Tentacular nature of the ‘column’ of the Cambrian diploblastic Xianguangia sinica

Abstract

Unveiling the body architectures of Cambrian problematic fossils would provide novel insights into the radiation of metazoan body plans during the ‘Cambrian Explosion’ and the ancestral traits of major living animal clades. Xianguangia sinica, from the celebrated Chengjiang biota (∼518 Ma), is a typical Cambrian problematicum with disputable body architecture, particularly about its ‘column’ part. The contradictory interpretations of the ‘column’ of X. sinica have led to at least three hypotheses regarding its affinity in the diploblastic clade. Here we depict the detailed anatomy of the ‘column’ based on new, exquisitely preserved material. The ‘column’ of X. sinica is formed by 18 longsword-shaped tentacle-sheath complexes that can either close or be in a flowering state. There is no partitioned cavity internally when the ‘column’ is closed, invalidating the homology with the true column of living sea anemones. Each tentacle tapers distally and includes a distal flexible portion at about one-fourth the length of the tentacle. The proximal portion is stiff, bearing a set of paired dark stains, and is enveloped by a single outer sheath. Pinnules carrying a row of large cilia are fringed on both sides along the whole length of the tentacles. The body plan of X. sinica is accordingly revised as consisting of a calyx and 18 unique tentacle-sheath complexes that radially surround the mouth. Our result corroborates previous observations that suggest a close relationship between Xianguangia, Daihua and Dinomischus, all of which are here formally assigned to the family Dinomischidae, a monophyletic clade recovered in our phylogenetic analyses. Xianguangia sinica likely employs cilia-bearing pinnate tentacles for sieving particle matter down to 21 µm, indicating that its nutrition source is suspended micro-planktonic organisms or other organic matter in the bottom water.

Keywords:

• early Cambrian
• dinomischiids
• body plan
• suspension-feeding
• Chengjiang biota

Introduction

Understanding the body plan of problematic fossils is a process of gradual improvement, constantly seeking new evidence with repeated verification (Cong, 2023). Xianguangia sinica is one of the problematic fossils exemplifying this process, known only from the early Cambrian Chengjiang biota (ca. 518 Ma) (Hou et al., 2017). The reconstructed body of X. sinica was initially divided into three parts: a basal holdfast, a middle ‘column’, and 16 distal soft tentacles (Chen & Erdtmann, 1991). It has been variably identified as a crown cnidarian (Chen & Erdtmann, 1991), ctenophore (Hou et al., 1999), lophophorate (Luo et al., 1999) or Ediacaran survivor (Han et al., 2010). Among these, the cnidarian hypothesis is the most popular, mainly gaining support from its radially symmetrical polypoid shape. This superficial similarity has led to a direct comparison between X. sinica and living anemone-like polyps. The longitudinal ridges observed in the ‘column’ of X. sinica were accordingly interpreted as impressions of mesenteries, thus hinting at the presence of a partitioned blind digestive tract (Chen & Erdtmann, 1991). Recently, an examination of 88 specimens revealed 18 feather-like tentacles protruding from the ‘column’ in X. sinica, which bore ciliated pinnules, and a novel phylogenetic analysis recovered Xianguangia as a stem-group Cnidaria (Ou et al., 2017).

Intriguingly, pinnules bearing large cilia were also identified in a newly named polypoid animal, Daihua sanqiong, from the Chengjiang biota (Zhao et al., 2019), of which the arrangement and dimensions are reminiscent of the ciliated pinnules observed in the tentacles of X. sinica. One well-preserved specimen of D. sanqiong in oral view further illustrated that pinnules occur along the entire length of the tentacle, consisting of a proximal sclerotized rod and a distal flexible portion (Zhao et al., 2019). Based on the distribution of pinnules, the structural scheme of the tentacle of D. sanqiong was inferred to be potentially applicable in X. sinica, i.e. the middle ‘column’ and the distal soft ‘tentacle’ of X. sinica should be reinterpreted as the same body unit (Zhao et al., 2019). In this new interpretation model, X. sinica lost the foundation for the cnidarian hypothesis and was recovered as a stem-group Ctenophora in an updated phylogenetic analysis (Zhao et al., 2019).

These two interpretations of the ‘column’ part of X. sinica prompt two contradictory evolutionary scenarios in the diploblastic grade, i.e. whether it is a stem cnidarian (Ou et al., 2017, 2022) or a stem ctenophore (Klug et al., 2021; Zhao et al., 2019). The central point for this dilemma is the biological nature of the ‘column’, which needs to be clarified before researchers can reconstruct the phylogenetic position of X. sinica. Here, we re-describe the architecture of the ‘column’ based on exquisitely preserved fossil specimens of X. sinica. Our findings provide crucial evidence to revise the body plan of X. sinica and recover its phylogenetic position.

Material and methods

In total 28 specimens, including 18 newly collected (prefix: YNGIP), were examined, all of which are deposited at Yunnan University, Kunming, China. Fossils that are not fully exposed were prepared with a fine needle under a stereomicroscope (Nikon SMZ 1000). Specimens were imaged under cross-polarized light at both high and low angles with a Canon EOS 5DS R digital camera mounted with macro lenses, either Canon EF 100 mm or Canon MP-E 65 mm (1–5X). Selected photos were imported into Adobe Photoshop CC 20.0.0 to adjust levels and brightness. Interpretative drawings were made from specimens and photos. Schematic drawings were drafted in Adobe Illustrator CC 23.0.1.

