New insights into the Devonian sea spiders of the Hunsrück Slate (Arthropoda: Pycnogonida)

Studied material

We accessed and studied in detail 46 slabs preserving 63 fossil specimens (listed in Material S1) representing at least four species: Palaeoisopus problematicus (51 specimens), Palaeopantopus maucheri (three specimens), Flagellopantopus blocki (one specimen), and Pentapantopus vogteli (three specimens, with possibly an additional one) and some undetermined pycnogonid material (four specimens). We could not locate the holotype and only specimen known of Palaeothea devonica and as such, it could not be included in our study. Part of the fossil material described here has already been presented in previous works while some other fossils are figured and described for the first time (refer to Material S1). The fossils are preserved as flattened slates with variable levels of pyritization. They are hosted in the collections of the Museum für Naturkunde, Berlin (collection numbers MB-A-), the Institut für Geowissenschaften, Section Palaeontology, University of Bonn (collection numbers IGPB-), the Naturhistorisches Museum Mainz (collection numbers NHMMZ PWL), and the Bayerische Staatssammlung für Paläontologie und Geologie, Munich (collection numbers SNSB-BSPG-).

X-ray microtomography

Seventeen slabs were selected for X-ray microtomography, scanned either with a Nikon XTH 225ST at the University of Bristol, a Phoenix|x-ray v|tome|xs at the Rheinische Friedrich-Wilhelms-Universität Bonn, or a Yxlon FF85 Modular at the Museum für Naturkunde in Berlin. The current during the scan ranged from 70 to 300 µA, voltage from 30 to 215 kV and exposure time from 1 to 2 s. A tungsten reflection target and copper and tin filters were used in some cases. The voxel sizes obtained ranges from 10 to 126 µm depending primarily on the size of the slab hosting the specimen (see Material S2 for detailed parameters for each scan). Often the poor contrast between the matrix and specimens (where not pyritised) or the resulting streak artefacts (where pyritised) precluded segmentation. As such, we instead present microtomography data using two-dimensional maximum intensity projections over a few slices. The software ORS Dragonfly (build 2021.3.0.1087, Montreal, Canada) was used to reconstruct these projections.

Reflectance transformation imaging

Reflectance transformation imaging (e.g., Béthoux, Llamosi & Toussaint, 2016; Decombeix et al., 2021; Sabroux et al., 2023) was used to image the surface of the fossils. This was performed for 46 slabs (60 specimens), using an RTI-dome, i.e., a hemispheric rig placed over the fossil with an automated lighting series of LEDs evenly spaced around its concave surface. Here, we used a 32-cm diameter light dome equipped with 52 LEDs distributed over three rings. Light orientation was recorded on each image by a reflective metal ball. Photos were taken with a Nikon D850 mounted with NIKKOR 40, 60 and 105 mm lenses or a Canon EOS 700D digital camera, with Canon 50 mm and Canon EF 100 mm macro lenses (see Material S3 for details).

Images were computed with the software RTIbuilder (Cultural Heritage Imaging, San Francisco, CA, USA) using the Highlight Based (HSH Filter) operation sequence. The resulting models were then visualised using the RTIviewer software (Cultural Heritage Imaging, San Francisco, CA, USA). RTIviewer allows the user to vary the direction of illumination at will. “Specular enhancement” allows the contrast between illuminated and shadowed surfaces of the item to be enhanced by estimating the normal for each pixel and using this to render surfaces using synthetic specular highlights. Finally, “normals visualization” enables visualisation of all normals at once with false colours depending on the position of the light source that most strongly illuminates a given part of the item.

Systematic palaeontology

The synonymy lists were written following the recommendations of Matthews (1973). The terminal claw of appendages is counted here as one podomere, and podomere morphological terminology adopted is the same as in Sabroux et al. (2023) and Siveter et al. (2023).

Class Pycnogonida Latreille, 1810

Palaeoisopus problematicus Broili, 1928

Figures 1–19.

v* 1928 Palaeoisopus problematicus: Broili, plate 1.

1932 Palaeoisopus problematicus: Broili, pp 45-54, figs. 1-5.

vp 1932 Palaeoisopus problematicus: Helfer, pp 70, 71, figs. 53, 55 (Helfer, 1932).

vp 1933 Palaeoisopus problematicus: Broili, pp 33-46, plates 1-5 (Broili, 1933).

1944 Palaeoisopus problematicus: Størmer, p. 146 [text], fig. 29a (Størmer, 1944).

1954 Palaeoisopus problematicus: Hedgpeth, pp 197-199 [text], 201 [text], 202 [text], fig. 3.

1955b Palaeoisopus problematicus: Hedgpeth, pp P171-P173 [text], fig. 123 (Hedgpeth, 1955b).

1957 Palaeoisopus problematicus: Dubinin, p. 881, fig. 1А-В.

1958 Palaeoisopus problematicus: Tiegs & Manton, p. 321 [text], fig. 18a (Tiegs & Manton, 1958).

vp 1959 Palaeoisopus problematicus: Lehmann, pp 98-101, tabs 10, 11.

1978 Palaeoisopus problematicus: Hedgpeth, pp 23-25 [text], 30 [text], 33 [text], figs 1, 2B (Hedgpeth, 1955a).

vp 1980 Palaeoisopus problematicus: Bergström et al., pp 10-31, 32 [text], 33 [text], 47-49 [text], figs. 1-23, 34 [affinities].

v 1998 Palaeoisopus problematicus: Bartels et al., pp 153-155, figs. 130-132.

v 2012a Palaeoisopus problematicus: Kühl et al., p. 70 [text], figs. 74, 75.

v 2012b Palaeoisopus problematicus: Kühl et al. p. 74 [text], figs. 74, 75. (cop. Kühl et al., 2012a, 2012b).

2017 Palaeoisopus problematicus: Südkamp, pp 77–79, figs. 118–120.

Examined material: HOLOTYPE: SNSB-BSPG 1928 VII 11: 1 juvenile–OTHERS: IGPB-AR-339: 1 sp., -AR-340: 1 sp., -M-142: 1 sp., -HS206:1 sp., -HS207:1 sp., -HS456: 1 sp., -HS457: 1 sp., -HS582: 2 spp., -HS636: 1 sp., -HS660: 1 sp., -HS694: 1 sp., -HS942: 1 juvenile, -HS1039: 3 spp.,– MB-A-46: 1 sp., -47: 1 sp., -288: 1 sp., -313: 1 sp., -3969-1: 1 sp. – NHMMZ PWL 1986/3: 1 juvenile, 1992/178-LS: 1 sp. 1994/54-LS: 1 juvenile, 1 sp., 1994/55-LS: 1 sp., 1994/56-LS: 1 sp., 1994/133-LS: 1 sp., 1995/17-LS: 1 sp., 1995/35-LS: 4 spp., 1996/18-LS: 1 sp., 1997/44-45-LS: 3 spp., 1998/122-LS: 2 spp, 1998/155-LS: 1 sp., 2000/46-LS: 1 juvenile, 2003/272-LS: 2 spp., 1 juvenile, 2008/141-LS: 1 sp., 2013/8-LS: 1 sp. – SNSB-BSPG 1932 I 63: 1 sp., 1932 I 67: 1 sp., 1967 I 306: 1 sp., 2021 IV 1: 1 sp.

