarzhevsky, Reference Voronova and Missarzhevsky1969 (Fig. 3.1), Protohertzina anabarica Missarzhevsky, Reference Missarzhevsky1973 (Fig. 3.7, 3.8), and Protohertzina unguliformis Missarzhevsky, Reference Miranda, Hirano, Mills, Falconer, Fenwick, Marques and Collins1973 (Fig. 3.9) indicates that the key horizon yielding the present materials is part of the SSFs Anabarites trisulcatus-Protohertzina anabarica Assemblage Zone, which has been estimated to be ca. 535 Ma (Steiner et al., Reference Steiner, Li, Qian, Zhu and Erdtmann2007, Reference Steiner, Qian, Li, Hagadorn and Zhu2014) within the Fortunian Stage (Peng et al., Reference Peng, Babcock and Cooper2012).
Small shelly fossils from the Zhangjiagou Lagerstätte. (1) Anabarites trisulcatus Missarzhevsky in Voronova and Missarzhevsky, Reference Voronova and Missarzhevsky1969, UMCU.XXAN14068; (2, 3) Anabarites ex gr. A. trisulcatus Missarzhevsky in Voronova and Missarzhevsky, Reference Voronova and Missarzhevsky1969, form 2; (2) UMCU.XXAN14069; (3) UMCU.XXAN14070; (4–6) Anabarites hexasulcatus (Missarzhevsky, Reference Missarzhevsky1974); (4) UMCU.2014WB526; (5) UMCU.2014WB528; (6) UMCU.2014WB238; (7, 8) Protohertzina anabarica Missarzhevsky, Reference Missarzhevsky1973; (7) UMCU.2014HZ001; (8) UMCU.2014HZ002; (9) Protohertzina unguliformis Missarzhevsky, Reference Missarzhevsky1973, UMCU.2014HZ003; (10, 11) long tubes with annular rings, internally hollow; (10) UMCU.2013XXG002; (11) UMCU.2013XXG001; (12) Acanthocassis orthacanthus (Yang and He, Reference Yang and He1984) He and Xie, Reference He and Xie1989, with a long stalk and five major branches, UMCU.XJ5-3A; (13) a multi-branched specimen, possible internal mold, UMCU.2014XXG001; (14) a maikhanellid shell, UMCU.XXMH1416; (15, 16) siphogonuchitid spicule bundles; (15) UMCU.XXLG1406; (16) UMCU.XXLG1410. Scale bar represents 1 mm (1–13), and 410 μm (14–16).
Materials and methods
All specimens were extracted from the phosphatic limestone (Fig. 1.3) from the lower part of the 2nd member of the Kuanchuanpu Formation at Zhangjiagou section in southern Shaanxi, South China (Figs. 1.1, 2). The rock samples were dissolved in acetic acid following procedures described in Müller (Reference Müller1985). Rock fragments were immersed in diluted acetic acid (~8%), and residues were retrieved regularly after seven days of reaction. The undissolvable residues were air dried, and microfossils were handpicked using a binocular microscope. Selected microfossils were mounted on aluminum stubs of pin type for observation under scanning electron microscopy (SEM). SEM images were further processed using Adobe Photoshop CS5, with the background cleared away and the brightness and contrast of the whole images adjusted. Photographs of the Zhangjiagou section (Fig. 1.1, 1.2) and the hand specimens of the phosphatic limestone (Fig. 1.3) were taken using a Canon 5DsR digital camera (with a Canon EF 11-24 mm F/4 lens), and were also processed using Adobe Photoshop CS5, with brightness and contrast of the whole images adjusted. The location map and stratigraphic column of the Zhangjiagou section were produced using Adobe Illustrator CS5.
The data matrices of the phylogenetic analyses are provided in the Supplementary Data (Data matrices 1–4), and were analyzed using software TNT (Goloboff et al., Reference Goloboff, Farris and Nixon2008). For each calculation, the TNT settings remained almost the same. For example, all characters were equally weighted, the gap mode was treated as missing, and the TNT memory was enlarged to 10,000 trees. In the phylogenetic analysis of cnidarians, collapsing rule 3 (max. length=0) is adopted, consistent with the default collapsing rule of PAUP (Swofford, Reference Swofford2002), because PAUP was adopted in the original phylogenetic analysis of Dong et al. (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016) with default settings. In the phylogenetic analysis of cycloneuralians, collapsing rule 1 (min. length=0) is adopted, and this is the default collapsing rule of TNT. Traditional search (heuristic search with 1,000 random stepwise addition replicates saving up to 10 trees per replicate, followed by TBR branch swapping) was adopted. The cladograms were redrawn with Adobe Illustrator CS5.
Repositories and institutional abbreviations
The current specimens are now deposited in the collection of the University Museum of Chang’an University (UMCU), Xi’an, China.
