The Cambrian Odaraia alata and the colonization of nektonic suspension-feeding niches by early mandibulates

(a) Systematic palaeontology

Arthropoda [27]

Mandibulata [28]

Order Hymenocarina [29] (last emended [8])

Family Odaraiidae [30]

Type genus: Odaraia Walcott, 1912 [31]

Type species: O. alata Walcott, 1912 [31]

Included taxa: O. alata [31]; Balhuticaris voltae [18]; Fibulacaris nereidis [26]; Jugatacaris agilis [32]; Nereocaris exilis [16]; Nereocaris briggsi [17]; Pakucaris apatis [33]. Other putative members: Vermontcaris montcalmi [34].

Diagnosis (last emended in Briggs [15]): hymenocarines without a dorsal hinge line, carapace; antennulae small, with fewer than seven antennal podomeres or entirely absent; body multi-segmented; endopods multipodomerous with homonomous podomeres; caudal rami suboval when present.

Genus Odaraia Walcott, 1912 [31]

Type species: O. alata Walcott, 1912 [31]

Diagnosis (emended from Briggs [15]): odaraiid with two large compound eyes flanking three small sensory organs; antennae absent; mandibles containing a row of six to seven large teeth and an incisor process; pair of lobe-like elements posterior to the mandibles and an isolated, highly sclerotized tooth in the oral area; thoracic legs with podomeres bearing two pairs of prominent, elongated spines and two pairs of smaller spines closer to podomere boundaries, and numerous tiny spines along the ventral margins of the podomeres; terminal segment fused with the caudal rami extending posteriorly into a dorsal fin. Caudal rami with anteroposterior differentiation, each bearing two small spines on their lateral edges.

(b). Description

Head and ocular segment. The head is large, anteriorly convex and extends frontally beyond the frontal margin of the carapace (figure 1a,e and electronic supplementary material, figure S3). The head includes the ocular segment, deuto- and trito-cerebral segments, mandibular segment and maxillar segment, all fused together. The ocular segment bears two large compound eyes originating laterally from short eye peduncles that project anterolaterally (figure 1a,b and electronic supplementary material, figure S2a). Three reflective structures are located on the dorsal side of the head, between compound eyes, along an anterior bulging area (figure 1b and electronic supplementary material, figures S4 and S5a,b), previously interpreted as sensory organs (median eyes in Briggs [15] or protocerebral frontal organs in Ortega-Hernández [35]). These structures were illustrated to be preserved either below or embedded into what has been interpreted as an anterior or ocular sclerite (ACS in Ortega-Hernández & Budd [36]). We cannot confirm the presence of a distinct sclerite boundary in the material studied herein that would differentiate it from the rest of the head, including in the two Odaraia specimens illustrated by Ortega-Hernández [35]. A disarticulated specimen, though, shows the head bearing a pair of pedunculate eyes with a characteristic round section that could potentially correspond to the frontal edge of the sclerite (electronic supplementary material, figure S4). Two small bulges visible in one specimen may represent potential peduncular lobes (electronic supplementary material, figure S6a,b).

Hypostome/labrum. Several specimens bear a ventral triangular to trapezoidal bulge on the frontalmost section of the head, before the first cephalic appendages (the mandibles) (figures 1c,d and 2p ), most probably representing a hypostome (see §5).

Mandibles. The presence of mandibles had already been suggested by Briggs [15] in the form of tooth-like structures (electronic supplementary material, table S2), usually adaxial to round structures interpreted as adductor muscle scars, but other details remained unknown. Here, we reinterpret some of these hypothetical muscle scars as the base of the mandibles based on new material showing the contiguity of these structures with the teeth, thus affirming the presence of mandibles (figures 1a–f , 2p and 3a , and electronic supplementary material, figures S2b,c, S3b, S4a–d, S5a–c, S6a,b, S7c–e, S8a,b, S9a,c and S10f–i). Mandibles are large and occupy almost the entire width of the head. They are stout, slightly wider than long and have an overall convex, recurved morphology (figure 1f,g and electronic supplementary material, figures S5c, S6c–e, S7c–e, S11f and S12a,b,g). The dorsal side bulges (electronic supplementary material, figure S4c). The posterior side ends in a singular blunt process (figure 3c and electronic supplementary material, figure S1c and S8c). The ventral side of the mandible is probably invaginated, providing a surface to which musculature could attach (figure 3c and electronic supplementary material, figure S1c). The frontal side of the mandible extends into a flat surface (figure 1g and electronic supplementary material, figure S5c and 7c) that curves slightly inwards distally (figure 3e and electronic supplementary material, figure S10f–i). This surface ends in a row of teeth (electronic supplementary material, figure S7e,f) that curve slightly inwards (figure 1h and 3f,g and electronic supplementary material, figures S2b,c and S10f–i). Teeth are distributed into two groups. The first group bears six to seven teeth in a single row, of which two to three may be smaller than the remaining four teeth, which are equal in size each (figures 1e,g,h and 3b,d,e,f,g and electronic supplementary material, figures S1c, S7e and S9c). The second is a single, larger tooth, a pars incisor, positioned more posteriorly (figure 3b,d and electronic supplementary material, figure S9c). One specimen exhibits a series of small setae along the mandible (figure 3h ), but their exact morphology and position are not clear.