Diagnosis. Body robust. Proboscis broad, about as long as the distance between the cephalic anterior margin and the posterior margin of the second trunk segment. Dorsal surface of the second and third trunk segments with tubercles disposed in a V, distal margin of all trunk segments thickened. Abdomen long, five-segmented, carrying an anus on the ventroproximal part of the ultimate segment, distal telson lanceolate. Chelifores massive. Palps with nine articles, the proximalmost podomere (coxa 1) divided in three rings, the distalmost ring individualised from the two proximalmost rings, sixth podomere (tibia) carrying a dorsal spur, no claw. Ovigers’ proximalmost podomere (coxa 1) divided into three rings, penultimate article with robust spine on distal margin, ultimate article a claw. Four pairs of legs consisting of nine (first pair of legs) or 10 (second to fourth pairs of legs) podomeres including a subchelate, terminal main claw; coxa 1 divided into rings; distal leg podomeres (patella to propodus) laterally flattened, second to fourth legs with pairs of long ventral setae.

Description and interpretation. Members of this species are very large for pycnogonids, reaching up to 400 mm in legspan, and are among the best-known fossils in the Hunsrück Slate; more than 80 specimens are known. The investigations of Bergström, Stürmer & Winter (1980) detailed the general morphology of Palaeoisopus problematicus, with a complete description of the trunk, the abdomen and the legs. We accessed 51 specimens, including not yet published material (Material S1), to compare and revise the descriptions of Bergström, Stürmer & Winter (1980; based on 57 specimens) in the light of new evidence.

The body of P. problematicus is composed of four trunk and five abdominal segments (Figs. 1, 2). The cephalon is fused with the first trunk segment as in modern pycnogonids. The fusion of the cephalon and the first trunk segment confers a roughly square shape. The first trunk segment carries the lateral processes to which the first pair of walking legs (WL1) articulates (Fig. 2D). These processes are approximately cylindrical in shape, with lateral thickening and ornamentation composed of rounded tubercles (Fig. 2D). Anteriorly and ventrally to the lateral processes, the cephalon also carries lateral extensions to which the palps articulate (Fig. 2D), their distal border broadened. Ovigers insert ventrally, at about the same level as the lateral processes of the first trunk segment (Fig. 3). The anterior margin of the cephalon on which the chelifores articulate also presents an ornamented thickening. The posterior margin of the cephalon is thickened and ornamented with tubercles.

The proboscis is visible on some ventrally preserved specimens (Fig. 3), the tip of it being posteriorly directed. Some specimens preserved laterally (Fig. 4) suggest that the proboscis is folded ventrally (its tip directed posteriorly), though the proboscis cannot be unambiguously identified in these specimens. If correct, this position is potentially a resting position and the proboscis was mobile, as seen in modern species presenting the same orientation of the proboscis (such as many modern Ascorhynchidae Hoek, 1881). The proboscis is tubular in shape overall, without clear proximal or distal narrowing. According to Bergström, Stürmer & Winter (1980) the dorsal surface of the proboscis is divided into three fields, suggesting that it is composed of three antimeres like extant species: this observation was made based on a single specimen that we could not examine.

The ocular tubercle region (Fig. 5) is positioned near the anterior margin of the cephalon. It has six rounded features (Figs. 5, 6). The anteriormost is roughly reniform, and there are two positioned postero-laterally to the first, surrounding a third central one. Posterior to these, are two further lateral rounded features with a broader base. These six features are concentrated in a region on the cephalon. More posteriorly, a seventh feature is also visible. Bergström, Stürmer & Winter (1980) proposed that this posteriormost feature is a posterior ocellus; these authors also interpreted the anterior, reniform features as an anterior eye, and the posteriormost pair of features as laterally positioned eyes (Fig. 6). If correct, this arrangement would be very different from extant species that instead present anterior and posterior eye pairs, and this interpretation does not explain the identity of the three features directly posterior to the “anterior eye”. It seems unlikely that any of these represent any eye at all, since eyes of modern sea spiders generally do not protrude from the cuticle, or only as a low convex lens (Brenneis, 2022) which is unlikely to preserve as markedly as the structures here following the dorsoventral compression typical of Hunsrück fossils. Many species in the extant genus Rhynchothorax Costa, 1861 (family Rhynchothoracidae Thompson, 1909) possess a disposition of more or less elongated tubercles very similar to the features seen in P. problematicus (Fig. 6; some of them can be absent in some species). We suggest that the features of P. problematicus are tubercles, homologous to those in Rhynchothorax. The two tubercles postero-lateral to the anteriormost correspond in position to the two lateral sense organs (sensu Lehmann, Heß & Melzer, 2017). These sensory structures have been identified in all sea spider families except for Pycnogonidae Wilson, 1878 (Brenneis, 2022). We cannot address whether these were already functional as lateral sense organs (indeed, their exact function remains unknown even in modern species; Lehmann, Heß & Melzer, 2017).

Trunk third and fourth segments have two to three pairs of large tubercles arranged in a V tapering posteriorly (Figs. 2, 7). These are followed posteriorly by thinner tubercles which are more variable in disposition and/or preservation. The fourth trunk segment carries two pairs of dorsal tubercles, and many thinner tubercles may be present more posteriorly. The distal margin of each trunk segment is thickened and presents additional ornamentation composed of about 10 tubercles along its length. The lateral processes are very short on the first trunk segment and are longer posteriorly. Their orientation is almost orthogonal to the trunk axis on the first trunk segment, and increasingly diagonal posteriorly, on the fourth trunk segment the axis of the lateral processes is almost parallel to the trunk. The distal margin of the lateral processes is thickened and ornamented with tubercles.

The abdomen (Figs. 8, 9) is attached on a posterior thickening of the fourth trunk segment. It carries four rectangular segments that are increasingly elongated distally, and a terminal fifth, lanceolate, very elongated segment (about 40% of the total abdomen length). The fifth abdominal segment carries a ventrally positioned anus at approximately a quarter of the total segment length (Figs. 8I–8P). For Bergström, Stürmer & Winter (1980), the anus indicates the posterior end of the fifth and ultimate abdominal segment, so that Palaeoisopus problematicus possesses five true abdominal segments, with the ultimate one fused with the post-anal telson; but it was suggested alternatively (e.g., Vilpoux & Waloszeck, 2003; Dunlop & Lamsdell, 2017) that the whole fifth segment actually corresponds to a telson, so that P. problematicus only has four abdominal segments, and uniquely presents an anus carried ventrally on the telson. At least the four distal abdominal segments (and probably also the first abdominal segment) carry short setae along their lateral margin. Thickened ridges are present medioventrally and mediodorsally. Dorsally each trunk segment has an anterior thickened margin. This thickened anterior margin may be present ventrally as well, as observed in specimen 1 of the slab IGPB-HS1039 (Figs. 8I–8L), although it is not visible in specimen IGPB-HS660 (Figs. 8M–8P). Both ventral and dorsal sides show a free arthrodial membrane (see Fig. 8). It is larger ventrally, giving the abdomen great ventral mobility (Figs. 9, 10A); dorsally, the arthrodial membrane is smaller, and provided limited dorsal mobility to the abdomen (Figs. 9D–9F), except at the abdomen’s base where dorsal mobility seems important (Figs. 9A–9C). Each abdominal segment has setae along its lateral margins, which in some cases (Figs. 9G–9I) appear to be relatively long. They are arranged in at least one, possibly several, rows. Low numbers of additional setae can sporadically be found on the ventral or dorsal posterior margins of abdominal segments (Figs. 8M–8P, 9G–9L). In specimen IGPB-HS660 (Figs. 8M–8P) there is clear evidence of a pair of short setae on the ventroposterior margin of the fourth abdominal segment.