Results
Olivooides multisulcatus
Materials assigned to Olivooides multisulcatus Qian, Reference Qian1977 are less common in our collection, and only three specimens are presented here (Fig. 4.1–4.3). They represent a progressive developmental sequence from a late embryonic stage (Fig. 4.1) to a large individual (Fig. 4.3). Olivooides multisulcatus differs from O. mirabilis (Yue in Xing et al., Reference Xing, Ding, Luo, He and Wang1984) mainly by their different sizes, with the embryos and hatched stages being much smaller than those of O. mirabilis (Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014). In addition, the hatched stages of O. multisulcatus have five rows of triangular thickenings on their post-embryonic tissues (Fig. 4.3).
Olivooides multisulcatus Qian, Reference Qian1977 (1–3) and Olivooides mirabilis (Yue in Xing et al., Reference Xing, Ding, Luo, He and Wang1984) (4, 5) from the Zhangjiagou Lagerstätte. (1) An embryo with five principal lobes and five subordinate lobes in pentaradial symmetry, UMCU.XXPT004; (2) an embryo with initially developed triangular thickenings, UMCU.2014XXPT006; (3) a hatched individual, latero-apical view, UMCU.2014XXPT008; (4) an embryo, UMCU.XXPT005, lateral view; (5) a fragment of the oral part, UMCU.2014XXPT009. Scale bar represents 500 μm (1–3), 856 μm (4), and 1.32 mm (5).
Olivooides mirabilis
Olivooides mirabilis is represented by two specimens here, including a pre-hatching embryo (Fig. 4.4) and a fragment of a large individual (Fig. 4.5). The fragment (Fig. 4.5) is part of the abapical end, with three distal annuli and one circlet of 10 terminal lobes preserved. It should be noted that the diameter of this fragment is ~3.16 mm. If the maximum diameter of the hatched stages is positively related with the total length, the completely preserved length is estimated to be ~8.6 mm, with at least 50 annuli. This is a very large individual, and it implies that the adults of O. mirabilis, and perhaps those of Olivooides multisulcatus and Quadrapyrgites quadratacris, could have reached centimeter-scale dimension with more than 50 annuli on their post-embryonic tissue.
Pseudooides prima
Only three embryos assignable to Pseudooides prima Qian, Reference Qian1977 were recovered (Fig. 5.1–5.3). The embryos have a so-called “germ-band” area made of 12 compartments separated into six pairs by a median furrow and five lateral furrows. These compartments are arranged in biradial symmetry, defining the fundamental symmetry pattern of P. prima. The remaining part of the embryos is undifferentiated. The fossils are usually internally hollow or filled with secondary mineral matter. Generally, they are slightly collapsed and, therefore, the outer surface is much wrinkled.
Embryos of Pseudooides prima Qian, Reference Qian1977 (1–3) and hatched stages of Hexaconularia sichuanensis He and Yang, Reference He and Yang1986 (4, 5) from the Zhangjiagou Lagerstätte. (1) UMCU.XXPT230; (2) UMCU.XXPT220; (3) UMCU.XXPT200; (4) UMCU.2013ZS001; (5) UMCU.2013ZS002. Scale bar represents 200 μm (1–3), 270 μm (4), and 593 μm (5).
Quadrapyrgites quadratacris
The general morphology of Quadrapyrgites quadratacris has been described by Steiner et al. (Reference Steiner, Qian, Li, Hagadorn and Zhu2014) and Liu et al. (Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a), and will not be repeated here. Liu et al. (Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a) reconstructed the post-embryonic development of Q. quadratacris and demonstrated the growth mode of the terminal lobes and annuli. The current material also exhibits a difference in the number of annuli of the post-embryonic tissue, ranging from two (Fig. 6.1, 6.2), three (Fig. 6.3), four (Fig. 6.4), five (Fig. 6.5), six (Fig. 6.6), seven (Fig. 6.7), 10 (Fig. 6.8), 11 (Fig. 6.9), 14 (Fig. 6.10), to at least 18 (Fig. 6.11). Combined with the previously reported data (Liu et al., Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a; Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014), the number of annuli on the post-embryonic tissues of Q. quadratacris extended continuously from one to at least 18, and this is interpreted to be a consecutive post-embryonic developmental sequence. It should be noted that the specimen with 18 annuli (Fig. 6.11) is exceptionally large, ~3.28 mm long and 1.2 mm maximum width. Its abapical end is broken, thus the total length and number of annuli are uncertain.
Hatched stages of Quadrapyrgites quadratacris (Li, Reference Li1984) Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014 from the Zhangjiagou Lagerstätte. (1, 2) With two annuli; (1) UMCU.2014XQ001; (2) UMCU.XQ002; (3) with three annuli, UMCU.XQ005; (4) with four annuli, UMCU.XQ006; (5) with five annuli, UMCU.XQ007; (6) with six annuli, UMCU.XQ008; (7) with seven annuli, UMCU.XQ009; (8) with 10 annuli, UMCU.2014XQ015; (9) with 11 annuli, UMCU.2014XQ016; (10) with 14 annuli, UMCU.2014XQ017; (11) with at least 18 annuli, UMCU.2014XQ018. Abbreviation a1 represents 1st annulus. Scale bar represents 500 μm.
