Comparisons with other flap-bearing panarthropods

The species most morphologically similar to Mobulavermis adustus n. gen n. sp. is Kerygmachela kierkegaardi (Fig. 4.3–4.7), with both taxa bearing broad ventrolateral swimming flaps, a roughly cylindrical body, and an elongate caudal spine devoid of flaps (Park et al., Reference Park, Kihm, Woo, Park, Lee, Smith, Harper, Young, Nielsen and Vinther2018). The similarity of the two genera allows for inferences on the anterior morphology of Mobulavermis, such as the very likely presence of sessile, ventrally situated compound eyes, a forward-facing mouth, and large raptorial frontal appendages. These inferred features, in addition to the broad flaps and the absence of walking limbs, suggest that like Kerygmachela, Mobulavermis was primarily or exclusively pelagic, although more discussion of its ecology awaits the discovery of specimens preserving the anterior morphology. Mobulavermis differs from Kerygmachela in a few aspects, however, such as Mobulavermis bearing many more pairs of flaps, which are densely spaced and more broadly shaped. In addition, the flaps of Mobulavermis gradually reduce in size until they become essentially flush with the caudal spine, which is much shorter and possibly more flexible than that of Kerygmachela. Compared with Utahnax, the swimming flaps of Mobulavermis appear much more derived—Utahnax may represent a kind of transitional form between the cylindrical limbs of siberiids and the expanded flaps of later kerygmachelids. The anterior regions of the flaps in Utahnax show evidence of evaginations of the body cavity (shown as darker staining of the surface) extending deep into the anterior of the flap, similar to the condition in the lobopodous limbs of siberiids and Opabinia regalis Walcott, Reference Walcott1912 (Lerosey-Aubril and Ortega-Hernández, Reference Lerosey-Aubril and Ortega-Hernández2022). In addition, the anterior region of the flaps in Utahnax preserve some three-dimensionality, suggesting that the anterior region may have been thicker than the posterior region (they also merge onto the same plane before the merge with the body, suggesting this topographic change is not evidence of a separate, hidden structure), which is essentially flat and unornamented. Thus, it could be proposed that in the Kerygmachelidae, ventrolateral flaps were developed not by uniform dorsoventral compression of the lobopod limbs, but by the extension of a flattened flap-like posterior margin, alongside the more gradual flattening of the lobopodous limb as a whole, such that the surface would be essentially flat in later genera (Kerygmachela and Mobulavermis show no obvious signs of a differentiated anterior region or any indication of the extent of the body cavity into the flaps).

The description of Mobulavermis lends further credibility to the hypothesis that the flaps of Kerygmachela and Utahnax are unrelated to the dorsal flaps of Pambdelurion, Opabiniidae, and a number of peytoiid radiodonts (Lerosey-Aubril and Ortega-Hernández, Reference Lerosey-Aubril and Ortega-Hernández2022). Studies on preserved musculature in panarthropods from the Sirius Passet of Greenland have suggested that musculature was absent in the flaps of Pambdelurion but present in the lobopodous ventral limbs (Young and Vinther, Reference Young and Vinther2017). This casts doubt on the swimming capabilities of Pambdelurion and, by extension, the Opabiniidae, which share homologous dorsal flaps and the controversial ventral lobopods (Budd and Daley, Reference Budd and Daley2012); recent discoveries, including the ventral lobopods of Mieridduryn bonniae Pates et al., Reference Pates, Wolfe, Lerosey-Aubril, Daley and Ortega-Hernández2022b, and small, enigmatic structures possibly attributable to lobopodous legs in the holotype of Utaurora comosa Pates et al., Reference Pates, Botting, Muir and Wolfe2022a, suggest that the lobopods of Opabinia are genuine. That the dorsal flaps are homologous to the dorsal flaps of Pambdelurion also suggests Opabinia may have moved primarily by walking on its lobopod legs rather than as an active swimmer, as previously believed. However, the same study (Young and Vinther, Reference Young and Vinther2017) on the musculature of Pambdelurion also failed to identify flap musculature in Kerygmachela (the sole specimen cited in the study as belonging to Kerygmachela more closely resembles Pambdelurion judging from its broad, anteroposteriorly tapering trunk, lack of dorsal trunk nodes, large and anteriorly widened pharynx, and wide gap between the bases of the frontal appendages).

