A Silurian short-great-appendage arthropod

A new arthropod, Enalikter aphson gen. et sp. nov., is described from the Silurian (Wenlock Series) Herefordshire Lagerstätte of the UK. It belongs to the Megacheira (=short-great-appendage group), which is recognized here, for the first time, in strata younger than mid-Cambrian age. Discovery of this new Silurian taxon allows us to identify a Devonian megacheiran representative, Bundenbachiellus giganteus from the Hunsrück Slate of Germany. The phylogenetic position of megacheirans is controversial: they have been interpreted as stem chelicerates, or stem euarthropods, but when Enalikter and Bundenbachiellus are added to the most comprehensive morphological database available, a stem euarthropod position is supported. Enalikter represents the only fully three-dimensionally preserved stem-group euarthropod, it falls in the sister clade to the crown-group euarthropods, and it provides new insights surrounding the origin and early evolution of the euarthropods. Recognition of Enalikter and Bundenbachiellus as megacheirans indicates that this major arthropod group survived for nearly 100 Myr beyond the mid-Cambrian.

Some so-called short-great-appendage arthropods (¼Megacheira [17]), such as leanchoiliids, are characterized by a first (great) head appendage with a short peduncle connected by a knuckle/elbow joint to a distal 'claw', the three podomeres of which each extends distally into a long flagellum [18,19]. Megacheirans have only been recorded from Cambrian deposits. Here, we describe a new genus and species of megacheiran with such a great-appendage morphology: Enalikter aphson from the Silurian Herefordshire fauna, representing another major arthropod group to be recognized from this Lagerstätte. Fossils from exceptionally preserved lower Palaeozoic biotas, such as the Herefordshire example, have the greatest potential for revealing the earliest stages of arthropod diversification, the stem region of the arthropod phylogenetic tree. Phylogenetic analysis of Enalikter and the re-evaluated Devonian taxon Bundenbachiellus refines the topology of this stem region, providing new insights into immediately pre-euarthropod crown-group morphologies.

Material and methods
Specimens of Enalikter were serially ground at 20 mm intervals. Each ground surface was captured digitally and, through using the SPIERS software suite, the resulting tomographic dataset was rendered and studied as a three-dimensional virtual fossil [20,21]. Interpretation on-screen of the virtual fossils was facilitated by variable magnification, unlimited rotational, virtual dissection and stereoscopic-viewing capabilities; they were also examined through hard-copy images.
Analysis of the phylogenetic position of Enalikter and Bundenbachiellus was performed using a modified version (see the electronic supplementary material, note S1) of the panarthropod character matrix of Legg et al. [22], which represents the most comprehensive morphological matrix available. The Legg et al. analysis included recent re-interpretations of head appendage innervation [23,24], added to which we have now also taken into account the subsequently published conclusions of Tanaka et al. [25]. A dataset of 314 taxa and 753 characters was analysed using maximum-parsimony in TNT v. 1.1 [26], which generated 36 most parsimonious trees (MPTs). The strict consensus tree is provided (see electronic supplementary material, figure S1), and also a summary of the topologies from the phylogenetic analyses (figure 2; electronic supplementary material, figure S2).
Holotype: Oxford University Museum of Natural History (OUMNH C.29631) complete outstretched specimen, length 24.4 mm from anterior margin of cephalic shield to posterior margin of telson (figure 1a-c,k,o,x).
Datasets from serial-grinding tomography of the specimens are housed in the Oxford University Museum of Natural History.
Other species. None. Generic and specific diagnosis. Head shield subrectangular, lacking a narrow, raised margin. Head bearing a boss-like structure ventromedially, extending anteriorly into a curved whip-like process. Trunk limb exopods with long, narrow, non-overlapping filaments lacking spines. Telson with a needle-like process medially, and two pairs of blade-like processes laterally.
Description. The head shield is about 1.5 times as long as wide, subrectangular in outline and dorsoventrally shallow, partially covering the first trunk segment (figure 1e,j). Surface sculpture is apparently lacking.
Appendage 1 originates at about 20% of the head length (figure 1h). It is uniramous, comprising a short peduncular section of probably two podomeres, plus three closely originating and tapering flagella (podomere numbers unresolved). One flagellum is about half as long as the other two-the ventralmost on both the best-preserved, outstretched specimens (figure 1e,h); an elbow/knuckle joint is lacking between peduncle and flagella. Appendage 2 is biramous and originates at about 55% of the head length. The limb base is very short, anteroposteriorly flattened, and bears a conspicuous spine-like endite. The endopod is finger-like, evenly tapered, and comprises at least three podomeres; the exopod is similar but much more slender (podomeres unresolved), and slightly shorter (figure 1p). Appendage 3 arises at about 85% of the head length. It is biramous and similar to appendage 2 but slightly larger, with a more robust, blunter endite; the first of the (at least four or five) podomeres of the endopod bears a median ridge (figure 1s).