Post-mandibular structures. Three structures are present posterior to the mandibles, which belong to the feeding apparatus. The first is a small, trident-shaped singular element situated centrally between the mandibles, here referred to as 'central tooth'. This tooth-like structure is three-dimensional, highly sclerotized (figures 1f–i , 2a and 3p and electronic supplementary material, figures S6f–i, S7g,h and S9d) and has four denticles of similar size, three situated posteriorly and one anteriorly, while its median section evaginates (figure 1i and electronic supplementary material, figures S6i, S7g and S11f). The second structure is a pair of semicircular lobes with a round outline positioned invariably below the mandibles (figure 1g and electronic supplementary material, figures S6f–i and S7d,e). These are preserved faintly, implying that they are thin, and preserved in a different level of the matrix than surrounding structures, indicating that they are isolated. We cannot reject the possibility that these structures might be connected to the central tooth, although the superposition of these features is also likely. These posterior lobes may be interpreted as paragnaths (see §5). The third structure is an elongated cephalic appendage. The point of attachment is difficult to infer, but its position posterior to the mandibles suggests it is probably a maxilla (e.g. cephalic appendage in figure 2c,p and electronic supplementary material, figures S11d and S13). This appendage is differentiated from the thoracic limbs and is subdivided into seven to eight podomeres, with at least the distalmost bearing numerous setae (figure 3h–m and electronic supplementary material, figure S1d).

Carapace. Bivalved carapace without hinge line (electronic supplementary material, figure S6b) that covers between half and two-thirds of the total body length. The carapace valves extend ventrally, partly enveloping the head and trunk (figure 1a and electronic supplementary material, figures S1a,b and S7a,b). The dorsal side of the carapace may bear a small posterior process as described in Briggs [15].

Trunk. The trunk is composed of 30−35 homonomous tergopleural rings that decrease in width posteriorly (figure 1a and electronic supplementary material, figures S2a, S9a,b and S10a,b) with a small trapezoidal sternite on their ventral side (figure 2a,b,p,q and electronic supplementary material, figure S13). Each segment bears three or five rows of small spines on its dorsal side: one row of spines along the axial area with more prominent spines posteriorly, and one or two additional smaller spine rows, laterally (electronic supplementary material, figures S10n, S11, S14e and S22). A dark trace originating from the head and extending across the ventral side of the animal subdivided into nodules could represent a ventral nerve cord (figure 2a,b and electronic supplementary material, figure S11a,g) and could be potentially present on other specimens preserved as a dim dark tissue transversal across the body, although its affinity is more ambiguous, in these cases (e.g. figure 1b and electronic supplementary material, figure S4).