There is at least one specimen of Kerygmachela, MGUH 25.053, which undoubtedly preserves muscle tissue, attributable to Kerygmachela by the large trunk tubercules and the mostly cylindrical trunk. In this specimen, elongate bundles of muscle fibers are found outside the boundaries of the trunk, tapering outwardly (Budd, Reference Budd1999). For each metameric segment of the body, two of these bundles are developed on either side of the body, with one pair associated with the dorsal tubercules and the other offset behind, set between successive sets of dorsal tubercules. Some of these bundles are slightly curved posteriorly. Their exact origin on the body is undetermined, but they appear to originate laterally from the body. Comparison of these muscle fiber bundles with the elongate high-relief structures, supposedly lobopodous ventral limbs in the holotype and paratype of Kerygmachela, shows that some of these limb-like structures are associated with the dorsal tubercules (holotype, Fig 4.3, 4.4), while in other specimens (paratype, Fig. 4.6), they are set between the dorsal tubercule rows. This would make little sense if they were interpreted as lobopodous limbs, both their inconsistent position and the apparently doubled number of limbs per segment. Triangular limb-like structures with positive relief are also found in MGUH 32.059 (Park et al., Reference Park, Kihm, Woo, Park, Lee, Smith, Harper, Young, Nielsen and Vinther2018, supp. fig. 8f–h), where they are associated with the dorsal tubercules, oriented perpendicular to the body, and apparently more elongate than those of the holotype and paratype of Kerygmachela.

In the complete specimen MGUH 32.048, comparable structures are seen in relief (although they are particularly shallow in this specimen) and under high-angle polarized light. In this specimen, both positional modes of these structures can be identified, associated with the dorsal tubercules and offset between them; these “non-neural impressions” (Park et al., Reference Park, Kihm, Woo, Park, Lee, Smith, Harper, Young, Nielsen and Vinther2018) extend far into the pleural region, beyond the extent of the supposed diminutive lobopod limbs (Budd, Reference Budd1993, Reference Budd1999). Non-neural impressions are further identified in other specimens under high-angle polarized lighting, showing reflective patches that connect to the trunk and extend a great distance into the pleural region (Park et al., Reference Park, Kihm, Woo, Park, Lee, Smith, Harper, Young, Nielsen and Vinther2018, supp. fig. 6). While the suggestion that one or all of these high-relief triangular structures in the holotype and paratype of Kerygmachela represent lobopodous ventral limbs, the deep pleural extension of the “non-neural impressions,” which appear to correspond to the supposed “lobopod limbs,” suggests that all of these structures may be better interpreted as muscular extensions into the ventrolateral flaps rather than distinct ventral legs. The interpretation that these two muscle-fiber bundles power a single flap rather than a pair of structures on either side of the body aligns with the known intrinsic limb musculature of Pambdelurion, which consists of a distinct promotor and remotor muscle which taper anteriorly, extending distally from the trunk (Young and Vinther, Reference Young and Vinther2017, fig. 9). Considering the limbs of Pambdelurion are homologous to either the flaps or the proposed ventral lobopods of Kerygmachela, a similar musculature should be expected, and available material suggests that these muscles belonged to the flaps. Restudy of the type material of Kerygmachela also elucidated further detail on the structures originally interpreted as setal-blade gills. While Lerosey-Aubril and Ortega-Hernández (Reference Lerosey-Aubril and Ortega-Hernández2022) described these striations as confined mostly to the distal parts of the body flaps, the holotype and paratype material show no such preference (Fig. 4.3, 4.4, 4.6). The striations, or wrinkles, are present down the entire length of the flap, ending at the trunk, directed mostly anteroposteriorly rather than posterodistally. These wrinkles overlying the paired limb-like structures (which are here interpreted as internal flap musculature) may be responsible for the appearance of “annulations” on limbs (labeled “striated limb tip” in Budd, Reference Budd1999). These wrinkles are unlikely to be related to, or analogous to, the strengthening rays of radiodonts, which must have evolved later. Their exact function is unknown, but they perhaps served a similar respiratory role to the setal-blade gills of radiodonts.