Our phylogenetic dataset is adopted from a previous study (Zhao et al., 2019), with several characters excluded, rescored or added. This updated dataset contains 278 characters and 86 taxa, including 44 living diploblastic taxa. Bayesian inference under the mkv + gamma model (Lewis, 2001) was conducted in MrBayes 3.2.7 (Ronquist et al., 2012), with 10,000,000 generations and with an option to stop automatically when the average deviation of split frequencies was < 0.01.

Terminology

Several terms used in describing the morphology of Xianguangia sinica are confusing. Initially, the body of X. sinica was divided into three parts and termed ‘pedal disc’ for the proximal part, ‘column’ for the middle cylindric region and ‘tentacle’ for the distal outgrowth, respectively (Chen & Erdtmann, 1991). The ‘pedal disc’ was subsequently renamed the ‘holdfast’ with a basal disc and a basal pit (Ou et al., 2017). Recently, the term ‘holdfast’ was changed to ‘calyx’, while the ‘column’ and ‘tentacle’ were defined as proximal and distal parts of the same body unit, respectively (Zhao et al., 2019). In the present study we adopt most of the terminology from the most recent study (Zhao et al., 2019). For some uncertain or controversial structures, we recommend neutral descriptive term to avoid direct, subjective interpretation. As such, the terms ‘cushion plate’ and ‘nerve’ in Zhao et al. (2019) are changed to ‘dark stain’ and ‘dark line’ instead, respectively. The ‘column’ used in this study has the same definition as in the previous studies, referring to the problematic middle part of the body.

Institutional abbreviations

NIGPAS, Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences, Nanjing, China; YKLP and YNGIP, the invertebrate palaeontological collections of Yunnan University, Kunming, China.

Systematic palaeontology

Phylum, Class, Order Uncertain

Family Dinomischidae Conway Morris, 1977

Amended diagnosis (amended after Conway Morris, 1977, p. 834). Polypoid, sessile organism bearing a calyx encircled by tentacle-sheath complexes, with an organic skeleton. The basal portion of the calyx varies from a circular pit to a slender stalk. The tentacle carries sets of pinnules, each with a row of cilia, and preserves a set of paired dark stains. The outer sheath flanking each tentacle present.

Remarks

Only one genus, Dinomischus, was assigned to the family Dinomischidae in previous studies (Chen et al., 1989; Conway Morris 1977; Peng et al., 2006), and although the genus Siphusauctum was considered closely related to Dinomischus, a separate family, Siphusauctidae, was erected for it (O'Brien & Caron, 2012). Dinomischus has long been interpreted as an entoproct comprising a calyx and sclerotized bracts (Chen et al., 1989; Conway Morris, 1977). A recent study revealed a potential blind digestive tract and a partitioned cavity in some well-preserved Dinomischus specimens, indicating that it is less likely to be a bilaterian entoproct which possesses a typical U-shaped gut (Zhao et al., 2019). Paired dark stains and pinnules identified recently in the bract of Dinomischus are similar to that of Daihua, and the body plan of Dinomischus has thus been revised as comprising a calyx and pinnate, sclerotized tentacles (Zhao et al., 2019). The present study further reveals that similar dark stains, as well as ciliated pinnules, appear in the longsword-shaped tentacles of Xianguangia. The tentacle in Siphusauctum also bears paired sets of dark stains, but lacks pinnules and is non-sclerotized. The outer sheath surrounding each tentacle has been identified in Xianguangia, Daihua, Dinomischus and Siphusauctum as well, but it likely varies in number. It is reasonable to formally assign the genera Xianguangia and Daihua, together with Dinomischus, to the family Dinomischidae, a monophyletic group recovered in our Bayesian phylogenetic analyses. The 18 pinnate, sclerotized tentacles, plus a calyx with an organic skeleton, could be recognized as synapomorphies for this family. Although our phylogenetic results support the previous hypothesis that dinomischiids are a total-group ctenophore (Klug et al., 2021; Zhao et al., 2019), alternative phylogenetic analyses with different scores on the characters ‘cushion rows’ and ‘large cilia’ recover dinomischiids as either a stem-group ctenophore or in a polytomy with ctenophores and cnidarians. The phylum-level affinity of dinomischiids is indeterminate, and new fossil taxa with anatomical transformations would provide critical information to verify its position in the tree of life.

Genera included

Dinomischus Conway Morris, 1977; Xianguangia Chen & Erdtmann, 1991; Daihua Zhao et al., 2019.

Occurrence

Cambrian unnamed Series 2, Stage 3, Chengjiang biota, South China (Xianguangia, Daihua, Dinomischus); Miaolingian Series, Wuliuan Stage, Burgess Shale, Canada, and Kaili biota, South China (Dinomischus); Miaolingian Series, Barrandian area, Czech Republic (cp. Dinomischus).

Genus Xianguangia Chen & Erdtmann, 1991

Type species

Xianguangia sinica Chen & Erdtmann, 1991

Diagnosis

As for the type and only species.

Remarks

The different architecture of the calyx separates Xianguangia from Daihua and Dinomischus (Table 1). Xianguangia only possesses an invaginated pit in the basal portion of the calyx, whereas the latter two taxa have either a blunt tip or a slender stalk, respectively (Zhao et al., 2019). The oral surface expands to various degrees among these three taxa, forming three oral domes with slender bracts in Daihua, and an elongated oral cone in Dinomischus (Zhao et al., 2019). Xianguangia shares with Daihua and Dinomischus an outer sheath enveloping each tentacle rod. A single sheath has been identified in Xianguangia and Dinomischus, while two possible paired sheaths have been suggested in Daihua (Zhao et al., 2019).