Legs. Each trunk segment bears a pair of biramous legs. Endopods are thin and taper in width distally (electronic supplementary material, figure S9a,b) and are subdivided into 19−20 homonomous podomeres (figure 2c,f,r and electronic supplementary material, figures S2a and S13). The six to seven proximalmost podomeres are shorter (figure 2r ) than the remaining ones. The 10 distalmost podomeres are longer, with the distalmost podomere being a short pad (figure 2h ). All podomeres bear a large round to conic endite (figure 3f,n and electronic supplementary material, figure S10l,m). These previously unrecognized endites are covered in thin spines of different lengths on their surface, many facing distally (figure 2n,o and electronic supplementary material, figure S10m), as well as a set of eight larger spines. Four of these spines originate from each of the lateral sides of the endite (figure 2m and electronic supplementary material, figures S1f and S15b), extending laterally or ventrolaterally (figure 2h–j,m ). On each side of the endite, of the four spines, two are smaller, although slightly larger than the other small spines covering the surface of the endite (figure 2n ). Each of these two spines is positioned at the proximalmost and distalmost section of the endite, respectively. The remaining two spines are considerably larger: they are highly elongated (figure 2h,m and electronic supplementary material, figures S10k and S11d,e) and slightly curved (figure 2c,m ), with one originating more proximally than the other. Each of the spines faces a different lateral to ventrolateral direction; spines extending from each podomere interlock between limbs, creating a mesh-like structure (figure 2i,j , figure 4 and electronic supplementary material, figures S1e and S15b). Exopods are ovate, as long or slightly longer than the endopods (figure 2a,d–g and electronic supplementary material, figure S14) and terminate in a small blunt process (figure 2e ). It is possible that the proximalmost section of the exopod is fused to the endopod, leaving only the distalmost (approx. 10) podomeres of the endopod free (figure 2g and electronic supplementary material, figure S16b). The exopod bears a series of cuticular extensions (i.e. lamellae) connected to a singular rod (figure 2d,k,l and electronic supplementary material, figure S18b) at least on its proximalmost surface, but their total extension is not clear.

Gut. The gut (e.g. figure 1a and electronic supplementary material, figures S14a–e and S17c) is usually preserved three-dimensionally and terminates at the ventral side of the terminal trunk segment as the anus (electronic supplementary material, figures S4a and S22). Under direct light, two specimens (e.g. electronic supplementary material, figures S3a and S12b,e,f) exhibit a pair of large bright structures surrounding the frontal section of the gut and extending into the limb bases, interpreted as gut diverticula.

Terminal segment. The terminal segment of the trunk is approximately five times longer than the preceding trunk segments (figure 1a ). It has a conical shape and terminates into three fin-like structures: two oriented laterally and one oriented dorsally (electronic supplementary material, figures S16a, S17c, S19a and S22). The dorsal element is an extension of the terminal segment (electronic supplementary material, figures S8a–d and S14). It is thin and semicircular and bears a small spine on its dorsal edge. The other two fin elements are caudal rami that do not show any articulation to the terminal segment, implying that these structures are fused together (electronic supplementary material, figure S4b,d). The caudal rami are suboval but are differentiated into two sections, with the frontal one thicker than the posterior one (electronic supplementary material, figures S3c,d, S5a,b, S9a,b,e and S22). The frontal section bears two spines on its lateral edge, one located at the distal edge (electronic supplementary material, figure S22). The posterior section of the caudal rami may bear a series of setae (electronic supplementary material, figure S1c,d) and has an overall rectangular shape. One specimen exhibits distinct features that could represent haemolymph cavities or another type of internal tissue (electronic supplementary material, figure S22).

(a). Odaraia’s significance for early mandibulate evolution

Based on its cephalic anatomy, Odaraia can be confidently regarded as a mandibulate, aligning with previous reconstructions [15]. Incorporating the new morphological evidence of cephalic structures, Odaraia further strengthens the placement of odaraiids as early mandibulates, together with other hymenocarines and the Cambrian group Fuxianhuiida (figure 5) as previously suggested [20].

Most hymenocarines have only a pair of first antennae (or none, as in Odaraia) and mandibles, which implies the presence of a limbless, intercalary segment in between [10]. It follows, though, that a second pair of antennae evolved from this limbless segment later within crown crustaceans, a transition considered non-parsimonious [11,13]. Accordingly, there have been three lines of argument to tackle this conundrum. The first one is that hymenocarines may have a highly reduced second pair of antennae [14] (cf. [10]), and, in fact, Briggs [15] already described two pairs of appendages anterior to the mandibles (figs. 61, 66 and 103b in Briggs [15]), albeit on limited evidence. The second one is that some morphological reconstructions that identified mandibles may not be completely accurate ([11], cf. [10]), questioning the presence of this structure across all hymenocarines. Finally, the putative mandibles could belong to other cephalic segments [11], making hymenocarines a paraphyletic group of mandibulates and non-mandibulate arthropods that underwent an extraordinary case of convergent evolution [11]. Our study confirms the presence of mandibles and, therefore, of an intercalary segment based on both morphological and positional evidence.