Eyes are absent. Ventromedially, a boss-like structure (figure 1d,h,r) extends anteriorly into a recurved, whip-like process that is subconical proximally, more slender and tapering distally, and presumably flexible, although in all three specimens it ends beneath the mouth. The more ventral part of the boss is subcylindrical and terminates in a flat, disc-like surface with a central subcircular mouth that faces posteroventrally. A short, narrow, sediment-filled space immediately inside the mouth is interpreted as a buccal cavity and/or very short oesophagus (figure 1r); it connects sharply with a broader, sediment-filled cavity, interpreted as the stomach. The latter is directed dorsally before bending posteriorly in a J-shape into the intestine/midgut (figure 1q,r,b1).
The rest of the body, comprising a trunk and a telson, is about 14 times as long as wide. The trunk, which consists of 12 segments, is roughly parallel-sided, and is subcircular in transverse section in OUMNH C.29632 (figure 1m,u-w,a1), though both outstretched specimens display dorsoventral compression (see Discussion). Each tergite is dome-like (figure 1t,v) and lacks paratergal folds (tergopleurae). The sternite is a subcircular to subrectangular button-like structure, with a central node and a tuberculate marginal rim (figure 1f,i,u). At the anterior and posterior margin of each tergite and its associated sternite, there is a prominent, transverse, tuberculate ridge that encircles the trunk. In between these occur weaker, less persistent ridges (figure 1m,u,v,a1) representing articulations, which in places display a wedged concertina-like form, indicating segment pinching (figure 1m,u). These areas presumably represent arthrodial tissue, which enabled lateral flexure up to at least 908 between segments (figure 1t-v). Evidence of vertical trunk flexure is limited, and is at most gently upwards posteriorly (figure 1b). The gut is preserved discontinuously along the narrow trunk, but there is no evidence of midgut glands. Transverse, soft-tissue traces are evident posteriorly, some (? tendinous bars) coinciding with segment boundaries (figure 1x). rspb.royalsocietypublishing.org Proc. R. Soc. B 281: 20132986 The first trunk appendage (figure 1d,e,h) is biramous, with a short, stout, simple limb base that lacks endites. The endopod is stenopodous, similar but larger than that of head appendage 3, with at least six or seven podomeres, the second(?) of which is raised medially. The exopod consists of a slender, tapering shaft bearing at least eight filaments (each probably from a separate podomere). The filaments are long, slender, non-overlapping and apparently suboval in section; the most proximal is the stoutest, and they become shorter distally. Trunk segments 2-12 each bear a biramous appendage pair similar to the first trunk appendage (figure 1a-c). Some endopods preserve two slender spinose/setal terminal projections, which were presumably present on all trunk limbs. The exopods are recurved dorsomedially in both outstretched specimens. They preserve from 11 to 17 filaments (see figure 1o for a typical biramous limb). These filaments are long enough to overlap at least partially those of the following appendage ( figure 1b). The trunk appendages increase in size from the first to about the fifth, and are similar in length on successive segments ( figure 1a,g). The endopods of the more posterior trunk appendages are slightly more slender.
The telson is ovoid in dorsal view (figure 1a,l) and about 1.3 times as long (medially) as wide; in lateral view, it is wedgelike, increasing in height posteriorly (figure 1b,n,w). Ventrally, a slightly raised, posteriorly narrowing subtriangular axial region is bounded by a very weak abaxially convex furrow (figure 1u). A narrow, prominent tuberculate ridge and parallel furrow, similar to those on the trunk segments, encircle the anterior margin of the telson. Posteriorly, the telson bears two pairs of long, blade-like processes (figure 1l,n); each originates adjacent to the midline, tapers to a point, and is laterally flattened and suboval in section. The dorsal processes project posterodorsally at about 308. The ventral ones curve evenly dorsally through about 608, their tips crossing immediately outside those of the dorsal pair. There is no evidence for or against mobility in any of these processes. A medial, needlelike process projects posterodorsally from between the ventral pair. The anus lies posteroventrally, as indicated by a faecal stream ( figure 1w,z,a1). The telson extends parallel to the trunk (figure 1w) or may be inclined upwards at about 308 (figure 1b).

Discussion
The preservation of Enalikter (figure 1; electronic supplementary material, figure S4) in full three-dimensional form is unique for a stem euarthropod. The trunk of OUMNH C.29632 (figure 1 m,t -w,y,z,a1) is subcircular in cross-section, it bends laterally through 1808, and the exopod filaments curve around to hug the bend, in a lowered, presumed 'in repose' position ( figure 1t,y). The other two specimens (figure 1a,g) have a flatter trunk section, yet retain upstanding to outstretched limbs, with straight to slightly sinuous, vertically radiating exopod filaments ( figure 1k,o). Operation of the trunk and filaments by hydraulic pressure might account for such differences of inflation and disposition, though equally it might reflect the early onset of decay.