Eopriapulites sphinx
Eopriapulites sphinx Liu and Xiao in Liu et al., Reference Liu, Xiao, Shao, Broce and Zhang2014b is a millimeter-sized animal. The adult body length/width ratio is about more than 10, thus the animal looks a little slender (Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016). A detailed description of E. sphinx was given in Liu et al. (Reference Liu, Xiao, Shao, Broce and Zhang2014b) and Shao et al. (Reference Shao, Liu, Wang, Zhang, Tang and Li2016), and so is not repeated here. The ontogeny and developmental mode of E. sphinx is currently unknown because its embryonic and younger juvenile stages are lacking.
The current specimens (Fig. 7.1, 7.4) are proposed to be trunk parts of very large juveniles or adults. They are regarded as conspecific with Eopriapulites sphinx due to co-occurrence and identical morphology of the trunk annuli (Fig. 7.2, 7.3, 7.5, 7.6). UMCU.2014XXSY018 (Fig. 7.1) is ~2.55 mm long and 550 μm wide, whereas UMCU.2014XSY016 (Fig. 7.4) has a width of 750 μm. These specimens are preserved in the Orsten-type preservation, and this type of preservation usually includes individuals no larger than 2 mm in size, and larger specimens would be fragmented during taphonomy (Maas et al., Reference Maas, Braun, Dong, Donoghue, Müller, Olempska, Repetski, Siveter, Stein and Waloszek2006). Evidently, the original animals that yielded these fragments should have a body length far more than 2 mm in size, possibly even centimeter scale.
Eopriapulites sphinx Liu and Xiao in Liu et al., Reference Liu, Xiao, Shao, Broce and Zhang2014b from the Zhangjiagou Lagerstätte. (1) A trunk part representing part of a large individual, UMCU.2014XXSY018; (2) close-up of (1), showing the dense annuli; (3) close-up of (1), showing the zig-zag-shaped structures on the annuli; (4) a trunk part of a very large individual, UMCU.2014XSY016; (5, 6) close-up views of (4), showing the dense annuli. Scale bar represents 500 μm (1, 4), 194 μm (2), 63 μm (3), 232 μm (5), and 91 μm (6).
Phylogenetic perspectives
Pseudooides prima
Pseudooides prima is reported with only embryonic stages exclusively from the Fortunian Stage of South China (Steiner et al., Reference Steiner, Zhu, Li, Qian and Erdtmann2004b; Donoghue et al., Reference Donoghue, Bengtson, Dong, Gostling, Huldtgren, Cunningham, Yin, Yue, Peng and Stampanoni2006b). The information on P. prima comes exclusively from the outer surface because no internal structures have been discovered yet. The affinity of Pseudooides has been debated for a long time (Donoghue et al., Reference Donoghue, Cunningham, Dong and Bengtson2015). Steiner et al. (Reference Steiner, Zhu, Li, Qian and Erdtmann2004b) interpreted the paired compartments as a possible “germ band”, and compared them with the “germ band” of modern arthropod embryos. Based on this assumption, Steiner et al. (Reference Steiner, Zhu, Li, Qian and Erdtmann2004b) connected the embryos of P. prima and some co-occurring arthropod or arthropod-like fragments in a consecutive ontogenetic sequence. The occurrence of arthropod fossils in the Fortunian Stage is dubious, although a number of SSFs can be interpreted as arthropod fragments (Yuan et al., Reference Yuan, Xiao, Parsley, Zhou, Chen and Hu2002). However, an arthropod affinity of P. prima is unlikely and was challenged by Donoghue et al. (Reference Donoghue, Bengtson, Dong, Gostling, Huldtgren, Cunningham, Yin, Yue, Peng and Stampanoni2006b, Reference Donoghue, Cunningham, Dong and Bengtson2015) because the anatomy of P. prima is unusual for any modern arthropods.