The chelifores (Figs. 10, 11) have prominent chelae. Despite their often-superb preservation, we are puzzled by their segmentation. Bergström, Stürmer & Winter (1980) counted (with some hesitation) five podomeres, three of which are scapes: this differs significantly from what is observed in any modern pycnogonid, which generally have three podomeres (including one scape), although sometimes one, two or four (with one or two scapes) are present, but never five. From the material we examined, we could not confidently identify five podomeres; four (including two scapes) is the minimal number and seems as likely, if not likelier, than five. The robustness of this interpretation is, however, limited by the ambiguity of the ornamentations of the scapes: it is often very difficult to determine whether ornamentation demarks the margin of a segment, or is in more of a median position (see highlighted margins and lines in Figs. 10 and 11). The proximal margin of the chelifores’ first podomere has a ridge ornamented by tubercles, roughly perpendicular to the segment axis; more distally, is another, diagonal ridge (the distal margin of the first scape?) that is ornamented with large tubercles, some of which have marks of possible setal insertions (Fig. 11J). This is followed by another ridge (the proximal ridge of the second scape?) roughly perpendicular to the axis of the purported second scape. Still more distally, is another ridge diagonal to the axis of the appendage. This may either represent a median ridge (two scapes) or the joint between the second and the third scapes (three scapes). It is followed distally by one final ridge–the distal margin of the ultimate scape. Because it is not possible to distinguish proximal and distal margins of the purported joints between the purported second and third scapes, and because it is more conservative (in light of chelifore morphology of extant pycnogonids), we favour the two-scapes hypothesis (Fig. 14A).

The palps (Figs. 10, 12) are preserved laterally and generally directed anteriorly. They are composed of two proximal rings (Figs. 10, 12D, 12E) which together form the proximalmost podomere following the interpretation of Siveter et al. (2023) for walking legs. The second podomere possesses an angular projection at its distal margin (Figs. 12A, 12B, 12D). The third and fifth podomeres are about 1.5 times to twice as long as the others. The seventh podomere typically has a dorsal spur, its cuticle sometimes appearing coarse (Fig. 12G) or bearing short setae (Fig. 10). There is no distal claw. The second to fifth podomeres have a thickened distal margin. In total, we count a total of 10 podomeres in the palp of P. problematicus. This number differs from the typical (extant) sea spider bodyplan, which has nine podomeres, or a reduction from this fundamental pattern (Sabroux et al., 2023). Even Haliestes dasos has nine palpal podomeres, excluding the terminal claw (Siveter et al., 2023). Typically, in modern species, the palps have a first short coxa 1, followed by a longer coxa 2 (Sabroux et al., 2023). Since the ovigers of P. problematicus have three coxal rings (see below), and palps only two, we hypothesize that the second podomere of the palp of P. problematicus could be the distalmost ring of coxa 1, that is individualised. We therefore propose to apply the morphological terminology of Sabroux et al. (2023) and Siveter et al. (2023) as presented in Figs. 14B, 14D. Whether this additional podomere was mobile does not seem to be addressable here.

The ovigers (Figs. 10, 13) are positioned ventrally, as in extant species. Ovigers are never completely preserved in any studied specimen, making it difficult to count podomeres. The limbs bear a distal claw (Figs. 13P, 13R, 13S). In specimen 3 of NHMMZ PWL 1995/35-LS (Figs. 13A, 13F, 13K, 13P), nine podomeres plus a podomere composed of three proximal rings (the coxa 1) can be counted; a part of the limb (from the seventh proximal podomere) is hidden under a leg. In the hidden part of the oviger, there is room for at least one more podomere, suggesting at least 11 podomeres (in agreement with Bergström, Stürmer & Winter, 1980). Also, the fourth podomere appears slightly larger than more distal ones. In the right oviger of specimen NHMMZ PWL 1994/133-LS (Figs. 13E, 13J, 13O, 13T), a similarly longer podomere is visible as the eighth most distal podomere. This supports the 11 podomeres hypothesis (fig. 14C). Most extant families (Ammotheidaec Dohrn, 1881; Ascorhynchidae Hoek, 1881; Callipallenidae Hilton, 1942; Colossendeidae Jarzynsky, 1879; Nymphonidae Wilson, 1878; Pallenopsidae Fry, 1978; Rhynchothoracidae Thompson, 1909; see, for example Sabroux et al., 2023) also have 11 podomeres, which has been suggested as the plesiomorphic state among Pantopoda (Bamber, 2007). We consider the count herein uncertain, however, due to the absence of any completely preserved and uninterrupted oviger. There is evidence of setae on the purported eighth podomere (Figs. 13C, 13H, 13M, 13R), on the fifth and sixth podomeres (Figs. 13A, 13F, 13K, 13P), and probably on some of the tarsal podomeres (Figs. 13B, 13G, 13L, 13Q). The distal margin of the last tarsal podomere (purported tenth podomere) forms a sort of vertical lamella ornamented with spines in at least some specimens (Figs. 13R, 13S), although this is not apparent in all specimens (Figs. 13A, 13F, 13K, 13P). Ovigers are one of the appendages in which sexual dimorphism, when present, is the most marked morphologically (e.g., Sabroux et al., 2023), and this could be also the case here, although a simple difference in preservation seems as likely.

The first pair of walking legs (WL1; Fig. 15) has four proximal rings (see also Figs. 2, 7) interpreted as coxa 1 by Siveter et al. (2023). These are followed by eight plain podomeres: the next podomere, coxa 2, is roughly trapezoidal, enlarging distally. It carries at least two ornamented ridges extending proximo-distally. Their exact position is difficult to interpret due to the flattening of the fossil that deformed the legs, but these are likely to be dorsal and anterior in position. The distal margin of coxa 2 is thickened. Coxa 3 is short, roughly cylindrical, its distal margin thickened markedly with a dorsal notch probably enabling the levator-depressor swing of the femur and patella. The femur is short, roughly triangular, its broader side ventral, marking a strong geniculation in the legs, and its distal margin thickened. The patella, tibia, tarsus, and propodus are strongly flattened laterally, as made clearly visible in specimens that have their legs preserved dorsally (Figs. 15L–15Q). The patella is characterised among these podomeres by its tapering proximal extremity. The ventral margin of the patella, tibia and tarsus distally presents a cluster of short setae near the insertion of the arthrodial membrane of the following podomere. Dorsally, these podomeres carry a (single?) row of short setae set on small tubercles. The propodus has at least two rows of denticles on its distal lamellar margin (Figs. 15I–15K). The claw is subchelate. In modern pycnogonids, the distal claw of legs (generally referred to as the main claw) is movable, so it is very much possible that the distal claw of P. problematicus WL1 was able to close on the distal indented margin of the propodus, making it a raptorial and/or cutting appendage.