Table 1. Morphological differences in three genera of Dinomischidae.

Genus	Calyx		18 sclerotized tentacles with ciliated pinnules
	Basal calyx	Oral surface	Outer sheath	Distal tentacle
Xianguangia	Basal pit	No expansion	Single	Present
Daihua	Tapering blunt tip	Three circumoral domes with paired bracts	Two pairs	Present
Dinomischus	Slender stalk	Oral cone	Single	Absent

Xianguangia sinica Chen & Erdtmann, 1991

(Figs 1–3)

Figure 1. Xianguangia sinica with tentacle-sheath complexes closed. A, B, laterally preserved specimen: A, YKLP 13478a, part; B, YKLP 13478b, counterpart. C, a close-up image showing the outer sheaths, tentacle rods and pinnules. D, interpretative drawing of C. E, YKLP 13829a, the preservation of the whole body suggests light sclerotization. F, a fluorescent image of the distal tentacle region, showing pinnules fringed on the distal tentacles. Abbreviations: Cal, calyx; Cc, circumferential constriction; Ds, dark stain; Dt, distal tentacle; Os, outer sheath; Pin, pinnule; Ter, tentacle rod; T1–T8, tentacle-sheath complex 1 to 8. Scale bars: A, B, E = 5 mm; C, F = 1 mm.

Figure 2. Xianguangia sinica with tentacle-sheath complexes splayed out. A, YKLP 13496a, part, showing a displaced calyx and tentacle-sheath complexes. B, YKLP 13496b, counterpart, preserved with discrete tentacle-sheath complexes. C, a merged image of part and counterpart, showing a complete morphology of the body. D, YKLP 13477a, part, a specimen in oral-aboral view showing a total of 18 tentacle-sheath complexes. E, close-up of the tentacle-sheath complexes T10–T11, showing the smooth outer sheaths and relief tentacle rods with rows of paired dark stains. F, a fluorescent image showing the details of dark stains along the tentacles. G, interpretative drawing of F. Abbreviations: Cal, calyx; Ds, dark stain; Fu, furrow; T1–T18, tentacle-sheath complex 1 to 18. Scale bars: A–D = 5 mm; E–G = 2 mm.

Figure 3. Xianguangia sinica with dark patches. A–C, YNGIP 90001, A, a complete specimen in lateral view showing a pair of dark patches preserved in the lower central region of the ‘column’; B, a close-up image of the dark patch region; C, a fluorescent image showing the patches likely originated from the oral surface. D, YNGIP 90002a, a close-up image showing three dark patches. E, YNGIP 90003, a close-up of the dark patch region, showing two lateral patches inclined towards the middle one. Abbreviations: Ca, cavity; Cal, calyx; Dp, dark patch; Ds, dark stain; Mo, mouth; Pin, pinnule. Scale bars: A, D, E = 2 mm; B, C = 1 mm.

*1991 Xianguangia sinica Chen & Erdtmann; Chen & Erdtmann: 64–65, 73–74, pl. 3, figs 1–4.

1996 Xianguangia sinica Chen & Erdtmann; Chen, Zhou, Zhu & Yeh: 93–94, figs 96–99.

1997 Xianguangia sinica Chen & Erdtmann; Chen & Zhou: 30, fig. 24.

v.1999 Xianguangia sinica Chen & Erdtmann; Hou, Bergström, Wang, Feng & Chen: 50–51, figs 54–56.

1999 Xianguangia sinica Chen & Erdtmann; Luo, Hu, Chen, Zhang & Tao: 86, pl. 22, fig. 1.

2002 Xianguangia sinica Chen & Erdtmann; Chen, Luo, Hu, Yin, Jiang, Wu, Li & Chen: 8, 39, pl. 21, fig. 4.

v.2003 Xianguangia sinica Chen & Erdtmann; Hou & Bergström: 62, 64, fig. 8(2).

2004 Xianguangia sinica Chen & Erdtmann; Chen: 164–165, figs 246–247.

v.2004 Xianguangia sinica Chen & Erdtmann; Hou, Aldridge, Bergström, Siveter, Siveter & Feng: 54–55, fig. 9.1.

2014 Xianguangia sinica Chen & Erdtmann; Lei, Han, Ou & Wan: 968–970, fig. 3.

2017 Xianguangia sinica Chen & Erdtmann; Ou, Han, Zhang, Shu, Sun & Mayer: 8835–8839, figs 1–3, S1–S3.

v.2017 Xianguangia sinica Chen & Erdtmann; Hou, Siveter, Siveter, Aldridge, Cong, Gabbott, Ma, Purnell & Williams: 80–81, fig. 11.4.

v.2019 Xianguangia sinica Chen & Erdtmann; Zhao, Vinther, Parry, Wei, Green, Pisani, Hou, Edgecombe & Cong: 1113–1114, fig. 1.

Holotype

NIGPAS 108506 (Chen & Erdtmann, 1991, pp. 73–74).

Figured material

YKLP 13477 (Fig. 2D), YKLP 13478 (Fig. 1A–D), YKLP 13496 (Fig. 2A–C, E–G), YKLP 13829 (Fig. 1E, F), YNGIP 90001 (Fig. 3A–C), YNGIP 90002 (Fig. 3D), YNGIP 90003 (Fig. 3E).