Here, we propose a new hypothesis, namely that the material of Pseudooides prima is actually synonymous with embryos of the taxon of co-occurring hexangulaconulariids. Hexangulaconulariids exhibit biradial symmetry, and contain two genera: Arthrochites and Hexaconularia (Conway Morris and Chen, Reference Conway Morris and Chen1992). In the Kuanchuanpu Formation of Zhangjiagou section, P. prima co-occurred with Hexaconularia sichuanensis (Fig. 5.4, 5.5). Material of H. sichuanensis is biradially symmetric (Van Iten et al., Reference Van Iten, Zhu and Li2010), while the pinched “germ band” area of P. prima is also biradially symmetric. Animals with biradial symmetry are less common in the Kuanchuanpu Formation, thus P. prima and H. sichuanensis, both with biradial symmetry, are possibly conspecific, at least being close relatives. In Olivooides and Quadrapyrgites, the embryonic tissues were retained in the hatched stages as the apical parts. Likewise, the pinched “germ-band” area of P. prima embryos might be compared with the apical area of H. sichuanensis, both of which are biradially symmetric. But evidently, there is a large morphological gap between them, and a detailed investigation of that matter, including recovery of more embryos of later embryonic stages of P. prima, is urgently demanded. Since hexangulaconulariids were regarded as an intermediate form between conulariids and Olivooides (Bengtson and Yue, Reference Bengtson and Yue1997), a similar phylogenetic position is assumed for Pseudooides, and Pseudooides might also be a coronate scyphozoan.
Olivooides and Quadrapyrgites
Olivooides and Quadrapyrgites have exclusively been reported from the Cambrian Fortunian Stage of South China (Donoghue et al., Reference Donoghue, Cunningham, Dong and Bengtson2015). Due to comparable morphology and ontogenetic sequence, Steiner et al. (Reference Steiner, Qian, Li, Hagadorn and Zhu2014) proposed that Olivooides and Quadrapyrgites should be sister groups, constituting the taxon Olivooidae Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014, and might be early cycloneuralians.
As radiate animals, the affinity of olivooids falls better within Cnidaria (but see Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014; see discussion below). Under the cnidarian hypothesis, the preserved tubes of olivooids are interpreted as periderm (exoskeleton) that completely embraces the internal soft part anatomy. Among modern cnidarians, only coronate scyphozoans have this type of periderm, thus Olivooides was regarded as a coronate scyphozoan when it was interpreted as a fossil embryo for the first time (Bengtson and Yue, Reference Bengtson and Yue1997; Yue and Bengtson, Reference Yue and Bengtson1999). As the sister group of Olivooides, the taxon Quadrapyrgites should have the same affinity. The coronate scyphozoan hypothesis has been cited by other researchers and now takes a central place for the affinity of Olivooides and Quadrapyrgites (Bengtson and Yue, Reference Bengtson and Yue1997; Dong et al., Reference Dong, Cunningham, Bengtson, Thomas, Liu, Stampanoni and Donoghue2013, Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016; Liu et al., Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a).
The coronate scyphozoan hypothesis is largely dependent on similarity between olivooids and the periderm of modern coronate scyphozoans, because no convincing internal soft-part anatomy of olivooids has been reported. In addition to the possible periderm shared between olivooids and modern coronate scyphozoans, Dong et al. (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016) proposed a second feature, the periderm teeth, as a possible uniting character. Dong et al. (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016) regarded the “inner walls” within the embryos as possible ectoderm, and most importantly, the “paired pentaradial projections” as periderm teeth. In their phylogenetic analysis, Olivooides and Quadrapyrgites are resolved as sister groups of coronate scyphozoans, and the monophyly of conulariids, Coronatae, Olivooides and Quadrapyrgites is well supported by two synapomorphies (the presence of an all-embracing periderm and the presence of periderm teeth). However, the periderm teeth interpretation might be dubious. According to Jarms (Reference Jarms1991), the periderm teeth of coronate scyphozoans should be located on the inner side of the periderm. If this is applied to the specimens of Olivooides, the periderm teeth should be located on the inner side of the stellate tissues or on the inner side of the striated tissues. But in Dong et al.’s (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016) interpretations, the periderm teeth are located on the inner side of the “inner walls” (possible ectoderm according to their interpretation), while the periderm, on which the periderm teeth should be located, is not preserved at all. Furthermore, as part of the periderm, the periderm teeth should have the same fossilization potential, but in most cases, the inner side of the tubes is smooth without any signs of the development of periderm teeth. Therefore, the periderm teeth interpretation for the “internal pentaradial projections” is dubious.
In order to test the phylogenetic positions of Olivooides and Quadrapyrgites, we carried out a phylogenetic analysis. The selected taxa and characters follow Dong et al. (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016). First, we re-analyzed Dong et al.'s dataset (Datamatrix 1), and got 10 most parsimonious trees (MPTs), Tree Length (TL)=127, Consistency Index (CI)=0.740, Retention Index (RI)=0.738. The strict consensus tree presented (Fig. 8.1) is consistent with Dong et al.’s result (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016, their fig. 10A). In order to test the bearing of the presence of periderm teeth (the 88th character) upon the topology of the phylogenetic tree, we deleted the coding of the 88th character, and re-analyzed the revised datamatrix (Datamatrix 2). We got 40 MPTs, TL=126, CI=0.738, RI=0.734. The strict consensus tree is presented (Fig. 8.2). The monophyly of Olivooides, Quadrapyrgites, Conulariids, and Coronatae is still supported, but the internal relationships are collapsed.