Walking legs two to four (WL2-4; Fig. 16) differ from WL1 in presenting an additional podomere compared to WL1, as seen in Haliestes dasos (this additional podomere is referred to as a metatibia; Siveter et al., 2023). We propose that these two “additional” podomeres are homologous, and therefore, we follow the same morphological terminology. The coxa 1 presents three (WL2) or two (WL3-4) rings; it is shorter than broad. Coxa 2 is distinctly more elongated than in WL1, about 2.5 to 4.5 times as long as broad. It is ornamented like WL1, including rounded tubercles and short setae. Coxa 3 and the femur are short, about as long as broad, and the femur is triangular as seen in WL1. The patella is much more elongated than that of WL1: about twice as long as broad. The tibia, metatibia, tarsus, and propodus are all longer than broad. Podomeres from the patella to the tarsus presents two rows of long ventral setae (Figs. 16M–16P), as well as short dorsal setae. The propodus also carries setae, but it is unclear whether these are in one or several rows (Figs. 16E–16H). The terminal claw is subchelate.

Juveniles: six specimens of much smaller size are regarded as juveniles of this species, including the holotype (Figs. 1, 17, 18). They are often found in association with crinoids (Hapalocrinus frechi Jäkel, 1895 and Parisangulocrinus minax (Schmidt, 1934); see details in Material S1). Besides their smaller size, they differ from the adults in having a larger propodus on WL1, and the WLs are ornamented with only a few ventral setae rather than a dense range of ventral setae like in adults. Ventral setae of WL1 are also much longer than in adults and are disposed in a loose range along the length of WL1 rather than in a distal bouquet.

Locality and age. All fossils originate from the middle Kaub Formation in the Central Hunsrück Basin (consistent with their preservation), in the vicinities of Bundenbach and Germünden (see details in Material S1). Specimens IGPB-HS206, -HS207, -HS456, -HS457 and NHMMZ PWL 1992/178-LS are more precisely attributed to the Eschenbach quarry, IGPB-HS582, -HS636, -HS660, -HS694, NHMMZ PWL 2003/272-LS and 2008/141-LS to the Eschenbach-Bocksberg quarry, IGPB-M-142 to the Schmiedenberg quarry, and IGPB-HS942, -HS1039 and NHMMZ PWL 1998/155-LS to the Obereschenbach quarry. This dates the specimens to the Early Emsian.

Palaeopantopus maucheri Broili, 1929

Figures 20–24

v* 1929 Palaeopantopus maucheri: Broili, pp 272-274, 277, fig. 6, plate 5.

v 1930 Palaeopantopus maucheri: Broili, pp 209-213, figure.

v 1932 Palaeopantopus maucheri: Helfer, fig. 54.

1954 Palaeopantopus maucheri: Hedgpeth, pp 197 [text], 198 [text], 202 [text], 203 [text], fig. 4.

1955a Palaeopantopus maucheri: Hedgpeth, pp P169 [text], P170 [text], fig. 122.

1957 Palaeopantopus maucheri: Dubinin, p. 881, fig. 1Г, Д.

1958 Palaeopantopus maucheri: Tiegs & Manton, p. 321 [text].

1978 Palaeopantopus maucheri: Hedgpeth, pp 25 [text], 32 [text], fig. 3.

vp 1980 Palaeopantopus maucheri: Bergström et al., pp 33-41, 46-48 [text], figs. 24-29.

1998 Palaeopantopus maucheri: Bartels et al., pp 155 [text; pro parte], 156 [text] [non fig. 133].

v 2017 Palaeopantopus maucheri: Südkamp, pp 78, 79, fig. 121.

Examined material. HOLOTYPE. SNSB-BSPG 1929 V 3: 1 sp. – OTHER. SNSB-BSPG 1930 I 501: 1 sp., MB-A-45: 1 sp.

Diagnosis. Body robust, with lateral processes touching. Cephalon inconspicuous dorsally. Trunk unornamented. Abdomen at least three-segmented, with basal ring-like structures. Proboscis ventral. Chelifores, palps and ovigers present. Legs cylindrical in section, composed of at least 10 podomeres including a terminal claw, first podomere (coxa 1) divided into rings, second podomere (coxa 2) short, podomeres 4 to 8 elongate, joint between podomeres 4 and 5 immovable.

Description and interpretation. We accessed the three specimens described by Broili (1929: Fig. 20; 1930: Fig. 21) and Bergström, Stürmer & Winter (1980: Fig. 22) and are aware of no additional material for this species. We propose that the specimen identified by Bartels, Briggs & Brassel (1998), IGPD-HS437, does not belong to Palaeopantopus maucheri (see the dedicated paragraph in this section). Bergström, Stürmer & Winter (1980) also suggested that specimen MB-A-45 (Fig. 22) could be another species, distinct from P. maucheri, but RTI and microtomography reveal a body plan very similar in all well-preserved structures, except for a smaller size that may indicate different sex or life stage. The combination of RTI and X-ray microtomography data allows us to go further into details with the description of this species.

The trunk is compact, four-segmented, and broad. Lateral processes are very short and touching, their distal margin thickened (Fig. 20). The posterior margin of each trunk segment is slightly broadened. There is no evidence of a cephalon in dorsal or ventral view (Figs. 20, 22), though cephalic appendages and proboscis are present (see below).

The abdomen is composed of at least six elements (Figs. 20, 21), including three short, proximal elements, a median element of medium size, and two longer, distal elements. These two long distal elements are clearly identified as abdominal segments, with their joint visibly bent in the two specimens where the abdomen is visible. The first three or four, proximal elements are reminiscent of the coxal rings of the walking legs (see below). It is not clear if their demarcation represents true segmentation and these four elements may correspond to one to four abdominal segments. Bergström, Stürmer & Winter (1980) suggested that the anus was terminal (hence regarding P. maucheri as a telson-less pycnogonid), based on their interpretation of internal pyritization as a trace of the digestive tract. However, this internal pyritization is also present in the palps and ovigers (Fig. 20G), while in extant pycnogonids the digestive tract never extends into these appendages (Frankowski, Miyazaki & Brenneis, 2022). Thus, we do not accept the internal pyritization to represent a digestive tract and so P. maucheri could have possessed a non-terminal anus (and thus a telson). The lanceolate tip of the abdomen, starting about after the first proximal fourth of the last abdominal segment length, could correspond in shape to an abdomen quite like that of Palaeoisopus problematicus. To our knowledge there is, instead, no instance of extant pycnogonids (which all show a terminal anus) with a long distal portion of the abdomen slender and lanceolate.