Mobulavermis can be differentiated from the radiodonts by its nearly cylindrical, extremely elongate trunk (judging from the portion of the trunk that is preserved), contrasted with the strongly posteriorly or anteroposteriorly tapering, rather compact bodies of radiodonts (Potin and Daley, Reference Potin and Daley2023); by the absence of any associated sclerotized appendages, carapace elements, or mouthparts; and by the absence of any ornamentation on its relatively well-preserved flaps (some peytoiids with dorsal flaps such as Aegirocassis Van Roy et al., Reference Van Roy, Daley and Briggs2015, have trunks that are blocky in cross section and only weakly tapering posteriorly, but this condition is very dissimilar to the present fossils and always accompanied by obvious setal blades). Radiodonts exhibit a diverse array of flap morphologies, which can be separated into a few different archetypes. Anomalocaridid and amplectobeluid radiodonts have elongate, mostly lobate flaps divided into a clear anterior zone, which is robust and strengthened by vein-like strengthening rays running anterodistally from the median of the flap to the anterior margin, and a posterior zone, which is mostly smooth and featureless (Zeng et al., Reference Zeng, Zhao and Zhu2023). In Ramskoeldia platyacantha Cong et al., Reference Cong, Edgecombe, Daley, Guo, Pates and Hou2018 and a number of isolated body flaps from the Emu Bay Shale (Daley et al., Reference Daley, Paterson, Edgecombe, García-Bellido and Jago2013), the flaps are broad and roughly triangular, with a posterior margin essentially perpendicular to the trunk axis, with a wide anterior zone ornamented with strongly anterodistally directed strengthening rays. This occasionally leads to their confusion with the body flaps of peytoiids (Zhu et al., Reference Zhu, Lerosey-Aubril and Ortega-Hernández2021). In peytoiids, flaps are generally short, either subcircular or triangular, and may be paired with dorsal flaps (Van Roy et al., Reference Van Roy, Daley and Briggs2015). The ventrolateral flaps of peytoiids are uniformly ornamented with strengthening rays that run from the base of the flap to the outer margin; in at least Peytoia Walcott, Reference Walcott1911a, there is also a posterior zone where the strengthening rays are absent or weak. The strengthening rays of peytoiids are well defined, while no such structures are observed in the fossils of Mobulavermis, nor do they match the morphology of anomalocaridid/amplectobeluid flaps with their clearly divided anterior and posterior zones. In addition, the condition of an elongate, unpaired caudal spine lacking an associated tail fan is not known in any known radiodont.