Suggested phylogenetic positions of Olivooides and Quadrapyrgites. (1) Strict consensus tree derived from analysis of the original dataset of Dong et al. (Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016); (2) strict consensus tree derived from analysis excluding character 88 (presence or absence of periderm teeth). Symbol † in this figure and in Figure 9 denotes extinct taxa.
Without the inclusion of the presence of periderm teeth, the internal relationships of Olivooides, Quadrapyrgites, conulariids, and Coronatae are not resolved in the current study (Fig. 8.2). However, the monophyly of Olivooides, Quadrapyrgites, conulariids, and Coronatae is still supported by one character—the presence of an all-embracing periderm. The interpretation of the periderm teeth is highly doubted, and should not be coded in the phylogenetic analysis currently.
Eopriapulites sphinx
The cuticular scalids of modern scalidophorans are internally hollow up to the tip and contain sensory cells that lead to a hole distally, whereas the hooks of nematoids are solid (Schmidt-Rhaesa, Reference Schmidt-Rhaesa1998). Modern scalidophorans have, plesiomorphically, both longitudinal and circular muscles in their subepidermal muscle tube. Contraction of the circular muscles during locomotion causes annular wrinkling of their cuticle, thus scalidophorans could be annulated. In contrast, nematoids lack circular muscles, retaining only longitudinal muscles, and they evolved internal cuticular longitudinal thickenings. These two features together result in a different mode of locomotion, a kind of wriggling in zig-zag shape. The trunk of nematoids is usually long and thin. However, annulation may occur in certain nematodes like in Desmoscolex frontalis (Decraemer, Reference Decraemer1986, his fig. 5B). Eopriapulites sphinx has a body consisting of internally hollow introvert scalids, collar scalids, and an annulated trunk, suggesting a systematic position within the Scalidophora. However, E. sphinx lacks characters that allow it to be assigned to any in-group Scalidophora, and this taxon is interpreted as a stem-lineage derivative of Scalidophora, an assignment that is also supported by phylogenetic analyses (Liu et al., Reference Liu, Xiao, Shao, Broce and Zhang2014b; Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016).
The exact phylogenetic assignment of Eopriapulites is founded on the assumption that the ground pattern of each node in the phylogenetic tree of Cycloneuralia is well resolved. But unfortunately, there are many conflicts and uncertainties about the ground pattern characters of Cycloneuralia and its in-groups, Nematoida, and Scalidophora. For example, it is uncertain whether the stem species (= last common ancestor) of Cycloneuralia has internally hollow or solid introvert scalids/hooks. The occurrence of both longitudinal and circular muscles in the body wall of Scalidophora is a plesiomorphic feature retained from the last common ancestor of Cycloneuralia/Nemathelminthes and Bilateria (Ax, Reference Ax2003; Nielsen, Reference Nielsen2012). Absence of circular muscles in modern nematoids is a secondary loss, hence an autapomorphy of Nematoida. Therefore, assignment of Eopriapulites to the Scalidophora based on internally hollow scalids and an annulated trunk might be incorrect because these two characters may be simply plesiomorphic states, referring to the stem species of Cycloneuralia. Eopriapulites would still belong to Cycloneuralia, but assignment to an in-group requires knowledge about more morphologically significant details.
In order to further test the phylogenetic position of Eopriapulites, we carried out a number of phylogenetic analyses. The coded characters followed those used by Zhang et al. (Reference Zhang, Xiao, Liu, Yuan, Wan, Muscente, Shao, Gong and Cao2015), and the data matrices are provided in the Supplementary Data (Datamatrices 3 and 4). The palaeoscolecid species, Palaeoscolex piscatorum Whittard, Reference Whittard1953, Chalazoscolex pharkus Conway Morris and Peel, Reference Conway Morris and Peel2010, Xystoscolex boreogyrus Conway Morris and Peel, Reference Conway Morris and Peel2010, and Guanduscolex minor Hu et al., Reference Hu, Li, Luo, Fu, You, Pang, Liu and Steiner2008, are replaced with Palaeoscolecida sensu stricto (Harvey et al., Reference Harvey, Dong and Donoghue2010). Markuelia hunanensis is not included in the analysis. Dong et al. (Reference Dong, Donoghue, Cheng and Liu2004) proposed that M. hunanensis should be a direct developer without larval stages, and under such condition M. hunanensis can be included in the phylogenetic analysis because the adults and the pre-hatching embryos have similar morphology. However, further development of M. hunanensis is unknown because we lack information on its post-embryonic, free-living stages (Haug et al., Reference Haug, Maas, Waloszek, Donoghue and Bengtson2009).