Among hymenocarines, mandibles have been recovered in Odaraia [15], protocaridids [6], Waptia fieldensis [8], Ercaicunia multinudosa [14] and Canadaspis perfecta [19,37]. This accumulating evidence endorses the authenticity of these structures, contrary to alternative interpretations (e.g. limb bases [11]), and showcases that mandibles were not unique to a specific group (e.g. waptiids), but were present across all major hymenocarine subgroups (e.g. odaraiids, protocaridids and clypecaridids). It follows, that some of the rounded cephalic structures in other hymenocarines (e.g. Pectocaris) are most probably not muscle scars [38], but instead genuine mandibles (see, for example, potential teeth in fig. 3f of Jin et al. [38]). The new material of Odaraia permits a unique three-dimensional reconstruction of the mandible (figure 5), including the exact number and shape of the teeth and a distinct pars incisor, previously regarded as a key mandibulate synapomorphy [39], rejecting previous raptorial reconstructions [16]. The stout, cup-shaped morphology of the mandible is analogous to that present across several crustacean groups [40,41], and teeth similar in number, size and general morphology are present in copepods [42], or in Cambrian small carbonaceous fossils of potential pancrustacean affinity (fig. 1i in Harvey et al. [43]). Perhaps most striking is the apparent lack of a mandibular palp, which was previously observed in the hymenocarine Waptia [8] and has been considered an ancestral mandibulate feature [13]. It is possible that poor preservation could have obscured this structure across hymenocarines, but also that the palp was lost secondarily in several hymenocarine groups like the odaraiids, as happened in other mandibulates groups (e.g. branchiopods).

Positional evidence comes from surrounding structures, which, independently from the mandible, also show mandibulate affinities. The posterior semicircular lobes observed in Odaraia are thin and unsegmented. These lobes could represent cuticular outgrowths, which, based on their position posterior to the mandible, would be homologous to the paragnaths of crustaceans and the hypopharynx of myriapods [44]. Paragnaths (or homologous structures) have already been recognized in several Cambrian crustaceomorph species (e.g. Oelandocaris oelandica [45] and Wujicaris muelleri [46]) but not in hymenocarines, and represent a strong mandibulate synapomorphy [44]. Their absence in other early mandibulates (e.g. fuxianhuiids) is most probably taphonomic, resulting from their small size and position. On the other hand, the singular tooth is so far, a structure unique to Odaraia. Its high sclerotization and three-dimensional shape suggest that this structure may have a feeding function (see discussion below), indicating that the mouth was most probably situated in between the mandibles, further supporting their post-tritocerebral position. Post-mandibular appendages (i.e. first and second maxillae) are increasingly being discovered in hymenocarines [6,8,14,37], including, although only inferred, in the odaraiid Pakucaris apatis [33]. While differentiated cephalic appendages posterior to the deutocerebrum can be found in other Cambrian non-mandibulate arthropods such as habeliids [47], the presence of a pair of dentate appendages (i.e. a mandible) followed by another pair of differentiated reduced appendages (i.e. maxillae) is a highly distinct mandibulate trait. Given the basal position of hymenocarines respective to other mandibulates (figure 5), the presence of true, post-tritocerebral mandibles is better supported by common ancestry than by convergence of different cephalic appendages into mandible-like morphologies (see further discussion in supplementary material).

Yet, the presence of an intercalary segment and the re-expression of the tritocerebral appendage in crustaceans still represent a non-parsimonious transition from the point of view of character optimization [11,13]. Developmentally, it is worth noting that the formation of limbs is independent of the number and position of limb segments (e.g. [48–50]). It follows that the re-expression of a limb in the crustacean head may not require a rearrangement of their entire cephalic morphology. In hexapods, in fact, in which this appendage is similarly not expressed, limb buds can still be seen early in development [51], and their formation can be induced through alterations of gene expression [52]. An intercalary segment has also been induced in arthropods that have a tritocerebral appendage [53], suggesting that the re-expression of this appendage is dependent on downstream expression that could have changed across early arthropod evolution.

In addition to clarifying the morphology of odaraiids and further supporting the post-tritocerebral identity of the mandibles, this re-examination of Odaraia supports recent reinterpretations of fuxianhuiids as early mandibulates [20]. Fuxianhuiids are characterized by a head shield, multipodomerous legs and a head with two pairs of appendages: the first antennae and what have been previously interpreted as raptorial specialized post-antennal appendages (e.g. [12]), but more recently homologized with the mandibles [20]. A further connection between fuxianhuiids, hymenocarines and euthycarcinoids, the latter considered stem-group myriapods [9], has been suggested by Aria et al. [20] based on shared cephalic and limb characters. This interpretation would imply that the tritocerebral appendage was present in a yet unidentified mandibulate ancestor and was lost convergently in the myriapod stem (containing hymenocarines) and in hexapods. However, this scenario was not supported by their phylogenetic results, which recovered a stem-mandibulate position for all hymenocarines and fuxianhuiids. This result was already acknowledged by the same authors, who suggested that the lack of cephalic information in odaraiids could have obscured a stronger connection between these three groups [20], a connection that is not phylogenetically supported with the new material of Odaraia, at present (figure 5).