Alternative interpretation of Parvibellus

The recently discovered lobopodian Parvibellus avatus Liu et al., Reference Liu, Dunlop, Steiner and Shu2022 from the Chengjiang Fauna of Yunnan, China, was initially interpreted as a lower stem group arthropod bearing at least 11 pairs of ventrolaterally protruding, slightly overlapping trunk flaps (Fig. 4.8) described as similar to the dorsal flaps of other Cambrian lobopodian genera such as Opabinia and unlike the clawed walking legs of lobopodians such as Aysheaia. The single known specimen of Parvibellus (total length 5.25 mm) also preserved a distinct, rounded head region with a large, ventrally projecting circular mouth lined with sclerotized tooth plates, no evidence supporting the presence of eyes, and two small, unsclerotized, antennae-like frontal appendages. The trunk of the specimen is wide and preserves a simple gut running from the mouth to between the terminal pair of trunk outgrowths, described as diminutive furcae. The authors excluded the possibility of the specimen representing a juvenile or larval radiodont by comparisons with known juvenile specimens of Lyrarapax unguispinus Cong et al., Reference Cong, Ma, Hou, Edgecombe and Strausfeld2014, which are virtually identical to adult specimens (Liu et al., Reference Liu, Lerosey-Aubril, Steiner, Dunlop, Shu and Paterson2018) (a suite of juvenile Stanleycaris hirpex Pates et al., Reference Pates, Daley and Ortega-Hernandez2018 fossil specimens show that the overall shape of the radiodont body remains largely unchanged through ontogeny, with the eyes developing precociously, although there is a period of anamorphic growth where trunk segments were added posteriorly [Moysiuk and Caron, Reference Moysiuk and Caron2023]). This is highly inconsistent with an interpretation of Parvibellus as a radiodont. Parvibellus as a total-group Onychophoran lobopodian or as belonging to the more basal aysheaiids was also rejected as the trunk of Parvibellus is short, with stout, laterally projecting limbs lacking claws or obvious annulation. In addition, while the overall shape of Parvibellus could be consistent with an aysheaiid, the mouth of Parvibellus is obviously ventral and lined with large triangular tooth plates, while the mouth openings of aysheaiids are devoid of obvious sclerites and directed far anteriorly (Whittington, Reference Whittington1978).

With these groups excluded, the authors allied Parvibellus with the “APOK” group (Anomalocaridids, Pambdelurion, Opabinia, Kerygmachela). As evidence toward this placement, the authors compared the lack of eyes in their specimen with traditional reconstructions of Pambdelurion and Kerygmachela, which were both originally described as lacking eyes. However, recent evidence has demonstrated that both Kerygmachela and Pambdelurion possessed large, ventral, sessile compound eyes—long and sickle-shaped in Kerygmachela (Park et al., Reference Park, Kihm, Woo, Park, Lee, Smith, Harper, Young, Nielsen and Vinther2018) and more bulbous in Pambdelurion (Fleming et al., Reference Fleming, Kristensen, Sørensen, Park and Arakawa2018). These eyes are evidenced by paired and consistently shaped reflective patches on the undersides of the head region, just posterior to the bases of the frontal appendages (in Pambdelurion, they flank the large ventral mouth on either side), making the argument that the lack of eyes is shared between Parvibellus and the traditionally defined gilled lobopodians unconvincing. This also excludes Parvibellus from aligning with the opabiniids Opabinia and Utaurora, while the paired and unfused frontal appendages exclude it from the possibly eyeless taxa recently described from the Castle Bank fauna of Wales (Pates et al., Reference Pates, Wolfe, Lerosey-Aubril, Daley and Ortega-Hernández2022b). The comparison between the appendages of Parvibellus and the dorsal flaps of opabiniids and gilled lobopodians is also in opposition with the emerging view that the dorsal flaps of Pambdelurion and Opabinia were simple extrusions of the body wall lacking muscular control and that the flaps of the kerygmachelids evolved swimming flaps independently.