The proboscis of P. maucheri is challenging to delimit. There is little doubt that the roughly pyriform, rounded structure extending posteriorly on the venter (Figs. 20, 22) is part of this proboscis. However, the slender, tapering structure projecting anteriorly from it (Fig. 22) puzzled Bergström, Stürmer & Winter (1980) just as it puzzles us. A similar structure can also be observed with RTI on the ventral side of the holotype (Figs. 20E, 20F) and its presence is confirmed by microtomography (Fig. 20G). Bergström, Stürmer & Winter (1980) ultimately interpreted this structure as a cephalic appendage, which is consistent with the internal pyritization seen in the holotype (Fig. 20G): in this case, it would be the oviger. Under this hypothesis, the proboscis is likely folded ventrally, with its distal mouth oriented posteriorly, and its base probably movable as in P. problematicus. However, the striking continuity of this structure with the proboscis cannot be ignored, especially in specimen MB-A-45 (Figs. 22A–22D). We suggest, as an alternative hypothesis, that this structure could also be the anterior part of the proboscis. If so, the slender, tapering projection is likely the distal part of the proboscis (and the rounded, pyriform part was proximal) as there is no evidence of any articulating petiole (such as found in Eurycyde Schiödte, 1857; e.g., Sabroux et al., 2019) or anterior cephalic projection (such as found in some modern Ascorhynchus Sars, 1877, e.g., Clark, 1963). Under this hypothesis, the round, broad proximal portion of the proboscis would be its base, implying that it inserts very posteriorly on the body, at the level of the third trunk segment (Figs. 20H, 22F). If correct, this condition is very different to extant species and it could perhaps be linked with the apparent disappearance of the cephalon in this species, which could in turn be linked with its ventralisation. Alternatively, perhaps the proboscis of P. maucheri possessed ventral ornamentation that would project posteriorly, as seen in the modern Anoplodactylus californicus-digitatus complex (Arango & Maxmen, 2006).

Cephalic appendages are all present. Chelifores are poorly preserved, and do not show any evidence of chelae (Figs. 20, 22). Palps are identified as such based on their anterior position, but their preservation is too poor to address their segmentation (Bergström, Stürmer & Winter, 1980 identified at least seven podomeres) although their base clearly shows annular structures. The tip of the palps does not show any evidence of a terminal claw. The ovigers are positioned posteriorly to the palps. They present a hook distally (a strigilis?) followed by at least two broad podomeres (Fig. 22). Bergström, Stürmer & Winter (1980) counted nine podomeres, though we can neither confirm nor reject this assertion. The base of the ovigers is not visible.

Walking legs (Fig. 23) have annular structures at their base. Bergström, Stürmer & Winter (1980) suggested that these corresponded to the lateral processes rather than to the appendages in this species (unlike P. problematicus) based on the presumed absence of lateral processes. However, as indicated above, we identify the lateral processes as short but present. In addition, it seems unlikely that such similar structures could be convergent between P. problematicus and P. maucheri (as well as in Halietes dasos and other Hunsrück fossils as demonstrated below). We therefore interpret these rings as coxa 1, as in P. problematicus. Due to the potentially high number of podomeres in the legs of P. maucheri, we refrain from applying the morphological terminology of Sabroux et al. (2023) and Siveter et al. (2023) for WL (but see Discussion). The two podomeres following coxa 1 are relatively short (the second is slightly longer than the third in many cases) and the following fourth and fifth podomeres are always directed on the same axis so that their joint may not have been movable. These two podomeres are up to twice as long as the proximal podomeres, the fifth longer than the fourth. The sixth podomere is narrower than the fifth, about three to five times as long as wide in the holotype. The seventh podomere is about 10% longer.

Distal podomeres are more difficult to define due to the distal fragmentation which either reflects anatomical segmentation or an artefact due to pyritization; the eighth podomere seems to be shorter, or as long as the seventh. It is followed by at least one, or possibly two, podomeres, which are then followed by a hook-like feature. This hook is variable in shape and its fragmentation suggests that it is composed of several, short podomeres, rather than made of a single curved podomere (such as the propodus of many extant species). Based on the number of fragments that can be counted, it composes at least four podomeres. However, taphonomic artefacts cannot be excluded to explain this fragmentation. These putative podomeres are followed by a terminal claw (Figs. 23A–23L, 23S).

Locality and age. All fossils originate from the middle Kaub Formation in the Central Hunsrück Basin, in the vicinities of Bundenbach (SNSB-BSPG 1929 V 3, 1930 I 501) and Germünden (MB-A-45; see also Material S1). This dates the specimens to the Early Emsian.

Flagellopantopus blocki Poschmann & Dunlop, 2006

Figures 25–27

v* 2006 Flagellopantopus blocki: Poschmann & Dunlop, pp 984-986, text-fig. 1.

V 2017 Flagellopantopus blocki: Südkamp, pp 79, 80, fig. 123.

V 2019 Flagellopantopus blocki: Sabroux et al., pp 1927 [text], 1929 [text], 1945 [text], Supplementary Material 3.

Material. HOLOTYPE. NHMMZ PWL 2004/5024-LS.

Diagnosis. Slender body. Proboscis pyriform (distally or totally?). Abdomen elongate, broad, four-segmented with a terminal flagelliform telson and two elongate processes on the second abdominal segment. Chelifores, palps and ovigers present, ovigers with a terminal claw. Four pairs of legs; first podomere (coxa 1) divided into rings, carrying one long dorsal process in the first pair of legs; second podomere (coxa 2) and podomeres 4 to 6 elongate; joint between podomeres 4 and 5 immovable.

Description and interpretation. This species described by Poschmann & Dunlop (2006) is, as far as we know, represented by a single specimen (Fig. 25; but see Discussion). Here, the CT-scan provides interesting insights on the body of F. blocki, and RTI helps us to improve our understanding of the structure of the legs relative to the accounts of Poschmann & Dunlop (2006) and Sabroux et al. (2019).

The trunk segments (Figs. 25E, 25F) are individualised rather than fused. The cephalon carries a mediodorsal tubercle interpreted as the ocular tubercle (Fig. 25F), which has not been observed before. The distal margin of each segment is inflated, and ornamented with rounded tubercles, including one larger dorsomedial tubercle on each segment margin. The first lateral processes are about as long as their basal width, the others about twice as long as the width of the base, separated by a distance shorter than their own diameter.

The proboscis is proximally obscured, even to X-ray imaging, but distally, it is pyriform in shape, tapering to a blunt tip. How the proboscis inserted on the cephalon is therefore not clear: its proximal part may be missing, or alternatively, the proboscis could have been simply detached during the decay of the specimen.

The abdomen is broad and segmented. It is composed of four large proximal segments about as long together as the last three trunk segments (the third abdominal segment slightly longer than the others), followed by a very long flagellum, more than three times the length of the body (from the cephalon to the abdomen). The flagellum is likely articulated, rather than rigid, as suggested by the curves along its length. Some traces of potential segmentation are also visible. The second abdominal segment carries two dorsolateral extensions near its distal margin that are likely to represent long setae.

Cephalic appendages are all present. Chelifores are long, with their chelae not preserved or absent. Palps reach the tip of the proboscis; their segmentation is poorly preserved. Ovigers (Fig. 26) have similarly poor preservation, although they have a strigilis (possibly four-segmented). This appears to have terminated with a small distal claw that is visible both with RTI and X-ray microtomography data.