Parsimony analysis (Datamatrix 3) including extant cycloneuralians plus E. sphinx and Palaeoscolecida sensu stricto yielded 429 MPTs (TL=249, CI=0.683, RI=0.898). The 50% majority rule consensus tree (Fig. 9.1) resolves E. sphinx and Palaeoscolecida sensu stricto as close relatives of extant Nematoida. Parsimony analysis (Datamatrix 4) with large sampling including extant and extinct cycloneuralians yielded 1080 MPTs (TL=338, CI=0.55, RI=0.854). The 50% majority rule consensus tree (Fig. 9.2) resolves Palaeoscolecida sensu stricto, together with some cycloneuralians from Burgess Shale-type lagerstätten, as close relatives of extant Nematoida, while E. sphinx and some other early cycloneuralians are resolved within crown-group Priapulida.
Suggested phylogenetic positions of Eopriapulites and Palaeoscolecida sensu stricto. (1) 50% majority rule consensus tree derived from analysis including extant cycloneuralians plus Eopriapulites sphinx and Palaeoscolecida sensu stricto; (2) 50% majority rule consensus tree derived from analysis including extant and extinct cycloneuralians.
From previous and the current analyses, it is evident that the phylogenetic position of Eopriapulites is heavily influenced by the taxa included in the analyses. Eopriapulites can be resolved as stem-lineage derivative of Scalidophora (Liu et al., Reference Liu, Xiao, Shao, Broce and Zhang2014b; Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016), stem-lineage derivative of Nematoida, or even crown-group Priapulida, thus its position is heavily unstable. Based on current knowledge of the ground pattern of Scalidophora, Nematoida, and Cycloneuralia, Eopriapulites has an uncertain position and should be assigned to total-group Cycloneuralia.
Discussion
Debate about the affinity of olivooids
Though traditionally accepted as possible coronate scyphozoans, many other hypotheses have been proposed for Olivooides and/or Quadrapyrgites, including the echinoderm hypothesis (Chen, Reference Chen2004), cubozoan cnidarian hypothesis (Han et al., Reference Han, Kubota, Li, Ou, Wang, Yao, Shu, Li, Uesugi, Hoshino, Sasaki, Kano, Sato and Komiya2016a, Reference Han, Li, Kubota, Ou, Toshino, Wang, Yang, Uesugi, Masato, Sasaki, Kano, Sato and Komiyab), stem-lineage cycloneuralian hypothesis (Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014), and stem lineage of Diploblastica hypothesis (Yasui et al., Reference Yasui, Reimer, Liu, Yao, Kubo, Shu and Li2013). The echinoderm hypothesis has been criticized by many authors (e.g., Dong et al., Reference Dong, Cunningham, Bengtson, Thomas, Liu, Stampanoni and Donoghue2013, Reference Dong, Vargas, Cunningham, Zhang, Liu, Chen, Liu, Bengtson and Donoghue2016; Liu et al., Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a) and can be effectively rejected. The discussion here focuses on the stem-lineage cycloneuralian hypothesis in light of the oldest known cycloneuralians recovered from the Fortunian Stage in recent years, and on the cubozoan hypothesis that competes with the scyphozoan hypothesis within a broader cnidarian interpretation.
Bengtson and Yue (Reference Bengtson and Yue1997) originally noted the similarity between Punctatus tubes and the loricate larvae of Priapulida. Steiner et al. (Reference Steiner, Qian, Li, Hagadorn and Zhu2014) compared the olivooid tubes with the pre-loricate larvae of modern priapulids. Modern pre-loricate larvae of priapulids do not feed and lack an anus, but an anus is developed immediately after the second loricate larval stage (Wennberg et al., Reference Wennberg, Janssen and Budd2009). The known material assigned to olivooids lacks an anus throughout the whole ontogeny (Yasui et al., Reference Yasui, Reimer, Liu, Yao, Kubo, Shu and Li2013; Steiner et al., Reference Steiner, Qian, Li, Hagadorn and Zhu2014). Annuli are added one by one during elongation of the tubes (Liu et al., Reference Liu, Li, Shao, Zhang, Wang and Qiao2014a). Based on the Zhangjiagou material (Figs. 4.5, 6.11), the adults of olivooids are estimated to have reached centimeter scale with more than 50 annuli. Accordingly, if olivooids actually represented cycloneuralians, the absence of an anus would imply growth to 50 annuli requiring at least 49 separate molts without any food intake. The physiological energetics of molting requires corresponding nutrition that does not support a non-feeding larvae hypothesis. Olivooids are interpreted here as being radially symmetric, contrasting completely with the bilateral symmetry of cycloneuralians. The oldest known cycloneuralians, exemplified by Eopriapulites and Eokinorhynchus (Zhang et al., Reference Zhang, Xiao, Liu, Yuan, Wan, Muscente, Shao, Gong and Cao2015; Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016), along with the younger Shergoldana (Maas et al., Reference Maas, Waloszek, Haug and Müller2007) had already developed an introvert armored with scalids. Olivooids lack not only an anus, but also the specific characteristics of cycloneuralians, such as an introvert with scalids. Therefore, the stem-lineage cycloneuralian hypothesis is not supported here.