Other material

YKLP 13429, YKLP 13473, YKLP 13474, YKLP 13475, YKLP 13476, YKLP 13830, YNGIP 90004 to YNGIP 90018.

Amended diagnosis (amended after Ou et al., 2017, pp. 8835–8836). Solitary, polypoid metazoan, with a calyx and a whorl of 18 tentacle-sheath complexes surrounding an oral region. Calyx with a digestive tract inside. Underside of the calyx smooth and convex with a prominent, circular pit. Putative oral region with a circumferential constriction, separating the tentacle-sheath complex from the calyx. Tentacle comprising a proximal tentacle rod and a distal flexible portion. Tentacle rod lightly sclerotized, bearing serially arrayed paired dark stains and enveloped by a single sheath externally. Both proximal and distal parts of the tentacle biserially flanked by flexible, ciliated pinnules. Cilia large, long and thick, densely arrayed along either side of each pinnule.

Locality

Yu’anshan Member of the Chiungchussu Formation, lower Cambrian (Series 2, Stage 3, ca. 518 Ma), corresponding to the Eoredlichia–Wutingaspis trilobite Biozone (Babcock & Zhang, 2001). Fossil specimens were collected from the classic Maotianshan section, and the sites included Ercai, Jianshan, Mafang and Erjie, southern Kunming, Yunnan Province, China.

Description

Gross morphology

Generally, most known specimens of Xianguangia sinica are laterally preserved, displaying a well-defined, distally tapering ‘column’ in the middle part of the body (Fig. 1A, B, E). A circumferential constriction (‘Cc’, Fig. 1E) delimits the ‘column’ and the calyx. A few longitudinal ridges (‘T1–T8’, Fig. 1A, E) in the ‘column’ are nearly parallel to one another.

Tentacle-sheath complex

One laterally preserved specimen displays a Xianguangia-type calyx that is cylindrical in shape (ca. 19.2 mm height and 20.9 mm wide) and a displaced ‘column’ part (Fig. 2A–C). Although the ‘column’ is not directly connected with the calyx as it was normally preserved, it shares the same structural composition and gross shape with the undetached ‘column’. The displaced ‘column’ part preserved 12 longsword-shaped structures (‘T1–T12’, Fig. 2C) that are splayed out around the top of the calyx (Fig. 2A–C). Each longsword-shaped structure has a clear, constant edge, further indicating that it is a separate unit. These units are ca. 29.5 mm long and a maximum of 3.4 mm wide (at the basal end). The same longsword-shaped structures, which are delimited by longitudinal deep folds that are formed by the tight contact between adjacent ones, can also be identified in these classical specimens preserved in lateral view (‘T1–T8’, Fig. 1A, B, E). These longsword-shaped structures are here termed tentacle-sheath complexes. A total of 18 tentacle-sheath complexes can be identified in a specimen compacted along the oral-aboral axis (Fig. 2D).

Two furrows with slight relief (‘Fu’, Figs 1E, 2D, E) that converge distally can be discerned on the tentacle-sheath complex under low-angle light. Occasionally, a shallow longitudinal groove extends along the central section of the tentacle-sheath complex (Fig. 2E). These observations together indicate that the complex was likely a three-dimensional wedge-shaped projection when alive. The sheath (‘Os’, Fig. 1A–C) is smooth on the surface and fits closely together with tentacles due to compression. Each sheath is wider than the tentacle (Fig. 1A–C), and likely surrounds a tentacle that was preserved either as slight relief (Fig. 2E) or as red-brown or brown-yellow in colour (Fig. 1A–C). The tentacle tapers distally to form a flexible portion at 65–75% of the whole tentacle length (herein termed distal tentacle, ‘Dt’, Fig. 1A, E, F), which is arrayed and spread at the end of the sheaths, while the proximal portion (herein named tentacle rod, ‘Ter’, Fig. 1B–D) of the tentacle is stiff and robust, suggesting light sclerotization.

Dark stains

Paired dark stains are serially arrayed along the tentacle rod, and distinctly identified in the distal region of T10 and T11 with teardrop-shaped relief under low-angle light (‘Ds’, Fig. 2E, F). The size of the stains is variable (up to 0.35 mm wide and 0.83 mm long in T11) and gradually decreases distally. A median dark line extends longitudinally between paired dark stains (Fig. 2F, G), but no branches connecting with the stains are identified.

When the tentacle-sheath complexes are closed to form a tapering ‘column’, dark stains are discernable as well (Figs 1A, B, E, 3A). In YKLP 13478, the inner side of the tentacle rod was exposed when the specimen was split into part and counterpart, showing about six to eight well-defined rows of paired stains longitudinally extending to the base of the ‘column’ (Fig. 1A, B). Dark stains appear to be present only in the tentacle rod (Fig. 1A, E, F). These stains have a teardrop-like shape generally with an obvious dark outline, but also vary in size.

Pinnules

Pinnules (‘Pin’, Figs 1, 3A) are fringed serially on the distal tentacle (Fig. 1E, F) and the proximal tentacle rod (Fig. 1A–D). They were preserved as irregular dark patches, sometimes fading slightly, probably due to decay and/or weathering. In YKLP 13478, each pinnule can reach ca. 1.8 mm long and 56 μm wide. The space between adjacent pinnules is ca. 71–123 μm. Overall, pinnules project to the upper right corner when they appear in the left half part and vice versa (Figs 1A–D, 3A), suggesting that they subtend to the oral-aboral axis when alive.

