Jaw elements in Plumulites bengtsoni confirm that machaeridians are extinct armoured scaleworms

Abstract

Machaeridians are Palaeozoic animals that are dorsally armoured with serialized, imbricating shell plates that cover or enclose the body. Prior to the discovery of an articulated plumulitid machaeridian from the Early Ordovician of Morocco that preserved unambiguous annelid characters (segmental parapodia with chaetae), machaeridians were a palaeontological mystery, having been previously linked to echinoderms, barnacles, tommotiids (putative stem-group brachiopods) or molluscs. Although the annelid affinities of machaeridians are now firmly established, their position within the phylum and relevance for understanding the early evolution of Annelida is less secure, with competing hypotheses placing Machaeridia in the stem or deeply nested within the crown group of annelids. We describe a scleritome of Plumulites bengtsoni from the Fezouata Formation of Morocco that preserves an anterior jaw apparatus consisting of at least two discrete elements that exhibit growth lines. Although jaws have multiple independent origins within the annelid crown group, comparable jaws are present only within Phyllodocida, the clade that contains modern aphroditiforms (scaleworms and relatives). Phylogenetic analysis places a monophyletic Machaeridia within the crown group of Phyllodocida in total-group Aphroditiformia, consistent with a common origin of machaeridian shell plates and scaleworm elytrae. The inclusion of machaeridians in Aphroditiformia truncates the ghost lineage of Phyllodocida by almost a hundred million years.

1. Introduction

Machaeridians were first described in 1857 and for the following 150 years they remained in taxonomic limbo, having been compared by different authors to echinoderms, barnacles, molluscs and annelids [1–4]. They are armoured with serially repeated inner and outer shell plates that either cover the dorsal surface only (Plumulitidae) or completely enclose the body (Turrilepadidae and Lepidocoleidae). Machaeridians are common as disarticulated remains, which are often most easily studied when silicified [5], but occur more rarely as complete or partially complete scleritomes [6,7]. They are united by a distinctive bilayered shell plate ultrastructure [2,8] and growth of the shell plates by marginal accretion [5].

The discovery of a partial plumulitid scleritome with soft tissues in part resolved the question of machaeridian affinities. The key specimen of Plumulites bengtsoni was a partial scleritome, with some shell plates lost from the body prior to burial. It preserves a through gut and parapodia with chaetal bundles, the latter confirming that machaeridians are in fact an extinct clade of armoured annelids. Further support for the annelid affinity of all machaeridian families came from a specimen of Lepidocoleus hohensteini from the Lower Devonian Hunsrück Slate [9]. One specimen of this species preserves soft tissues that are replicated in pyrite, demonstrating that lepidocoleids also have a segmented body and that the single shell plate series of lepidocoleids attached to alternating segments [9].

Nevertheless, the precise phylogenetic position of machaeridians relative to extinct and extant annelid groups remains mysterious, with competing hypotheses placing them in the annelid stem [10] or crown [7] group. A placement in the annelid crown group was proposed based on a proposed homology of shell plates and the elytrae of aphroditiform polychaetes (scaleworms and their close relatives) [7]. Scaleworm elytrae are most likely modified dorsal cirri [11]. They alternate along the body with cirriform dorsal cirri, much like the alternating shell plate morphologies of machaeridians. Alternation of cirrus morphology also occurs in other clades within Phyllodocida (the group to which aphroditiforms belong), such as in Syllidae [11], and so this alternation pattern may be a synapomorphy of Phyllodocida or one of its subgroups. Machaeridians have alternatively been proposed to belong to the annelid stem group [10], but this has been defended mostly by the presence of numerous autapomorphies in machaeridians (e.g. the mineralized scleritome), characters that are typically not relevant to resolving the phylogenetic position of a given taxon.