Not discussed by the authors, or included within the phylogenetic analysis and matrix accompanying the description, is the possibility that the specimen represents one of the slightly more basal siberiid lobopodians, more closely related to the gilled lobopodians and radiodonts than to the total-group onychophoran lobopodians such as Hallucigenia Conway Morris, Reference Conway Morris1977 or Microdictyon Bengston et al., 1986 (Smith and Ortega-Hernández, Reference Smith and Ortega-Hernández2014). Siberiids are very uncommon among lobopodians but easily identified by their broad, mostly uniform trunk regions, paired lobopodous limbs projecting ventrolaterally, absence of any claws on the trunk limbs, large raptorial frontal appendages, ventrally projecting mouth bearing a circumoral ring of triangular sclerites, and lack of eyes. This is in addition to the differences in preservation between Burgess Shale fossils and other Chengjiang fossils, which make some direct comparisons between Parvibellus and fossils of Opabinia problematic. Whereas in Opabinia the flaps lie on different bedding planes, with individual flaps marked by small and sudden changes in elevation (Whittington, Reference Whittington1975), the fossil of Parvibellus is preserved on a single bedding plane. In Parvibellus, dark outlines trace the shapes of the appendages and, although they may have been slightly warped taphonomically, show that the overlapping of the appendages, and overall shape and flexibility, are much more in line with the interpretation of these appendages as siberiid-like lobopodous walking legs rather than flattened, reverse-imbricated swimming flaps. The appendages of Parvibellus are elongate and highly flexible, preserved orientated at inconsistent angles; they mostly overlap each other by the curling and compression of appendages over-top of one another, not overlapping at their bases as is observed in gilled lobopodians and most radiodont taxa.

Parvibellus differs from the siberiids in a few notable ways, however. The lobopods of the siberiids bear repeated elongate papillae lining the posterior margin of the trunk appendages (Dzik, Reference Dzik2011), which perhaps served a respiratory function to support a more active, predatory lifestyle more comparable to the gill-bearing opabiniids and radiodonts, while Parvibellus lacks any such structures. The largest of the siberiids, Jianshanopodia, capable of reaching over 30 cm long judging from comparisons of ELI−J0005A to complete specimens of Jianshanopodia and Megadictyon, has elongate, tree-like papillae projecting from the posterior margins of its lobopods (Liu et al., Reference Liu, Shu, Han, Zhang and Zhang2006; Dzik, Reference Dzik2011). These additional branches to the main structure could have greatly increased the surface area if these papillae served a respiratory function, similar to the convoluted surface of the respiratory setal blades in Aegirocassis. Regarding the digestive tract observed in Parvibellus, it appears to lack any kind of diverticula structures, in contrast to the well-developed digestive tracts of other Chengjiang siberiids (Vannier et al., Reference Vannier, Liu, Lerosey-Aubril, Vinther and Daley2014) (although gut glands may be absent in Siberion). Finally, the frontal appendages of Parvibellus are small and undeveloped, while the frontal appendages of known siberiids are robust and bear raptorial endites or even claws at their tips (Vannier et al., Reference Vannier, Liu, Lerosey-Aubril, Vinther and Daley2014). The lack of these features in Parvibellus could, however, be explained by ontogenetic changes. The smallest known siberiid, Siberion lenaicus Dzik, Reference Dzik2011, is thought to have reached a maximum length of approximately 50–60 mm, judging from the more complete of the two known specimens (Dzik, Reference Dzik2011), while the sole specimen of Parvibellus measures only just over 5 mm, ~10 times smaller than Siberion. The diminutive size of the specimen would not immediately demand the growth of respiratory papillae as the existing surface area of the body would satisfy oxygen requirements. Restricting comparisons to only the Chengjiang Fauna, from which Parvibellus and the majority of siberiid species derive, possible adult forms become much larger, suggesting that Parvibellus could be a very early instar. The size and probably early developmental stage of the specimen could also explain the lack of any raptorial adaptations to the frontal appendages as individuals at this stage must have depended on a very different food source or prey than their large adult counterparts. The posteriormost portion of the specimen is poorly preserved, and a large area of it is missing. What is interpreted as a presumably paired furca-like structure may be interpreted as a reduced limb or part of a caudal fan, as observed in Jianshanopodia (Liu et al., Reference Liu, Shu, Han, Zhang and Zhang2006), or it may represent part of a medial lobe-like structure as seen in both Jianshanopodia and Siberion (Dzik, Reference Dzik2011). The interpretation of this structure as one of two paired furcae may also be correct. Because it probably represents a larval individual of an existing genus or species of siberiid lobopodian, and otherwise cannot be identified conclusively due to its probable immaturity, Parvibellus is here considered a nomen dubium and has not been included in the present phylogenetic analysis.