Dark patches

Finger-like dark patches are frequently preserved in the lower middle region of the ‘column’ (‘Dp’, Fig. 3). This structure differs from the surrounding region with its definite shape with dark colour and sometimes seems to extend downwards to the oral surface of the calyx (Fig. 3A–C), with the colour fading or turning to red-brown. The patches contain a few longitudinal lines in the most cases (Fig. 3B–E) and are occasionally infilled with sediment (Fig. 3E). The number of patches varies from two to three. When present as three, the two lateral patches sometimes incline towards the middle patch that is near vertical (Fig. 3D, E).

Remarks

The rough surface of the column-like body part of Xianguangia sinica appears to have parallel, ‘longitudinal ridges’ observed in laterally preserved specimens, which were generally regarded as evidence for the presence of mesenteries (Chen & Erdtmann, 1991; Ou et al., 2017). The present study illustrates that the ‘longitudinal ridges’ are actually discrete tentacle-sheath complexes, which have the ability to splay out freely. The assumed presence of mesenteries and digestive tract inside the ‘column’ is accordingly rejected. Instead, these structures are more likely to appear in the calyx (Zhao et al., 2019). The organic material preserved in the lower central part of the ‘column’ (Ou et al., 2017) is likely caused by the decay of the three dark patches, which likely derive from the oral surface. Paired dark stains are serially arrayed along the inner side of each tentacle rod, which is less likely to be a taphonomic artefact because of their regular arrangement and constant topological position across specimens. Given that dark stains and ciliated pinnules can be obscured by the tentacle-sheath complexes when specimens were preserved in lateral view, they have not been readily discerned in previous material of X. sinica. The preserved morphology of the stains and the pinnules is depicted here when tentacle-sheath complexes were decayed away or removed by splitting. It is therefore inappropriate to erect the present material as a new species because it has the same body composition and gross morphology as X. sinica.

Phylogenetic results

Our newly updated phylogenetic analysis resolves Xianguangia and Daihua as sister groups, which together with Dinomischus form a monophyletic clade that was previously named the family Dinomischidae (see Systematic palaeontology above). Our result supports a previously proposed scenario that dinomischiids and Siphusauctum form a paraphyletic grade along the stem ctenophores (Fig. 4).

Figure 4. Bayesian phylogenetic result. Phylogenetic position of Xianguangia recovered from Bayesian inference using mkv + gamma model on a dataset (containing 278 characters and 86 taxa) modified from Zhao et al. (2019). Xianguangia is resolved as the sister group to Daihua, which along with Dinomischus forms a monophyletic grade on the ctenophore stem. Numbers at the nodes are posterior probabilities, and the scale bar shows the expected number of substitutions per site. The fossil taxa are indicated with the dagger symbol. Also see Supplemental Material Figure S1A for the full result.

Discussion

The architecture of the ‘column’ and reconstruction

Our study provides novel evidence to revise the architecture of the ‘column’ of Xianguangia sinica, validating the composition of 18 tentacle rods (Fig. 2A–D), each of which is terminated by a soft, flexible portion (distal tentacle). Paired dark stains are regularly arranged along the inner side of each tentacle rod, as mentioned in a previous study (Ou et al., 2022). Our new material further confirms that sets of pinnules appear in both the proximal stiff portion and the distal flexible portion of the tentacles (Fig. 1). The evidence for pinnules carrying a row of large cilia has only been previously identified in a few specimens (Ou et al., 2017; Zhao et al., 2019). Considering the rarity of fossil specimens that were preserved with exquisite cilia, we infer that all pinnules carried large cilia when alive, resembling the reconstruction of Daihua sanqiong (Zhao et al., 2019).

We identified the presence of an outer sheath in the tentacles of X. sinica, which generally envelops the proximal tentacle rod and terminates at around one-fourth the length of the tentacle (Fig 1A, E). Each tentacle and its outer sheath are integrated into a structure named the tentacle-sheath complex. A set of tentacle-sheath complexes, 18 in number (Fig. 2D), forms a tapering column-like appearance when these complexes are closed, or a blooming flower-like appearance when they are splayed out. Therefore, the longitudinal ridges seen in the ‘column’ are the impression of the sclerotized tentacle rod and its associated outer sheath, instead of the putative mesenteries interpreted in the cnidarian hypothesis (Chen & Erdtmann, 1991; Ou et al., 2017). In light of these, the body plan of X. sinica is formally revised to bear two body parts, a calyx (formerly ‘holdfast’) and 18 tentacle-sheath complexes (formerly ‘column’ + ‘tentacle’) (Fig. 5A, B).

Figure 5. Body plan reconstruction. A, schematic drawing of the tentacle-sheath complex of Xianguangia. From left to right are shown the inside view, cross section and outside view. The tentacle consists of a proximal tentacle rod enveloped by a single outer sheath and a distal flexible portion. Ciliated pinnules emerge from tentacles laterally. B, lateral view of the whole-body shape with closed tentacle-sheath complexes.