Walking legs all share (Fig. 25B) the same structure as far as it is possible to identify from their preservation: the most complete leg in the holotype is the right third leg. Proximally, we suggest coxae 1 carry three rings, contra three podomeres as interpreted by Poschmann & Dunlop (2006), concomitant with the previous interpretation of Sabroux et al. (2019). However, RTI reveals that the remaining organisation of the leg is very different from the interpretation of these two references. The rings are followed by what we propose is a long podomere (about six times as long as wide), that was initially interpreted as two podomeres by Sabroux et al. (2019). This is followed by a smaller podomere at a geniculation. The fourth and fifth podomeres were interpreted as a single segment by Poschmann & Dunlop (2006) and Sabroux et al. (2019). RTI imaging, however, suggests that they are likely two podomeres sharing a single axis, their joint seemingly immovable (Fig. 25B). These are subequal in length, about five to six times as long as wide. The sixth podomere, slightly curved, is much slenderer and longer. The seventh podomere is probably only partially preserved. We refrain from using the morphological terminology of Sabroux et al. (2023) and Siveter et al. (2023) to assign identities to these podomeres (but see Discussion), as we did for Palaeopantopus maucheri with which it shares several structural resemblances. The third ring of the coxae 1 of the WL1 carries a thin projection, probably a seta.

Locality and age. The fossil is confidently attributable to the middle Kaub Formation of the Central Hunsrück Basin, but there is some confusion about its exact original location: according to Poschmann & Dunlop (2006), it originates from the Wingertshell Member of the Obereschenbach quarry; it is instead recorded by the hosting institution as originating from the Eschenbach-Bocksberg quarry. In either case, this dates the specimen to the Early Emsian.

Pentapantopus vogteli Kühl, Poschmann & Rust, 2013

Figures 28–30

v* 2013 Pentapantopus vogteli : Kühl et al., pp 557, 558, figs. 1, 2.

v 2017 Pentapantopus vogteli: Südkamp, pp 79, 80 [pro parte], fig. 122c, d.

Material. HOLOTYPE. NHMMZ PWL 2004/5024-LS. – PARATYPE. Same slab, 1 specimen. – OTHER. Same slab, one specimen.

Emended diagnosis (modified from Kühl, Poschmann & Rust, 2013). Slender body. Proboscis elongate. Abdomen not reaching beyond second coxa. Chelifores, palps and ovigers present. Four pairs of legs consisting of eight (first pair of legs) or nine (second to fourth pairs of legs) podomeres including a subchelate, terminal main claw; first podomere (coxa 1) divided into rings. Distal leg podomeres (femur to propodus) laterally flattened, ventrally setose, with ventral molariform indentations.

Description and interpretation. When first described (presented Fig. 28), Pentapantopus vogteli was considered to be the first 10-legged fossil pycnogonid (Kühl, Poschmann & Rust, 2013). While this condition is found in modern species (which are generically called polymerous, alongside 12-legged species; Arnaud & Bamber, 1987) it is extremely rare. In this description, we include a new specimen (Fig. 29), that was hidden amidst a crinoid (Parisangulocrinus zeaeformis (Follmann, 1887)) in the same slab where the holotype and paratype were found. Like P. vogteli, this specimen presents molariform indentations on the ventral side of the WL (Fig. 29), and presents an overall similar leg structure, justifying its taxonomic assignment. However, it differs from the original description in that this specimen has four rather than five leg pairs, and as many trunk segments. Reinvestigating the holotype and paratype with RTI and X-ray microtomography (Fig. 28), we find no evidence of five trunk segments: the paratype (preserved laterally) exhibits four pairs of legs and the holotype (probably preserved ventrally) presents seven legs. It seems more likely that either the supposedly fourth or fifth left leg in the holotype corresponds to the fourth right leg and was folded on the other side during the burial of the specimen. As such, P. vogteli is eight-legged.

The trunk of P. vogteli is four-segmented, the cephalon fused with the first trunk segment. The lateral processes are about as long as wide, separated by about their diameter.

The proboscis, which is clearly visible in a new specimen (Fig. 29), is elongated. It is preserved folded ventrally, suggesting it is either movable or directed vertically.

The abdomen is present, and is relatively short compared to other Devonian species, reaching beyond the lateral processes; it is not possible to determine whether it is segmented.

The cephalic appendages are all present. The chelifores are composed of at least three podomeres, including at least one scape. Palps and ovigers have their segmentation poorly preserved and hardly discriminable in the holotype; the left palp of the new specimen (Fig. 29) potentially bears a terminal claw (Fig. 28C).

The walking legs have their coxae 1 annulated at least in WL2 and 3 (and very probably also in WL1 and 4), the number of their rings being difficult to determine (beyond a minimum of three in each). WL1 present seven other podomeres (including a terminal claw) after the ringed coxa 1 in the holotype, compared to eight in WL2-4; WL1 of the new specimen appears similar in condition, as the portion of the leg covered by WL2 is not large enough to hide a complete podomere. Given the peculiarity of this assemblage, we refrain here from naming these podomeres following the morphological terminology of Sabroux et al. (2023) and Siveter et al. (2023; but see Discussion). WL1 podomeres following coxa 1 includes, from proximal to distal: a long second podomere (about five times as long as wide); a short third podomere (about as long as wide); a short fourth podomere (about twice as long as wide) with two or three ventral molariform indentations; the next two podomeres (the fifth and sixth) present ventrally four or five molariform indentations each; the seventh podomere which also possibly presents molariform indentations; and a long terminal claw (the eighth podomere) that is incompletely preserved in the type material. This claw is subchelate, movable (as indicated by the different positions it shows in left WL1-3 of the new specimen; Fig. 29) and probably able to close on the sole of the propodus, as observed in many extant species. In WL2-4, one additional podomere relative to WL1 is present somewhere between the fifth and the eight podomeres (probably in seventh position; see Discussion). The fifth to seventh (WL1) and fifth to eighth (WL2-4) podomeres present sparsely distributed long setae on the ventral surface that seem to occur in pairs (Fig. 29E).

Locality and age. The fossil is from the Central Hunsrück Basin, middle Kaub Formation, but here again there is some discrepancy between the literature and the data recorded by the hosting institution: it is either from the Wingertshell Member of the Obereschenbach quarry (Kühl, Poschmann & Rust, 2013); or from the Eschenbach-Bocksberg quarry (hosting institution data). In either case, the specimens date to the Early Emsian.

Pentapantopus ? vogteli ? Kühl, Poschmann & Rust, 2013

Figure 31

Material. NHMMZ PWL 2007/29-LS: 1 sp.

Description and interpretation. This pycnogonid is associated with two specimens of Bundenbachia beneckei Stürtz, 1886 (Fig. 31A). Its legs are not fully preserved and so it is not possible to address the presence of molariform indentations on the ventral margin of legs, as seen in the new Pentapantopus vogteli specimen (Fig. 29). Nevertheless, the two specimens have a very similar trunk outline. The number of annulations at the base of legs can be counted: four in WL1, five in WL2, four or five in WL3, and possibly three in WL4. This morphology is compatible with P. vogteli. Distal to the rings is a short structure that we identify as a potential second podomere; but this could equally be an additional disarticulated ring, or arthrodial membrane. The cephalon is remarkably well preserved, bearing a conspicuous ocular tubercle pointing anteriorly, in a very similar fashion to Haliestes dasos. Chelifores are present, possibly as well as palps. The segmentation of the abdomen is difficult to determine, its size and shape being very much like P. vogteli and H. dasos.