Phylogenetic relationships

The tree topology (Fig. 5) recovered by analysis of the modified matrix is broadly similar to previously recovered topologies from other studies on radiodont systematics. The trees were rooted with Aysheaia, a consistently basal panarthropod (Howard et al., Reference Howard, Hou, Edgecombe, Salge, Shi and Ma2020). The most-basal clade of total-group Euarthropoda, the Siberiidae (Dzik, Reference Dzik2011), contains Jianshanopodia Liu et al., Reference Liu, Shu, Han, Zhang and Zhang2006, Siberion Dzik, Reference Dzik2011, and Megadictyon Luo and Hu in Luo et al., Reference Luo, Hu, Chen, Zhang and Tao1999. Megadictyon is more poorly understood than the former two genera, with the type species initially known only from a section of the trunk bearing large, robust limbs and diverticulae, missing both the anterior and posterior sections of the body. Because of this ambiguity in the holotype, all subsequent, more-complete specimens have been attributed only tentatively to the genus, under Megadictyon cf. M. haikouensis (Liu et al., Reference Liu, Shu, Han, Zhang and Zhang2007). There are several more indeterminate siberiid lobopodian fossils from the Chengjiang Fauna (Vannier et al., Reference Vannier, Liu, Lerosey-Aubril, Vinther and Daley2014), but these were not included in the study. A new clade of three taxa, including Utahnax vannieri, Kerygmachela kierkegaardi, and Mobulavermis adustus n. gen. n. sp., is recovered in all trees under both equal weighting and implied weighting. In past studies, the informal name gilled lobopodians was used to join Kerygmachela and Pambdelurion. No formal name was established to unite the two genera, however, as the two genera formed a grade basal to the opabiniids, radiodonts, and arthropods rather than a distinct clade of closely related animals. The descriptions of Utahnax vannieri and of Mobulavermis adustus (Fig. 6) allow a more confident separation of Pambdelurion whittingtoni (and by extension, Omnidens amplus, which clades with Pambdelurion) from Kerygmachela and the establishment of a new formal clade between Siberiidae and the Opabiniidae (a monophyletic clade of Opabinia regalis and Utaurora comosa is present in all recovered trees, while the inclusion of the Castle Bank taxa is uncertain).

Figure 5. Phylogenetic results. (1) Majority-rule (50%) consensus of 299 most parsimonious trees under equal weighting (CI = 0.518, RI = 0.820), with newly named taxa highlighted and representative taxa bolded. Numbers beneath branches represent the percentage of trees in which that node was recovered, while branches without percentages are present in all trees. (2) Simplified trunk body plans of representative lobopodian and radiodont taxa (showing arrangements of annulation, setal blades, and limbs), in same order as bolded names. Models are shown facing away and to the right.

Figure 6. Digital life reconstruction of Mobulavermis adustus, shown swimming in a Cambrian ocean alongside small bivalved arthropods and early chordates (speculative anterior morphology based on Kerygmachela kierkegaardi).