There have only been three analyses that considered machaeridian affinities using discrete character data. The analysis of Sigwart & Sutton [12] explicitly considered the serial shell plates of machaeridians as homologous with the serially repeated shell fields of aculiferan molluscs, although machaeridian shell plates differ from aculiferan shell fields in lacking an inflexible, fixed orientation relative to the body [1], and also lack the cross lamellar ultrastructure characteristic of molluscan shells [2,5,8]. This analysis recovered machaeridians as members of the aplacophoran stem group and did not recover the monophyly of Machaeridia. The analysis of Vinther et al. [8] was performed in conjunction with the first description of machaeridian soft tissues but the position of machaeridians with respect to crown-group annelids and the stem-group annelid Canadia was contingent on whether or not shell plates were coded as homologous with dorsal cirri. If shell plates were not coded as modified dorsal cirri, then machaeridians formed part of a basal polytomy that collapsed all total-group annelids, but nonetheless supported the monophyly of machaeridians. If these structures were instead coded as cirrus homologues, machaeridians were recovered as part of total-group Aciculata, and thus crown-group Annelida. A third analysis likewise found uncertainty in the placement of Plumulites [13], where it behaved as a wildcard taxon among annelids.

The analysis of Vinther et al. [8] was substantially perturbed by the inclusion of tommotiids, which are putative stem-group brachiopods [14]. Dzik [4] proposed a close relationship between machaeridians and other Palaeozoic scleritomous animals, including tommotiids, siphogonuchitids, Wiwaxia and early polyplacophorans. In particular, he proposed that machaeridians could have been derived from a tommotiid ancestor (i.e. that machaeridians are phylogenetically nested within Tommotiida) as they share the presence of multiple differentiated elements in a single organism and a sclerite ornament of prominent concentric rugae. By contrast, however, tommotiids have a brachiopod-like microstructure, and show adapical growth rather than the marginal accretion in machaeridian shell plates [15].

These concerns aside however, if machaeridians are stem-group annelids [10], then a common annelid-brachiopod ancestor with tommotiid-like sclerites could conceivably be proposed, with sclerites becoming modified to machaeridian shell plates and then ultimately lost later in the annelid stem lineage. A scleritomous ancestor for annelids, molluscs and brachiopods has previously been discussed, such as in the hypothesis of Conway Morris and Peel [16], which sought to link halkieriids and Wiwaxia to the stem lineages of annelids, brachiopods and molluscs. Given that these three phyla are closely related, the hypothesis of a common origin of machaeridians and other scleritomous fossil taxa deserves some attention, especially considering the tumultuous history of the classification of scleritomous Cambrian and Ordovician animals.

Since the discovery of machaeridian soft tissues [8], numerous discoveries have added data pertinent to the affinities of machaeridians and other scleritome-bearing early lophotrochozoans. These include the radula-like nature of the mouthparts of Wiwaxia [17], articulated tommotiid scleritomes [18], tube-dwelling lophophorates [19] and a sachitid (stem aculiferan) with a radula closely resembling that found in the molluscan crown group [20]. Although these previous fossil discoveries have added much to our understanding of early lophotrochozoan evolution, the uncertainty in the phylogenetic position of machaeridians obfuscates their importance for understanding the evolution of annelids and their close relatives.

Here we describe a newly discovered complete scleritome of P. bengtsoni. Unlike the previously described specimen of this species, the new scleritome is complete anteriorly and demonstrates the possession of pharyngeal jaw elements.

2. Material and methods

The single specimen of P. bengtsoni preserving jaws was collected near the top of the southern part of Tigzigzaouine Hill, approximately 20 km northeast of Zagora, southeastern Morocco. The deposits from which the fossil was collected belong to the Fezouata Shales [21], Araneograptus murrayi biozone, Upper Tremadocian (Lower Ordovician). The specimen is deposited in the National Museum, Prague, Czech Republic.

Specimens were photographed dry and uncoated, with lighting provided by a Schott KL 1500 fibre-optic light source with movable polarizers fitted at the end of the goosenecks; a Cokin XPro X164 circular polarizer was mounted on the camera lens and crossed with the light source polarizer to maximize contrast. The part was lit from the northwest while the counterpart was illuminated from the southwest and mirrored horizontally in Adobe Photoshop CC 2015.4 to create a false-positive relief image, facilitating direct comparison of part and counterpart. Photographs were taken with a Hasselblad H4D-200MS medium-frame digital single-lens reflex camera and operated remotely in six-shot mode through Hasselblad Phocus 8.2.1 software to acquire images of 200-megapixel resolution using a Hasselblad HC Macro 4/120 mm II lens stopped down to f/9.5. Lens distortion was corrected using Hasselblad Phocus 8.2.1 software. Stacks of 11 images for each part and counterpart were taken in aperture priority mode, with manual focusing through the focal plane. After exporting the FFF-format digital negatives to TIFF from Hasselblad Phocus 8.2.1, the photographs were stacked in Zerene Stacker Pro 1.04 Build T201602151850 (64 bit) using the PMax pyramid stack algorithm. The stacked images were then post-processed in Adobe Photoshop CC 2015.4. The photographs were sharpened, followed by removal of the background. Levels were then manually balanced while holding down the ‘alt’ key to prevent clipping of pixels; the grey level was always retained at 50%. The high-resolution images were down-sampled into lower resolution for use in the plates. Specimens from Mazon Creek were photographed dry using cross polarized light using a Nikon D90 with a Nikon AF-S VR Micro-Nikkor 105 mm f/2 IF-ED lens or an AF Micro-Nikkor 60 mm f/2.8 D lens.