The proposed ‘cavity’ inside the ‘column’

In some specimens of Xianguangia sinica there is a dark region preserved in the lower middle region of the ‘column’, which was interpreted as the remains of organic material, indicating the presence of a gastric cavity (Ou et al., 2017, fig. 3A, B therein). However, considering the tentacular nature of the ‘column’, only a tentacular chamber would be possibly present when the tentacle-sheath complexes are closed. This indicates that the digestive tract cannot be present in the tentacle-enclosed chamber, as it would have no place to reside when the tentacle-sheath complexes are splayed out. It is very unlikely that there is a separate, unknown body part within the tentacle-enclosed chamber to accommodate the digestive tract. In the living cnidarian polyps, the true columnar part is the primary organ housing a digestive tract, of which the wall is delimited by the inner mesenteries (Brusca et al., 2016). The cavity and associated digestive tract are more likely accommodated in the calyx of X. sinica. This is corroborated by observations that a spacious cavity (‘Ca’, Fig. 3A), infilled with sediment, has been identified in the calyx of the adult specimens (Ou et al., 2017), and a partitioned cavity is present in some potential juvenile individuals (Zhao et al., 2019).

Our new specimens show at least three finger-like dark patches regularly occurring in the lower centre of the ‘column’ (i.e. the closed tentacle-sheath complexes). We suggest that the dark patches would be preserved as dark remains when the specimens experienced heavy decay and/or weathering. In terms of preservation and topological position, these dark patches are consistent with the dark organic material recognized in Ou et al. (2017), indicating they are the same structures. The consistent presence at the same position across multiple specimens indicates that the dark patches are real anatomical structures, rather than reflecting the random preservation of organic material, which probably rose from the oral surface of X. sinica. The position and number of these dark patches in X. sinica nullify their interpretation as a digestive cavity. Instead they are more comparable with the oral domes in Daihua sanqiong, which emerge from the oral surface and surround a central opening (Zhao et al., 2019). It is plausible that the oral dome could be extended upwards once the tentacle-sheath complexes closed. However, considering the fact that the domes on the oral surface are equally spaced in D. sanqiong, when the specimens are laterally compressed, the domes should be distributed regularly along the whole width of the oral surface instead of being concentrated in the central region. The centralized distribution of the dark patches in X. sinica might due to the relatively small size. Unlike the sclerotized oral dome in D. sanqiong, the patches in X. sinica appear to be unsclerotized and lack the pair of bracts that protrude from each oral dome and extend beyond the main body (Zhao et al., 2019). In Dinomischus venustus and Siphusauctum gregarium, only an oral cone is formed by the expansion of the oral surface (Zhao et al., 2019), which often extends to the distal part of the body and tapers to a narrow shape with a single opening (O'Brien & Caron 2012; Zhao et al., 2019). In contrast, X. sinica possesses about three separate dark patches that are restricted to the proximal region only, a condition similar to that of D. sanqiong.

Given that the dark patches are placed in the proximal region of the tentacle-sheath complexes and possibly close to the central mouth in the calyx, they are likely food sacs functioning to store food particles. A less plausible alternative is that they are remains of muscle blocks that control the closure of tentacle-sheath complexes, but no evidence supports this view currently.

Comparison

Our study confirms that Xianguangia possesses 18 feather-like, sclerotized tentacles and reveals that its longsword-shaped tentacle rod carries a row of paired dark stains and is enveloped by a single outer sheath (Fig. 5A, B). Xianguangia possesses unique tentacles disparate from most co-occurring sessile, solitary fossil taxa that have either smooth, soft tentacles like Cotyledion (Zhang et al., 2013) or branched, non-circumoral tentacles, such as Phlogites (Hou et al., 2006) and Herpetogaster (Caron et al., 2010). A putative hemichordate, Galeaplumosus, composed of a contractile stalk, a putative tube and possibly two branched arms (Hou et al., 2011) has been suggested to be a fragment of the ‘tentacle’ and the ‘column’ of Xianguangia (Ou et al., 2017). Based on our reconstruction of the body plan of Xianguangia (Fig. 5A, B), if the tube and stalk of Galeaplumosus were a fragment of the tentacle-sheath complex (‘column’, sensu Ou et al., 2017), this structure should taper towards the distal end rather than towards the proximal end. In addition, one tentacle rod would only correspond to one pinnate distal tentacle, in contrast to the two branched arms of Galeaplumosus. For these reasons, and given the absence of comparable paired dark stains in the tube and the stalk, Galeaplumosus should still be treated as a valid taxon.

Such tentacle-sheath complexes as in Xianguangia have only been noted, to our knowledge, in Daihua and Dinomischus (Zhao et al., 2019), across fossil and living taxa. Xianguangia shares striking similarities with them not only in the body architecture (a calyx with or without stalk and 18 tentacle-sheath complexes with ciliated pinnules), but in several aspects of the detailed morphology of the tentacle, such as structural composition (pinnules, dark stains), general shape (longsword-shaped, possibly light sclerotization) and topological position (surrounding the mouth), which forms the basis for the monophyly of these three taxa, within the family Dinomischidae (Fig. 4). The major morphological differences between Xianguangia, Dinomischus and Daihua occur in their basal calyx and oral surface (Table 1).

Evaluation of alternative hypotheses on the homologies of key features

Paired dark stains are arrayed along the inner side of tentacles, most notably along the tentacle rods. These structures have been interpreted as ‘cushion plates’ in Daihua and Dinomischus previously, likely beneath the ciliated pinnules (Zhao et al., 2019). However, whether these structures are homologous to the cushion plate of living ctenophores, which serves as a cytoskeletal polster of ctene consisting of fused macrocilia (Tamm, 2014), is controversial (Ou et al., 2022). One option to verify this interpretation is to find the macrocilia that appear on the dark stains directly in dinomischiids (Jakob Vinther, personal communication, September 2021). However, evidence for macrocilia emerging from the putative cushion plate has also never been found in well-preserved ctenophore fossil taxa (Conway Morris & Collins, 1996; Ou et al., 2015; Parry et al., 2021), indicating that the potential for the preservation of macrocilia on the dark stains is very low at best. When accepting the alternative interpretation (Ou et al., 2022) and recoding the ‘cushion plates’ as absent in Xianguangia, Daihua, Dinomischus and Siphusauctum, the monophyly of Dinomischiidae and its stem ctenophore affinity are still recovered in our phylogenetic analysis (Fig. 6A).