Locality and age. The fossil is recorded from Bundenbach in the Central Hunsrück Basin, and its preservation is typical of the middle Kaub Formation. This suggests the specimen dates to the Early Emsian.

Pycnogonida gen. sp.

Figures 32, 33

v. 1998 Palaeopantopus maucheri: Bartels et al., p. 155 [text; pro parte], fig. 133.

Material. IGPD-HS437: 1 sp.

Description and interpretation. This specimen (Fig. 32) was originally identified as the fourth representative of Palaeopantopus maucheri to be found (Bartels, Briggs & Brassel, 1998), but the two lateral projections inserted on the second abdominal segment suggest otherwise. While this specimen undoubtedly contrasts with other species described from Hunsrück (except, perhaps, Palaeothea devonica; see Discussion), its preservation does not allow for a sufficient list of diagnostic characters and, thus, for the description of a new species.

The trunk (Fig. 33) is four-segmented. The cephalon potentially has a low, rounded ocular tubercle. The posterior margins of the trunk segments widen, each possibly bearing a mediodorsal tubercle. The lateral processes are present, with a thickened distal margin.

The proboscis does not seem to be preserved.

The abdomen is multisegmented, with the two proximal-most segments being broad, about as long as wide, their posterior margin widening. The third segment is long (at least twice as long as broad), followed by up to three small segments about as long as wide. The second abdominal segment carries laterally two long processes, possibly setae.

Cephalic appendages: the chelifores are inserted anteriorly on the cephalon, but their number of podomeres is unclear. The chelae, if correctly identified, are comparatively large. The palps are either absent or unpreserved. The ovigers are positioned ventrally, with at least one posterior geniculation marking a joint, typically like the femoro-patellar geniculation of extant species (Sabroux et al., 2023).

The four pairs of walking legs consist of at least seven podomeres: three short ones, and four longer ones. The ultimate podomere visible in the right WL1 is either a terminal claw or it carries one. Proximally, we cannot be sure whether the specimen has coxa l rings like in other Devonian fossils.

Locality and age. The fossil is recorded from the Eschenbach quarry in the Central Hunsrück Basin. Its preservation is typical of the middle Kaub Formation. This suggests the specimen dates to the Early Emsian.

Undetermined pycnogonids

Figure 34

v 2013 “undetermined pycnogonid”: Kühl et al., p. 7, fig. 4.

v. 2017 Pentapantopus vogteli: Südkamp, pp 79, 80 [pro parte], fig. 122a, b.

Material. IGPB-HS942: 1 sp., NHMMZ PWL 2010/5-LS: 2 spp.

Three specimens with the typical outline of a sea spider are found associated with crinoids (Parisangulocrinus zeaeformis (Follmann, 1887); and Hapalocrinus frechi Jäkel, 1895; see details in Material S1), but cannot be confidently assigned to any known taxon, and are not complete enough to be described as a new species. Two of them were already described by Kühl, Poschmann & Rust (2013). There is not any solid evidence to refer these three specimens to the same species, except a general similar outline and size. Two of the three specimens present clear evidence of coxal rings (Figs. 34E–34L). The rest of the preservation is insufficient to allow any taxonomic assignment or species description.

Locality and age. All fossils originate from the middle Kaub Formation in the Central Hunsrück Basin, consistent with their preservation, and can therefore be attributed to the Early Emsian. Specimens in NHMMZ PWL 2004/5024-LS either originate more specifically from the Wingertshell Member of the Obereschenbach quarry (according to Kühl, Poschmann & Rust, 2013), or from the Eschenbach-Bocksberg quarry (according to the hosting institution records).

Ecology of the Devonian pycnogonids

Most of the fossils of the middle Kaub Formation were benthic organisms that could not escape the entrapping mudflows (Rust et al., 2016). Pelagic species, although present, are rarer in these layers of exceptional preservation, and were probably transported and rapidly buried (De Baets et al., 2013). Therefore, pycnogonids found in Hunsrück were probably benthic, consistent with extant species. Palaeopantopus maucheri, Flagellopantopus blocki, and specimen IGPD-HS437 share cylindrical legs with extant taxa, suggesting they were walking on the seabed (note that this does not exclude the fossils being able to occasionally “swim” in the water column; e.g., Morgan, 1972). Palaeoisopus problematicus and Pentapantopus vogteli, with their flattened legs and long ventral setae, were probably able to swim actively.

Adults of P. problematicus are found in great abundance (Südkamp, 2017), but juveniles are much rarer. While the possibility of a preservational bias toward larger specimens cannot be excluded, or that big and small specimens simply represent two different species with different frequencies, other scenarios can be considered. We note that four out of the six juveniles we examined were found associated with crinoids (Figs. 1, 17A, 17C, 17E, 17G, 18A–18H and Material S1). These associations were previously interpreted as indicating a life association (e.g., Bergström, Stürmer & Winter, 1980), but given that this almost never happens with adults this would have been limited to juveniles (the sole association of an adult P. problematicus of which we are aware is that of specimen NHMMZ PWL 2013/8-LS with the hexactinellid sponge Retifungus rugens Rietschel, 1970). Alternatively, these juveniles might have occasionally rested (or fed?) amidst crinoids’ arms, and those doing so–and which could not swim away rapidly–were those trapped by mudflows. This explains both the rarity of juveniles and their relative frequency among crinoids. It implies that they were not typically living among the benthic species commonly trapped in the Central Hunsrück Basin, but possibly just above in the water column, or in shallower seabed communities (Bartels et al., 2002). Such a scenario could also be proposed for P. vogteli which is always found in association with echinoderms (Figs. 28, 29), although the species might simply have been as rare as Palaeopantopus maucheri and Flagellopantopus blocki, which are unambiguously living on the seabed.

Was Palaeothea devonica a pantopod?

We could not locate the holotype of Palaeothea devonica. This specimen belonged to a private collection when described by Bergström, Stürmer & Winter (1980) and it may have been sold, or lost (C. Bartels, 2020, personal communication). To date, it is the sole specimen attributed to this species with confidence. Given their size, it is possible that the small unidentified sea spiders on the slab NHMMZ PWL 2010/5-LS observed by Kühl, Poschmann & Rust (2013; Figs. 34A–34H), or the smallest specimen of the slab IGPB-HS942 (Figs. 34I–34L), are related to this species. However, all of these specimens are only fragmentarily preserved. The holotype of P. devonica was never prepared (Otto, 1999) and was only studied and illustrated based on X-ray projections. The illustrations available of this fossil by Bergström, Stürmer & Winter (1980) provide very few informative details. For this reason, it is not possible for us to confidently attribute any specimen to P. devonica until the holotype can be identified and redescribed. If the disappearance of the holotype is confirmed, it will be necessary to identify this binomen as a nomen dubium.