To aid imaging of the jaw elements, latex replicas which were blackened with iron oxide were made and then coated using ammonium chloride and photographed using low angle illumination and a Leica stereomicroscope.

A camera lucida drawing of the part was made using a Leica MZ16A stereomicroscope with a variety of lighting directions. This drawing was then digitized using Adobe Photoshop CS6 and layered over an image of the counterpart. Details missing or ambiguous in the part were then added to create a composite interpretive drawing incorporating all preserved morphological information.

Three-dimensional tomographic datasets for jawed polychaetes were rendered using Drishti [22]. These datasets include Harmothoe imbricata [23], Hediste diversicolor [24] from the published literature and a new scan of Marphysa sp.

Our morphological dataset is derived from recent morphological matrices that focused on annelids [25] and molluscs [20] with additional characters for brachiopods and related taxa [19]. Phylogenetic analyses were performed using MrBayes 3.2.6 using the mkv + gamma model [26]. Two parallel analyses with four chains ran for 20 000 000 generations, with the first 25% of samples discarded as burn-in. Results were summarized as majority rules consensus trees.

The morphological character matrix and computerized tomography (CT) datasets are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.294g179 [27].

3. Results

The scleritome is articulated and almost complete (figure 1), with only posterior segments missing (possibly owing to disarticulation). The fifth outer shell plate and sixth inner shell plate are also missing owing to breakage and non-preservation. The specimen shows evidence of only one differentiated anterior plate (compared with two anterior plates in stratigraphically younger Plumulites species [7]). This anterior plate is followed by an inner shell plate, which is in turn followed by a strict alternating pattern of outer and inner shell plates. The scleritome is approximately 58 mm long, preserving an inferred total of nine inner and outer shell plates (figure 1).

Figure 1.

Whole scleritomes of Plumulites bengtsoni (National Museum, Prague, Czech Republic), part + counterpart. (a) Part, (b) counterpart, specimen lit from the southwest and horizontally mirrored and (c) interpretive drawing based on a camera lucida drawing of the part, with additional observations from the counterpart. Abbreviations for plates: ax, anterior plate; ix, inner shell plate; ox, outer shell plate, where × indicates the sclerite number in a given sclerite series. Note that the first anterior shell plate does not necessarily indicate the position of the first segment. (Online version in colour.)

A pair of structures in relief are displaced approximately 1 mm from the anterior margin of the anteriormost shell plate pair (figures 1 and 2d–f) and are visible in both the part (figures 1a and 2a,g,h) and counterpart (figure 1b). These structures are in close association with each other and preserve fine annulations consistent with growth lines. They are triangular in outline, with straight medial and posterior edges, with a lateral edge that is strongly curved. Each element is approximately 250 µm in both length and width and the midline separating them runs sub-parallel to the long axis of the scleritome. The relative size compared to the scleritome, their anterior position and orientation of these structures are consistent with them representing paired jaw elements.

Figure 2.

Jaws and the anterior region of Plumulites bengtsoni. (a) Close up of anterior region, part. (b) Camera lucida drawing of anterior region of the part. (c) Anterior region, counterpart, specimen lit from the southwest and horizontally mirrored. (d) Latex replica of the anterior region of the counterpart, whitened with ammonium chloride sublimate. (e) Close up of jaw in latex replica of counterpart, whitened with ammonium chloride sublimate, arrowheads indicate the position of prominent growth lines. (f) Jaw in specimen, counterpart. (g) Close up of jaw in latex replica of part, whitened with ammonium chloride sublimate. (h) Jaw in specimen, part. (Online version in colour.)