Figure 6. Summary of additional phylogenetic results. Additional Bayesian phylogenetic analyses under different homologous interpretations. A, the character ‘cushion rows’ is rescored as absent in dinomischiids and Siphusauctum; Xianguangia is recovered as a stem-group ctenophore. B, the character ‘large cilia’ is coded as only present in dinomischiids and Siphusauctum; Xianguangia is recovered as a stem-group ctenophore. C, the characters ‘cushion rows’ and ‘large cilia’ are rescored together; Xianguangia, Daihua and Dinomischus form a monophyletic grade, which is placed in a polytomy with cnidarians and a clade consisting of Siphusauctum and ctenophores. Numbers at the nodes are posterior probabilities. The fossil taxa are indicated with the dagger symbol. Also see Supplemental Material Figures S1B, S2 for the full results.

It is worth noting that the dark stains on the lobe-like tentacles of Siphusauctum carry large cilia directly (O'Brien & Caron, 2012), conforming in topological composition to the cushion plates seen in living ctenophores. The rows of broad, paired dark stains in Siphusauctum are morphologically similar to the putative ctene rows in the co-occurring scleroctenophores (O'Brien & Caron, 2012; Zhao et al., 2019). With the absence of pinnules and the development of dark stains on the tentacles, Siphusauctum seems to be in a transitional stage between dinomischiids and scleroctenophores. When accepting this potential homologous relationship and scoring the cushion plate as being present in Siphusauctum, plus adding the character ‘oral surface extension’ (hypothesizing that the main body region of ctenophores was derived from a progressive expansion of the oral surface) in the matrix of Ou et al. (2022) (including 126 characters of 42 taxa), Dinomischus and Siphusauctum are recovered as stem-group ctenophores, but Xianguangia and Daihua are resolved as sister groups, in a polytomy with ctenophores and cnidarians (Fig. 7B), not supporting the claim that Xianguangia is a stem cnidarian (Fig. 7A).

Figure 7. Summary of additional phylogenetic results based on an alternative view. Bayesian analyses on a dataset from Ou et al. (2022) (with 126 characters and 42 taxa) under mkv + gamma model. A, an original result showing that Xianguangia is recovered as a stem-group cnidarian. B, Siphusauctum and the associated character ‘oral surface extension’ are added into the original matrix, and character 47, ‘separate body cavities in polyps’, is deleted. Xianguangia is resolved as a sister group to Daihua, which is placed in a polytomy with cnidarians and a clade containing Dinomischus, Siphusauctum and ctenophores. Numbers at the nodes are posterior probabilities. The fossil taxa are indicated with the dagger symbol. Also see Supplemental Material Figure S3 for the full results.

Likewise, given that the unique cilia serially borne on the pinnule in Xianguangia and Daihua, of which the large size (up to 600 μm long) is incomparable with that in any living animal group except ctenophores, a previous study suggested that the large cilia in dinomischiids are potentially homologous to the compound cilia in living ctenophores (Zhao et al., 2019). However, the cilia have different roles in these two groups, which are used as a feeding structure in dinomischiids but as swimming ctene in living ctenophores. This hypothesis of homology is contested by a recent study which claims that taphonomic processes could increase the thickness of the cilia (Ou et al., 2022). When adopting this alternative view and coding the ‘large cilia’ as only present in dinomischiids and Siphusauctum, they still form a paraphyletic grade on the ctenophore stem group (Fig. 6B). Nevertheless, when the characters ‘cushion plates’ and ‘large cilia’ are both rescored, dinomischiids are resolved in a polytomy with ctenophores and cnidarians (Fig. 6C).

Suspension-feeding mode of life

The tentacle equipped with ciliated pinnules represents a unique feeding apparatus. The long, dense and evenly spaced cilia borne on pinnules indicate a suspension-feeding mode of life (Ou et al., 2017). Suspension feeding is one of the most important ecological strategies during the ‘Cambrian Explosion’, widespread in radiodonts (Lerosey-Aubril & Pates, 2018; Vinther et al., 2014), lobopodians (Caron & Aria, 2017; Yang et al., 2015), lophophorates (Zhang et al., 2009), and hemichordates (Nanglu et al., 2016). However, none of these has a comparable feeding apparatus to the tentacle of Xianguangia as well as Daihua and Siphusauctum. This suggests a high diversity and functional disparity of suspension-feeding apparatus already appeared in different metazoan groups during the ‘Cambrian Explosion’, likely corresponding to the quickly diversified ecological niches during this evolutionary radiation event.