Han et al. (Reference Han, Kubota, Li, Yao, Yang, Shu, Li, Kinoshita, Sasaki, Komiya and Yan2013, Reference Han, Kubota, Li, Ou, Wang, Yao, Shu, Li, Uesugi, Hoshino, Sasaki, Kano, Sato and Komiya2016a, Reference Han, Li, Kubota, Ou, Toshino, Wang, Yang, Uesugi, Masato, Sasaki, Kano, Sato and Komiyab) regarded some fossil embryos from the Kuanchuanpu Formation at the Shizhonggou section as cubozoans. This hypothesis was based on plausible internal biological structures, such as possible tentacles, frenula, oral marginal lappets, and gastric saccule-like humps, which might be more compatible with those of modern cubomedusozoans. Among these fossil embryos belonging to species of Olivooides and Quadrapyrgites, Han et al. (Reference Han, Li, Kubota, Ou, Toshino, Wang, Yang, Uesugi, Masato, Sasaki, Kano, Sato and Komiya2016b) recognized a series of the same internal structures of embryos of Olivooides multisulcatus and described them following the terminology of modern cubomedusozoans. This methodology implies that the embryos are direct developers, and they would hatch directly into juvenile cubomedusozoans without passing a planular phase and a polypoid phase. Following this hypothesis, the hatched stages of Olivooides and Quadrapyrgites have to be regarded as stalked medusoid stages, with the periderm embracing the internal cubomedusozoans completely. The current material from the Zhangjiagou section implies that the stalked stages are continuously developed, extending from small hatchlings with only one annulus to large individuals with ∼50 annuli. Medusoid stages in a stalked lifestyle occur only in modern staurozoans, which are interpreted as the sister group of all other Medusozoa (Miranda et al., Reference Miranda, Hirano, Mills, Falconer, Fenwick, Marques and Collins2016). Accordingly, the cubozoan hypothesis demands that the stalked medusoid forms occurred independently twice—once in early cubozoans, such as olivooids, and once in modern staurozoans. It also demands that the all-embracing periderm occurred independently twice—once in early cubozoans, such as olivooids, and once in modern coronate scyphozoans. This is a rather un-parsimonious evolutionary scenario for early cnidarian evolution, thus is not favored here.
Eopriapulites as an ancestral cycloneuralian
It has previously been proposed that the ancestral cycloneuralians were macroscopic, “priapulid-like,” introvert-bearing animals. Accordingly, macroscopic palaeoscolecids from the Burgess Shale-type lagerstätten were proposed to represent ancestral cycloneuralians or even ancestral ecdysozoans (Budd, Reference Budd2001, Reference Budd2003b; Harvey et al., Reference Harvey, Dong and Donoghue2010). The palaeoscolecids range from the early Cambrian to late Silurian, ca. 520–420 Ma (Harvey et al., Reference Harvey, Dong and Donoghue2010). A potential ancestor should have occurred at least no later than other members of its lineage. But based on the current data, at least three fossil cycloneuralians have occurred earlier than the earliest palaeoscolecids: Markuelia secunda Val’kov in Val’kov and Karlova, Reference Val’kov and Karlova1984 from the Pestrotsvet Formation of Siberia (ca. 521–525 Ma); Eopriapulites sphinx and Eokinorhynchus rarus from the lower Fortunian Stage of South China (ca. 535 Ma). The occurrence of these cycloneuralians implies that the cycloneuralians should have been rooted in the Fortunian Stage, or possibly even earlier, and that the ancestral cycloneuralian/ecdysozoan hypothesis of palaeoscolecids is challenged.
Eopriapulites and Markuelia occurred in the Terreneuvian Series, not long after eumetazoan origin and diversification, and earlier than the first occurrence of palaeoscolecids (Cambrian Series 2). Again, Eopriapulites and Markuelia, with their similar and seemingly plesiomorphic morphology, have been proposed to be early scalidophorans (Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016) or in-group scalidophorans (Dong et al., Reference Dong, Bengtson, Gostling, Cunningham, Harvey, Kouchinsky, Val’kov, Repetski, Stampanoni, Marone and Donoghue2010). However, the current analysis indicates that at least Eopriapulites might also be a close relative of Nematoida (Fig. 9.1). The uncertain phylogenetic assignments with Scalidophora or Nematoida imply that Eopriapulites has characters that are shared with both Scalidophora and Nematoida (i.e., with the stem of Cycloneuralia). This is very similar to the younger Shergoldana australiensis, which also has been assigned to total group Cycloneuralia. The specific character combination of the only known but exceptionally 3D-preserved specimen of this species, has both nematoid and scalidophoran characters, implying an assignment outside any in-group. An assignment to Nematoida or Scalidophora can even be ruled out based on its well-known morphology (Maas et al., Reference Maas, Waloszek, Haug and Müller2007).