Given that paired jaw elements have a restricted distribution among extant annelids, their discovery has implications for the systematic position of machaeridians within Annelida. Our phylogenetic analysis (figure 4) recovers the monophyly of Machaeridia, with plumulitids recovered as paraphyletic with respect to the clade composed of Lepidocoleidae and Turrilepadidae (Cuniculepadida [28]). Machaeridians are recovered in the crown group of Phyllodocida, in the total group of Aphroditiformia. The relationships recovered are otherwise similar to other analyses based on iterations of the datasets used to assemble our matrix. Cambrian fossil polychaetes are recovered in the annelid stem group as previously [25], and Cambrian and Ordovician sachitid/halkieriid fossils are crown-group molluscs recovered in the stem group of Aculifera [20]. Tommotiids and related taxa are recovered among the total group of Brachiopoda, consistent with previous analyses [19].

Figure 4.

Results of phylogenetic analysis incorporating machaeridians and numerous extinct and extant trochozoans (361 characters for 141 taxa). Analysis performed under the mkv + gamma model in MrBayes 3.2.6. Numbers at nodes are posterior probability and the scale bar is expressed in number of substitutions per site. Certain clades are collapsed here for clarity, see the electronic supplementary material, figure S1 for full results. Silhouettes from PhyloPic (credits: Michelle Site, B. Duygu Özpolat, Stanton F. Fink/Michael Keesey). (Online version in colour.)

4. Discussion

The new observations reported above for P. bengtsoni add key candidate synapomorphies shared with aciculate polychaetes and thus justify a reassessment of the phylogenetic position of machaeridians. A phyllodocidan affinity of machaeridians is supported by the presence of an alternating dorsal cirrus pattern and paired jaw elements while a close relationship to scaleworms is supported by the presence of elytron-like expansions of the dorsal cirrus structures in the form of serially repeated shell plates.

Jaws are a common feature of aciculate polychaetes and are readily recovered as scolecodonts, microfossils extracted from acid maceration of sedimentary rocks [29]. Jaws occur in two of the three clades of aciculates, the Eunicida (figure 3a) and the Phyllodocida (figure 3b–i), but not in the Amphinomida [30]. The jaws of Eunicida and Phyllodocida probably do not have a common origin as they do not develop from the cuticle in the same way, have a different chemical composition and function differently [31–34]. Thickened cuticle is also present in the pharynx of amphinomids [35], which is used to rasp prey items. This cuticle may be more robustly sclerotized in certain fossil amphinomids from the Carboniferous [36] in which the cuticular ridges of the pharynx are preserved in high relief. Consequently, although the last common ancestor of aciculates may not have possessed discrete jaw elements, sclerotization of the cuticle of the pharynx may have a single origin that was shared by the common ancestor of Aciculata.

Figure 3.

Comparative images of jaws in extant and fossil aciculate polychaetes. (a) Marphysa sp., CT scan, views from right to left: dorsal, ventral, left-lateral. (b) Hediste diversicolorm, CT scan, views from left to right: frontal, dorsal, ventral. (c) Glycera oxycephala, lateral view (d) Glycera oxycephala, frontal view (e) Goniada maculata, pharynx exposed by dissection. Images in (c–e) provided by Markus Böggemann. (f) Harmothoe imbricata, CT scan, views from left to right: frontal, left lateral, dorsal. (g) Hystriciola deliculata from Mazon Creek ROMIP 47966. Arrow indicates position of jaws. (h) Close up of (e) showing pharynx and jaws. (i) Dryptoscolex matthiesae ROMIP 48542. Arrowheads indicate the position of jaws. (Online version in colour.)

The jaws of eunicidans are complex, composed of multiple elements, including a dorsal set of multiple maxillae, which may or may not be paired (figure 3a) [37]. This is complemented by a ventral set of mandibles (figure 3a), with two pronounced cutting plates, which are mineralized in Labidognatha (Eunicidae and Onuphidae) and Lumbrineridae [37]. Phyllodocidan jaws are simpler, consisting either of a single lateral pair of elements (Nereidae (figure 3b, middle), Hesionidae, Chrysopetalidae), a ring of jaws in Glyceridae (figure 3c,d) and Goniadidae (figure 3e) or four dorsoventrally biting jaws (Aphroditiformia (figure 3f)). The jaws of these groups can be everted outside of the body and are used to grasp food items, either during scavenging or active predation [38].