The new material of Odaraia shows a triangular to trapezoidal plate positioned in front of the mouth and mandibles. A plate similar in shape, size and position can be found in fuxianhuiids, termed a hypostome [20] and comparable to the clypeus/epistome of extant mandibulates. The hypostome is partly fused to a plate-like bilobed labrum that extends above the mandibles [20], although both structures have traditionally been considered a single structure similarly termed hypostome (e.g. [12,54]). An alternative definition of the hypostome states that it is the result of a fusion between the ocular sclerite and the labrum [55]. This is clearly not the case in Odaraia, in which the ocular sclerite would be invariably positioned fronto-dorsally, separated from ventral plates. In Odaraia, however, there is no evidence of a separate labrum adjacent to the hypostome sensu [20]. Sclerotic plates partly or totally disassociated from the head have been reported in other hymenocarines and regarded as potential labral plates, although their affinity remains unclear [18]. It is possible that a labrum has similarly become disassociated from the remaining cephalic tissues and has not been preserved in the material available, although this remains unlikely given how prominent this structure is in fuxianhuiids and the present quantity of specimens of Odaraia preserving mandibles in situ. Briggs [15] described the presence of a labrum in Odaraia but was highly cautious about this interpretation, and based on its size and position, it is unlikely that this structure represented a labrum. The following two hypotheses remain: first, that this structure is, in fact, the labrum and that a more anterior hypostome is either not present or not preserved; and second, that both structures are either fused or that a separation between them is not visible in the material herein.

The labrum is a feature ancestral to all euarthropods [36,56] and thus should be present in hymenocarines, as inferred (e.g. 6). If both the labrum and the hypostome are indeed fused in Odaraia, this structure could have acquired its bulged morphology to assist in retaining particles during suspension feeding. Morphologically similar hypostome-labra appear in extant crustacean suspension feeders, such as remipedes [57] and some euphausiids [58], and could have similarly been present in other hymenocarines such as Ercaicunia (hypostome in Zhai et al. [14]). In benthic crawlers such as fuxianhuiids, the labrum could have expanded above the mandibles and acquired its highly sclerotized morphology to protect the mandibles from abrasion from the seafloor, converging in shape with that of trilobites or extant notostracans [41]. This morphological diversity, then, could have obscured the homologization across early arthropods and should be considered accordingly.

Overall, both odaraiids and fuxianhuiids share a potential fronto-dorsal ocular sclerite and a ventral labrum fused to or associated with a triangular-shaped hypostome (see electronic supplementary material, table S3), but the codification of these and other traits in the morphological matrix did not result in a stronger phylogenetic connection between both groups (figure 5 and electronic supplementary material, tables S4 and S20). Odaraiids and fuxianhuiids share further morphological similarities, mostly their body multi-segmentation and multipodomerous legs, but there are notable morphological differences, including the number and morphology of differentiated cephalic appendages (e.g. the raptorial palp and lack of maxillae in fuxianhuiids), the carapace, type of tagmatization (e.g. polypody in fuxianhuiids) or the presence of gnathobasic-like podomeres in fuxianhuiids (see electronic supplementary discussion). Nonetheless, the interrelations across early mandibulate groups have been weakly supported across phylogenetic analyses (e.g. [19,20]), implying that further morphological data, including potential transitional forms, are needed to further clarify this question.

(b). Suspension feeding in early mandibulates and their ecosystemic role

Based on their undifferentiated limbs, homonomous body segments and lack of raptorial appendages, many odaraiids have been interpreted as deposit [17] or suspension feeders [26]. Odaraia has been similarly cited as one of the most compelling examples of a suspension-feeding euarthropod at the Burgess Shale [15], but these ecological interpretations needed further support from unambiguous anatomical evidence. The principle of suspension feeding is that of capturing or retaining suspended food particles [59], which, in mandibulates such as crustaceans, is usually accomplished with a mesh-like structure [5,60], sometimes accompanied by secondary methods (e.g. mucus production). The apparent lack of these structures in odaraiids could be attributed to the smaller size of some of these taxa (e.g. Fibulacaris [26]) but is more difficult to justify for larger taxa (as originally mentioned, for example, in Odaraia [15] and Balhuticaris [18], which would be expected not only to offer a higher anatomical resolution but also to require large filtering surfaces [60], easily visible in other Cambrian suspension feeders of similar body size (e.g. [61]).