A new hypothesis proposed here is that Cycloneuralia might have originated in the Fortunian small shelly faunas rather than in the early Cambrian macrobenthos, implying that the ancestral cycloneuralians should have been microscopic, vermiform, introvert-bearing, and have characters more like Eopriapulites sphinx. Previous studies have shown that Markuelia might be more basal than Eopriapulites (Shao et al., Reference Shao, Liu, Wang, Zhang, Tang and Li2016), implying that pentaradially symmetric arrangements of introvert scalids occurred earlier than hexaradially symmetric forms. Therefore, the last common ancestor of Cycloneuralia might have possessed an introvert with internally hollow and pentaradially arranged introvert scalids. Internally hollow and pentaradially arranged introvert scalids may have been inherited by the last common ancestor of Scalidophora and possibly also by the lineage leading to modern Nematoida, therefore might have been lost autapomorphically early in this lineage.
Conclusions
The Fortunian Zhangjiagou Lagerstätte has yielded three-dimensionally phosphatized microfossils of radiate animals and cycloneuralians. Radiate animals include embryos of Olivooides multisulcatus, Olivooides mirabilis, and Pseudooides prima, as well as putative hatched stages of O. multisulcatus, O. mirabilis, Hexaconularia sichuanensis, and Quadrapyrgites quadratacris. These radiate animals represent the diversification of cnidarians in the Fortunian Stage. Cycloneuralians are represented by Eopriapulites sphinx and trunk fragments possibly related with Eokinorhynchus rarus (Zhang et al., Reference Zhang, Xiao, Liu, Yuan, Wan, Muscente, Shao, Gong and Cao2015, their unnamed forms I and II).
These exceptionally preserved microfossils provide important information on the early diversification of cnidarians and cycloneuralians. Based on these fossils, we propose (1) cnidarians have a high diversity in the Fortunian Stage of South China, and symmetry patterns include biradial (Pseudooides and Hexaconularia), triradial (Anabarites), tetraradial (Quadrapyrgites), and pentaradial (Olivooides and Qinscyphus Liu, Shao, and Zhang in Liu et al., Reference Liu, Shao, Zhang, Wang, Zhang, Chen, Liang and Xue2017) symmetry; (2) P. prima might be the embryonic stage of the co-occurring H. sichuanensis, with the biradial symmetry as a possible uniting feature; (3) the adults of Olivooides and Quadrapyrgites might have had a length exceeding 1 cm with more than 50 annuli on the post-embryonic surface, and the unusually large specimens of Olivooides and Quadrapyrgites can refute the non-feeding larvae hypothesis for Olivooides and Quadrapyrgites; (4) because of the lack of convincing internal soft-part anatomy, Olivooides and Quadrapyrgites might better be assigned to coronate scyphozoans, with the all-embracing periderm as a key uniting feature; and (5) the cycloneuralians, and possibly the ecdysozoans or nemathelminths, might have originated in the Fortunian small shelly faunas as part of the meiofauna rather than in the macrobenthos, with their ancestors being similar in morphology with Eopriapulites sphinx. As a consequence, ancestral cycloneuralians might have possessed pentaradially arranged and internally hollow introvert scalids. Such scalids may have been inherited by the last common ancestors of Scalidophora and Nematoida, used originally exclusively or mainly for locomotion, but were lost early in the crown-group of the Nematoida.
The final important point is that Cambrian cycloneuralians cannot all be treated simply as “priapulids” as pointed out by Maas et al. (Reference Maas, Waloszek, Haug and Müller2007) and Maas (Reference Maas2013) because such an assignment is not justified by any characters. Again, following Maas (Reference Maas2013), we emphasize that the morphology of Cambrian cycloneuralians, including Eopriapulites sphinx, does not give any hint that Cycloneuralia is the sister taxon of Panarthropoda, as indicated by the Ecdysozoa hypothesis.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (41572007, 41572009), the State Key Laboratory of Palaeobiology and Stratigraphy, Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences (173121), the Youth Innovation Promotion Association, Chinese Academy of Sciences (2016283), Quality Project of Chang’an University (0012-310600161000, 0012-310627171808), College Students’ Innovative Entrepreneurial Training Program of Chang’ an University (201710710062, 201710710063, 201710710240, 0012-310600161000, 0012-310627171808) and The Tenth “Challenge Cup” Competition of Chang’ an University (C-P-B-2, C-P-B-6, C-P-B-8). Two anonymous referees provided careful revisions and constructive suggestions to this paper. Jisuo Jin and Brock Glenn provided careful technical edits and language polishing. Correspondence should be addressed to HZ (hqzhang@nigpas.ac.cn) or YL (stotto@163.com).
Accessibility of Supplementary Data
Data available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.1cn6b