The pyritized but much larger arthropod (up to 228 mm [30]) Bundenbachiellus giganteus [29] (¼ Eschenbachiellus wuttkensis [31]; see [30]) from the Lower Devonian Hunsrü ck Slate is close in overall morphology to Enalikter. Insights from the new Silurian taxon are used here to reinterpret the younger Devonian form. Only one of the two specimens of Bundenbachiellus preserves the head ( [31], text figures 11-13; electronic supplementary material, figure S3), which was previously interpreted as bearing five appendages. A comparison with the better-preserved Enalikter indicates that the structures interpreted by Briggs & Bartels ( [31], p. 293) as a uniramous first (evident only on the left side) and a biramous second appendage, together represent a single triflagellate limb. It is likely that the following two (more posterior) head appendages of Bundenbachiellus were biramous, although only the endopod is clearly evident (see electronic supplementary material, figure S3). Comparison with the head of Enalikter suggests that the appendage interpreted as a fifth head limb in Bundenbachiellus may belong to the trunk. There would then be 12 pairs of biramous appendages in the trunk of Bundenbachiellus (although their correspondence to tergites is uncertain), as in Enalikter, and the posteriormost spines/appendages could be interpreted as telson processes (rather than a pair of spines and a caudal furca) such as those in Enalikter. Bundenbachiellus differs from Enalikter, however, in a number of ways: the head shield was semicircular (not subrectangular), surrounded by a narrow raised margin; there is no evidence of a whip-like process ventrally on the head; the trunk exopod filaments are leaflike (not linear) structures with fine spines on their inner margins; and there is no evidence of a medial, needle-like process on the telson. Additionally, the Devonian species is an order of magnitude larger than the Silurian one.
Enalikter and Bundenbachiellus fall in a clade of short-greatappendage (¼megacheiran) arthropods [32] that includes Leanchoilia from the lower Cambrian of Chengjiang and the middle Cambrian Kaili Lagerstätte, China, and the Burgess Shale, Spence Shale and Marjum Formation of North America; Alalcomenaeus from Chengjiang and the Burgess Shale; Actaeus from the Burgess Shale; and Oestokerkus from the lower Cambrian Emu Bay Shale, Australia [32][33][34][35][36][37] (figure 2; electronic supplementary material, text S1 and figure S1). Specifically, Enalikter is recovered in a clade (Enaliktidae) together with Bundenbachiellus. More broadly, it falls under a clade that is the most derived in the euarthropod stem and sister to Euarthropoda, and which also includes the megacheirans Haikoucaris and Parapeytoia from Chengjiang, and Yohoia from Burgess.
While the tergopleurae are reduced in some stem euarthropods-for example Haikoucaris [18]-enaliktids appear to be unique among stem euarthropods in lacking them entirely. Enaliktids are also distinguished among megacheirans in their lack (loss) of the knuckle/elbow joint between the peduncle and podomeres of the 'claw' (flagella), a hallmark of other megacheirans [48] (although this feature is only weakly developed in at least one other purported megacheiran, Occacaris [19]). A remarkable feature of Enalikter is the long, posteriorly recurved, whip-like anterior process on the head, which may be analogous to the spinose hypostomal structure found in parasitic eucrustaceans [49] (electronic supplementary material, text S2). The ventromedial, subventrally projecting boss-like feature to which the process is attached recalls similar structures interpreted as hypostomal homologues in the stem mandibulates Agnostus, rspb.royalsocietypublishing.org Proc. R. Soc. B 281: 20132986 Henningsmoenocaris and Martinssonia [50]; as in those taxa, a discrete, fully sclerotized hypostome is lacking in Enalikter. The flat, wide, circumoral disc-like surface in Enalikter bears comparison, variously, with the mouth/'Peytoia' cone of the panarthropod lobopodians Pamdelurian and Opabinia, and stem euarthropod radiodontids such as Anomalocaris and Peytoia, and the great appendage arthropod Parapeytoia [51 -55] (electronic supplementary material, figure S1). In those taxa, however, the oral cone surface is rigid and plated, unlike the disc surface of Enalikter, which lacks evidence of plates and was presumed fleshy (see electronic supplementary material, text S2). Enalikter inhabited the outer shelf/upper slope of the Anglo-Welsh basin, where water depths might have been up to some 200 m [2]. It is likely to have been a benthic or nektobenthic scavenger/detritivore (see electronic supplementary material, text S2).
Recognition of Enalikter and Bundenbachiellus in Silurian and Devonian rocks indicates that members of the stem clade Leanchoiliida survived for nearly 100 Myr (75 and 97 Myr, respectively [56]) after the mid-Cambrian Leanchoilia?   [34]), the hitherto stratigraphically youngest known short-great-appendage arthropod. Enalikter and Bundenbachiellus are some 55 and 77 Myr, respectively, younger than the next youngest stem euarthropod, anomalocaridids from the lower Ordovician Fezouta Lagerstätte (ca 480 Myr BP) of Morocco [57]; and the enaliktids represent only the second record of stem euarthropods in Silurian or Devonian strata, the other being that of Schinderhannes from the Hunsrü ck Slate [47]. Data on Enalikter and Bundenbachiellus highlight the importance of rare Silurian and Devonian Konservat-Lagerstätten for revealing the much later, mid-and upper Palaeozoic history of groups such as megacheirans that have previously been considered to be restricted to the Cambrian; more accurate knowledge of their true stratigraphic range is dependent on these critical taphonomic windows. Our study also highlights the advantage available in combining morphological data from different types of exceptional-preservation deposits.