Deuteropoda is recovered as sister to Radiodonta and is rooted in a polytomy of the recently described Erratus sperare Fu et al., Reference Fu, Legg, Daley, Budd, Wu and Zhang2022 and (under equal weighting) the much earlier discovered Cucumericrus decoratus Hou et al., Reference Hou, Bergström and Ahlberg1995. Both possess elongate unsclerotized trunk regions (save for the single large tergite that covers the head and much of the body in Erratus; a similar tergite is assumed to be present in Cucumericrus by phylogenetic bracketing but is not known from direct fossil evidence), with biramous appendages consisting of a lobopodous or distally sclerotized (Simonetta, Reference Simonetta2004) walking limb and a soft dorsolateral flap. They are followed by the Isoxyidae and then by the euarthropods, including Kylinxia zhangi Zeng et al., Reference Zeng, Zhao, Niu, Zhu and Huang2020 (Izquierdo-López and Caron, Reference Izquierdo-López and Caron2022). A probable group of “paranomalocaridids”, composed of Paranomalocaris and a recently figured undescribed species from the Kinzers Formation, were excluded from the matrix due to a relative absence of relevant applicable characters and the lack of any known remains besides the frontal appendages. The paranomalocaridids were similarly excluded from the analysis by Zeng et al. (Reference Zeng, Zhao and Zhu2023) with the additional reason of their radiodont affinities being questionable overall. Caryosyntrips Daley and Budd, Reference Daley and Budd2010, typically classed as a basal radiodont, is here recovered as basal to both Radiodonta and Deuteropoda. Caryosyntrips is recovered in a similar position by Moysiuk and Caron (Reference Moysiuk and Caron2019) and found in a polytomy with Deuteropoda and Radiodonta by Lerosey-Aubril and Pates (Reference Lerosey-Aubril and Pates2018) and by Moysiuk and Caron (Reference Moysiuk and Caron2021). The analysis by Zeng et al. (Reference Zeng, Zhao and Zhu2023) found Caryosyntrips as within Euarthropoda, rather than sister to it, or within Radiodonta.

A monophyletic Peytoiidae is rooted under equal weighting at two recently discovered fossils from the Chengjiang Fauna (Wu et al., Reference Wu, Pates, Ma, Lin, Wu, Zhang and Fu2022) attributed to Peytoiidae, but which both possess posteriorly oriented pectinate auxiliary spines, in contrast to the forward-facing or mesially oriented pectinate endites of all other peytoiids. The family name Peytoiidae is employed rather than “Hurdiidae” because, as pointed out by Greenfield (Reference Greenfield2023), the family name Hurdiidae has yet to be properly established following ICZN standards, and the name Peytoiidae (Conway Morris and Robison, Reference Conway Morris and Robison1982) is, as the oldest and only available name for the clade, the correct name to use. The known members of the Tamisiocarididae all form a polytomy with the base of Peytoiidae. Notably, Houcaris saron Hou et al., Reference Hou, Bergström and Ahlberg1995, as well as “Ramskoeldia” consimilis Cong et al., Reference Cong, Edgecombe, Daley, Guo, Pates and Hou2018, which is here included under the genus Houcaris (see Supplementary Information), are not recovered as tamisiocaridids as recently described (Wu et al., Reference Wu, Fu, Ma, Lin, Sun and Zhang2021a) but consistently as members of Amplectobeluidae.

Two species tentatively attributed to Innovatiocaris? in the description of Innovatiocaris maotianshanensis Zeng et al., Reference Zeng, Zhao and Zhu2023 form a monophyletic group within lower Amplectobeluidae, distant from the type species in Anomalocarididae, indicating they probably belong to a distinct genus. They are similar to the type species in their lack of posteriorly oriented auxiliary spines but easily differentiated by the outward kink separating the shaft and distal articulated regions, the hypertrophied first distal articulated zone endite pair, and the highly reduced or absent shaft endites in Innovatiocaris? multispiniformis Zeng et al., Reference Zeng, Zhao and Zhu2023 and Innovatiocaris sp., all distinctive traits of amplectobeluids. This morphological disparity warrants the establishment of a new genus, but this is outside the scope of the current study. Laminacaris Guo et al., Reference Guo, Pates, Cong, Daley, Edgecombe, Chen and Hou2019 is recovered as a basal amplectobeluid alongside Guanshancaris kunmingensis Wang et al., Reference Wang, Huang and Hu2013, alongside “Anomalocaris” sp. from the Balang Formation (Liu, Reference Liu2013), although the placement of “Anomalocaris” sp. should be viewed with caution until more complete material is described.