Living suspension-feeding invertebrates display about 11 types of particle capture mechanisms (Riisgård & Larsen, 2010). The stiff cilia are arranged subparallel to one another in specimens of Xianguangia, Daihua and Siphusauctum (Ou et al., 2017; Zhao et al., 2019), with a comb-like shape. They are evenly spaced, ca. 21 µm apart, in the pinnules of Xianguangia and Daihua, and ca. 36 µm apart in the tentacles of Siphusauctum (measured from Zhao et al., 2019, figs 1D, 2D, 4C therein). Considering the arrangement (dense, comb-like array) and morphology (straight), ciliated pinnules are likely used for sieving suspended particles, although the actual particle types may be complex. When only considering ciliary sieving, the minimum particle size should be equal to the spaced distance between adjacent cilia (Temereva & Malakhov, 2010). Therefore, suspended particles with a minimum diameter of 21 µm could be captured from the water column by Xianguangia and Daihua, and 36 µm in Siphusauctum. Alternatively, they may retain suspended particles smaller than the gaps between the cilia by employing some complex mechanisms, such as adhesion. These lower limits lie within the size distribution of microplankton, corresponding to suspended particles of nano- and micro-phytoplankton and -zooplankton (Stramski et al., 2004) (Fig. 8). The appearance of dinomischiid-like suspension feeders likely coincides with the sharp increase of plankton diversity in the early Cambrian (Vidal & Knoll, 1982), such as the unicellular phytoplankton (Butterfield, 1997).

Figure 8. Estimation of suspension-feeding particle size. The distance between adjacent cilia is c. 21 µm in Xianguangia and Daihua (black dotted line and box) and 36 µm in Siphusauctum (purple dotted line and circle), which corresponds to the range of the size of micro-plankton, indicating that Xianguangia, Daihua and Siphusauctum might feed on micro-planktonic organisms, such as micro-phytoplankton. The diagram on the left depicts the relationship between mesh size and food particle size, modified after Vinther et al. (2014, fig. 4 therein). The distribution of the particle size on the right is adopted from Stramski et al. (2004, fig. 1 therein). The hollow box and circle on the solid line terminated with arrows indicate the actual minimum size measured from Xianguangia, Daihua and Siphusauctum fossil specimens, respectively. The solid box and circle are the estimated minimum and maximum particle size mapping from the left diagram.

A similar inference could also be made when assuming Vinther et al.'s (2014) methodology is also applicable to ciliary sieving in Xianguangia, Daihua and Siphusauctum (Fig. 8), because they share a comparable sieved architecture with those suspension-feeding radiodonts, including stiff cilia/auxiliary spines/setae in a comb-like array (Lerosey-Aubril & Pates, 2018; Vinther et al., 2014). In this case, the suspension-feeding apparatus in benthic dinomischiids and Siphusauctum is optimized for ciliary sieving on smaller micro-planktonic organisms living in the water column and the bottom water, comparable with the filter system with a minimum mesh size of 70 µm in these nektonic radiodonts (Lerosey-Aubril & Pates, 2018).

Conclusions

The architecture of the ‘column’ is revised based on well-preserved specimens of Xianguangia sinica (Cambrian Epoch 2, Age 3, Chengjiang biota, ca. 518 Ma), which is formed by 18 tentacle-sheath complexes that can either close in a column manner or splay out in a blooming pattern. There is no partitioned cavity present within the tentacle-enclosed ‘column’, which is thus not homologous with the true columnar part of living sea anemones. Each tentacle is enveloped by a single outer sheath at the main proximal portion (the tentacle rod) and gradually tapers to a distal flexible portion (the distal tentacle). Paired dark stains are serially arranged along the inner side of each tentacle rod but their presence in the distal tentacle awaits more evidence. Ciliated pinnules are borne on both sides of the proximal and distal parts of each tentacle and are probably employed as a micro-plankton suspension-feeding apparatus.

Xianguangia shares similarities with Daihua and Dinomischus in bearing a common body plan consisting of a calyx and 18 unique tentacle-sheath complexes and possessing an organic skeleton. The genera Xianguangia and Daihua are thus formally grouped into the family Dinomischidae that is recovered as a monophyletic clade in our phylogenetic results. Dinomischidae is a group of diploblastic animals, nested in the basal branch of the total group of Ctenophora and subtending to the clade of Siphusauctum and the derived clade of Cambrian putative ctenophores, such as scleroctenophores, in our phylogenetic trees.

Associate Editor: Ken Johnson

Acknowledgements

We thank Jakob Vinther (University of Bristol), Luke A. Parry (University College London) and Qun Yang (Nanjing Institute of Geology and Palaeontology, CAS) for their insightful discussions and suggestions. We also thank two referees for their constructive suggestions that improved this manuscript significantly. This work was supported by the National Natural Science Foundation of China (42202023, 42072019), and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB26000000). YZ was supported by grants from the Jiangsu Funding Program for Excellent Postdoctoral Talent (2022ZB457) and the UZH Postdoc Grant (FK-22-129). PC especially thanks the continuous support from the Yunling Scholar and a joint project of YNU-Yunnan Government (202201BF070001-016). Open access funding was provided by the Universität Zürich.

Supplemental material

Supplemental material for this article can be accessed here: https://doi.org/10.1080/14772019.2023.2215787.

Additional information

Funding

We thank Jakob Vinther (University of Bristol), Luke A. Parry (University College London) and Qun Yang (Nanjing Institute of Geology and Palaeontology, CAS) for their insightful discussions and suggestions. We also thank two referees for their constructive suggestions that improved this manuscript significantly. This work was supported by the National Natural Science Foundation of China (42202023, 42072019), and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB26000000). YZ was supported by grants from the Jiangsu Funding Program for Excellent Postdoctoral Talent (2022ZB457) and the UZH Postdoc Grant (FK-22-129). PC especially thanks the continuous support from the Yunling Scholar and a joint project of YNU-Yunnan Government (202201BF070001-016). Open access funding was provided by the Universität Zürich.