The new material shows that the legs of Odaraia have podomeres with endites bearing ventrally and ventro-laterally directed short and long spines, hardly visible in the specimens studied by Briggs [15]. These spines imbricate with those on the surrounding legs, creating a mesh (figure 6) highly reminiscent of those in extant suspension-feeding arthropods (e.g. euphausiids [60] and anostracans [62]). Spinose endites had been previously inferred in the hymenocarine Pectocaris [63], but, likewise, their morphology was not clarified until much later [38]. This re-evaluation further highlights the role of taphonomy in morphological and ecological reconstructions, even when exceptionally preserved material is available.

Measurements of the space left between the larger spines of Odaraia (around 0.5 × 0.8 mm) indicate that this mesh would have been able to capture prey in the size range of mesozooplankton [61], as well as smaller particles when closed through passive forces [59] or by involving the smaller enditic spines. These smaller spines, as well as similar features (e.g. the setal brush of the maxillae), could also have aided in cleaning filtering surfaces and bringing food towards the mouth, a function common among extant suspension feeders [64,65]. Another structure aiding in feeding is Odaraia’s central tooth. Spines and microspines are found across all panarthropods [66], but this tooth is comparatively larger and appears in isolation. The function of the central tooth is more likely analogous to the functions of ossicles and gastric teeth of extant crustaceans [67], which grind food together with the mandibles. This structure appears unique among all described hymenocarines and other early mandibulates.

It has previously been suggested that a predatory lifestyle may have complemented suspension feeding in large odaraiids such as Odaraia [15] and Balhuticaris [18] to meet the energetic requirements of larger body sizes [60]. Predatory activities are supported by additional adaptations, for instance, the size and type of eyes. The gut diverticula of Odaraia have also been previously linked to predatory or macrophagous habits [68], and its mandibular teeth are similar in shape and number to those in notostracans ([41]; electronic supplementary material, figure S21) omnivorous copepods [69] and laevicaudatans [70], which have generalist or omnivorous diets. The presence of these features reinforces the idea that odaraiids contained microphagous (e.g. Fibulacaris) and macrophagous (e.g. Odaraia) suspension feeders that used a mesh to filter feed in currents but also captured larger prey (e.g. mesocosm plankton), a lifestyle currently present in notostracans, large anostracans [62,64] and other crustaceans [71].

Odaraia and other large-sized odaraiids (>100 mm) (e.g. [18]), closely related hymenocarines [25] and some radiodonts [61] could have constituted a community of large, highly motile suspension feeders. Odaraia’s nektobenthic lifestyle is inferred since the carapace encloses its body and limbs and its tail fan. Briggs [15] already highlighted the unique morphology of the tail fan among all known arthropods and inferred a stabilizing and steering function. Here, we note the presence of small setae along the caudal rami, homologous to those found in other hymenocarines (e.g. Balhuticaris [18]), which could act as mechanosensory or chemosensory organs, further adaptations towards a motile nektobenthic lifestyle.

These communities of large suspension feeders could have been key to the development of the Palaeozoic marine food webs. The evolution of these food webs was characterized by the colonization of multiple pelagic to benthic niches in the water column, in which arthropods played a major role [72,73]. The origin of these food webs has long been linked to a change from benthic-dominated Cambrian ecosystems to an increase in plankton abundance and diversity during the ‘Ordovician Plankton Revolution’, but there is increasing evidence that this change had already started by the Cambrian period [72]. Zooplankton communities were already diverse by the middle Cambrian [3,43,74], but the interconnections between benthos and plankton at that time are not fully understood, yet. It is possible, then, that several Cambrian groups were already establishing pelagic communities that were connecting food webs across the water column, including arthropods [75], chaetognaths, ctenophores [76] and cnidarians [77]. Today, suspension feeders play a critical role in connecting the energy flux at different levels of the water column, especially in larger-bodied taxa that enrich the water column through faecal pellets or food clustering [78], a role that could have been performed by large odaraiids and suspension-feeding radiodonts in the Cambrian.