A key difference between the jaws of phyllodocidans and eunicidans is that the maxillae of the latter are periodically shed and replaced [34,39], whereas those of the Phyllodocida grow throughout life and consequently display fine growth lines [34]. The mandibles of eunicidans do grow continuously and also show growth annulations, which are visible in fossil specimens from Konservat-Lagerstätten [40]. However, eunicidan mandibles are only a single element, unlike the multiple paired elements of P. bengtsoni. The presence of growth lines in multiple jaw elements in P. bengtsoni excludes the possibility that they are homologous with the jaws of Eunicida, instead inviting a comparison with the diverse jaws of phyllodocidans.

A single pair of jaw elements is a common condition in Phyllodocida, occurring in Nereididae (figure 3b), Nephtyidae, Hesionidae and Chrysopetalidae [38,41], a group that may be monophyletic, or paraphyletic with respect to other phyllodocidans. The presence of four jaw elements occurs in Glyceridae and aphroditiforms. In Glyceridae, the four jaws form a ring at the tip of the pharynx when everted and are not in closely associated pairs like the jaws of P. bengtsoni. In scaleworms, the jaws occur in two dorsoventrally biting pairs (figure 3f) that are closely associated, similar to the single set of paired elements preserved in P. bengtsoni. Assuming that the anteriormost shell plate belongs to either the first or second segment, the position of the jaws suggests that the specimen is preserved with the pharynx everted, displacing the jaws anterior to the head. Preservation of fossil scaleworms with the pharynx everted occurs frequently at Mazon Creek (Upper Carboniferous, Illinois, USA), such as in the scaleworm Hystriciola deliculata (figure 3g,h), in which the everted pharynx is approximately 26% of the total length of the specimen. Although only two elements are visible in this specimen of Hystriciola, it is possible that other elements remain entombed in the matrix as is common in the Mazon Creek scaleworms [42]. Consequently, although only a single pair of elements is visible in P. bengtsoni, we cannot easily exclude the possibility that an additional pair is present but not observed as it would be immediately ventral of the visible jaws and therefore buried in the matrix of the part. The close association of the paired jaw elements in P. bengtsoni is similar to the condition in extant and extinct scaleworms, including the Carboniferous Hystriciola (figure 3g,h), Dryptoscolex (figure 3i) and Harmothoe imbricata (figure 3f).

The analysis included herein (figure 4) incorporates these recent discoveries into a taxonomically inclusive analysis of molluscs, brachiopods, sipunculans and annelids. The taxon sample includes extant and fossil trochozoan taxa with which the machaeridians have historically been considered closely related, including aculiferan molluscs, tommotiids and other putative stem-group brachiopods, and annelid taxa that encompass the crown and stem group of Annelida. The recovery of monophyletic machaeridians deeply nested within the crown group of annelids confirms that the origin of the machaeridian scleritome is unrelated to the complex multi-element scleritomes present in other Palaeozoic fossil taxa, such as tommotiids and halkieriids. The hypothesis wherein machaeridians were placed in the annelid stem group [10] is rejected by analysis of our dataset with strong support of 0.95 for the monophyly of Aphroditiformia including Machaeridia (figure 4).

The placement of machaeridians in the crown group of Phyllodocida substantially reduces the length of the ghost lineage of this diverse group of polychaetes. The sister group of the Phyllodocida, the Eunicida, is well documented from the Early Ordovician onwards [29] but unequivocal records of the Phyllodocida were not known until the Middle Devonian fossil Arkonips [43]. The presence of crown-group phyllodocidans in the Ordovician confirms that major cladogenetic events in this taxon had already occurred by this time, as is the case in Eunicida, in which the majority of family-level divergences had occurred by the Middle or Late Ordovician [28,37].

Data accessibility

Datasets for microCT scans and the character-taxon matrix are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.294g179 [27].

Authors' contributions

L.A.P., G.D.E. and J.V. conceived the study, analysed and interpreted the fossils and contributed to the scoring and construction of the character matrix. L.A.P. made the figures, rendered the CT data, performed the phylogenetic analysis and wrote the initial draft of the manuscript with input from all co-authors. D.S. scanned the specimens of Hediste and Marphysa and assisted with the rendering of the CT data.

Competing interests

We declare no competing interests.

Funding

L.A.P. was funded by NERC grant no. NE/L501554/1 and a YIBS postdoctoral fellowship at Yale University.