Based on their observations, Bergström, Stürmer & Winter (1980) concluded that P. devonica had no basal leg annulations (= annulated coxae 1) and possessed three short coxae on each leg and chelifores with reduced chelae. Hence, they interpretated this species as a pantopod, even suggesting affinities with Ammotheidae Dohrn, 1881 sensu Stock (1994) (= Ascorhynchoidea Hoek, 1881 sensu Bamber, 2007) (a group now regarded as paraphyletic; Ballesteros et al., 2021; Bamber, El Nagar & Arango, 2024; Sabroux, Corbari & Hassanin, 2023). This morphological interpretation is questionable. Re-examination of the radiographs published by Bergström, Stürmer & Winter (1980) shows that the so-called coxae are very narrow, more than usual for these podomeres (Fig. 35); we could as well interpret these structures as annulated coxae 1. The abdomen is poorly preserved, but its broad base and the outline on the radiographs of Bergström, Stürmer & Winter (1980) indicate it was comparatively large. Affinities of P. devonica with Pantopoda are therefore dubious at best.

Interestingly, P. devonica shares many characters with Flagellopantopus blocki and/or the specimen IGPD-HS437: this includes the thickened posterior margin of the trunk segments, the four dorsomedian tubercles on trunk segments, and the number of annulations at the leg bases (in F. blocki). It is also worth noting that the abdomen of P. devonica could also have the same dimensions as the massive abdomen of F. blocki or the specimen IGPD-HS437, although no unequivocal conclusion can be drawn from available illustrations. Perhaps P. devonica was of the same species as either one or the other of these two pycnogonids, though its smaller size would imply it was a juvenile or a different sex. Perhaps it could be another, third species. Unfortunately, these hypotheses cannot be further addressed in the absence of the holotype of P. devonica.

Variability of the Pycnogonida walking leg ground plan

The most emblematic characteristic of modern sea spiders that makes them (almost) unmistakable in the field is their slender morphology with a narrow, often inconspicuous body onto which long legs articulate. This typical aspect has led to the sea spiders being coined as “no-bodies” (e.g., Brenneis et al., 2017), or pantopods (Pantopoda Gerstäcker, 1863; literally “all made of legs”). Therefore, including pantopods within a broader phylogenetic framework among Pycnogonida and Arthropoda in general, involves a thorough comparison of the appendages. In most cases, the Hunsrück sea spiders have poorly preserved cephalic appendages; in contrast, walking (or swimming) legs were more massive and therefore often better preserved, sometimes exquisitely.

Leg type 1: Pantopoda

The leg pattern of modern sea spiders (as, presumably, in all pantopods) is very stable (Fig. 36), with rare exceptions such as the parasitoid Nymphonella tapetis. From proximal to distal, the WL are composed of three short coxae; three long podomeres: femur, patella, tibia (more often referred to femur, tibia 1 and tibia 2 in taxonomic literature); a variably elongated tarsus and propodus; and a terminal claw often flanked by two auxiliary claws. Hunsrück fossils show that this ground plan (to which we will refer to as leg type 1; Table 2) has not always been the sole leg pattern for sea spiders.

Table 2:

Characteristics of the three leg types in Pycnogonida.

Leg type 3 is presented following two alternative hypotheses (i) and (ii). See also Fig. 36.

Leg characters	Leg type 1	Leg type 2	Leg type 3 (hypothesis i)	Leg type 3 (hypothesis ii)
Coxa 1	Simple	Annulated	Annulated	Annulated
Femur	Simple	Simple	Divided	Simple
Femur and patella	Mobile	Mobile	Mobile	Immobile?
Metatibia	Absent	Present in swimming legs 2–4	Absent	Present in all walking legs
Rows of long ventral setae	Absent	Present	Absent	Absent
Transversal section	Cylindrical	Laterally flattened	Cylindrical	Cylindrical

DOI: 10.7717/peerj.17766/table-2

The most conspicuous and constant feature that differentiates Hunsrück fossils from leg type 1 is the annular structure at the base of legs. It was identified in Palaeoisopus problematicus and Palaeopantopus maucheri (e.g., Bergström, Stürmer & Winter, 1980) and we also show clear evidence of its presence in Flagellopantopus blocki and Pentapantopus vogteli. As explained above, we suspect that Palaeothea devonica also had basal annulations. This structure is also observed at the base of the legs of Haliestes dasos (Siveter et al., 2023). Siveter et al. (2023) argued that these rings are likely homologous with the coxae 1 of leg type 1, relying on two arguments: 1) the position of legs in H. dasos suggests that the joint between the purported coxa 1 and 2 of this species has a promotor-remotor swing, which is typical of the coxa 1-coxa 2 articulation (Manton, 1978)–note that the Hunsrück material does not provide further evidence to support this observation; and 2) that the following podomere could be identified as coxa 2 through a ventrodistal protuberance that is typical of coxa 2 of extant pantopods, where the gonopores open. While Bergström, Stürmer & Winter (1980) interpreted the annulations at the base of legs of P. problematicus the same way as Siveter et al. (2023), they proposed something different for P. maucheri, namely that they were “without doubt” part of the lateral processes seen also in P. maucheri. This hypothesis relies on the absence of lateral processes, an observation which we do not accept, as detailed above. We therefore regard the coxal rings as homologous among all the sea spiders that present them.

Affinities of the Hunsrück sea spiders

Sea spiders have been classified in up to four orders: Palaeoisopoda, Palaeopantopoda, Nectopantopoda and Pantopoda (see Table 1; Hedgpeth, 1954, 1978; Bamber, 2007). Hedgpeth (1954) included extant species in Pantopoda, a name that was until that time used as a synonym of the class Pycnogonida. The order Palaeopantopoda was named some 25 years before by Broili (1930) for the species Palaeopantopodus maucheri. Later, when he recognised Palaeoisopus problematicus as a sea spider, he also placed this species in the order (Broili, 1932). Hedgpeth (1978) divided Palaeopantopoda into two infraorders, Palaeopantopodina Broili, 1930 and Palaeoisopodina Hedgpeth, 1978, which were much later raised to the ordinal level (i.e., Palaeopantopoda and Palaeoisopoda) by Bamber (2007). Bamber (2007) also established a fourth, monospecific order, Nectopantopoda, for Haliestes dasos. This classification is still in use (see for example Bamber, El Nagar & Arango, 2024), but this reflects more the scarcity of attempts to review it than a real consensus.

This work demonstrates that P. problematicus, P. maucheri and Flagellopantopus blocki all had a long, segmented abdomen, with often (if not always) a terminal telson. We suggest this was also probably the case for Palaeothea devonica and Pentapantopus vogteli. This character is arguably plesiomorphic as it mirrors the state in all potential fossil outgroups and it is plesiomorphic to arthropods as a whole. Consequently, the reduced, unsegmented, telson-less abdomen of pantopods is a robust autapomorphy for the group. All fossils studied here presents the typical annulated coxae 1 (it is not unambiguously observed only in the specimen IGPB-HS437). This also clearly distinguishes these fossils from Pantopoda, which have legs of type 1 (i.e., without annulated coxae 1; see Table 2, Fig. 36).

Leg morphology allows us to identify two distinct groups among the Devonian fossils. On one hand are sea spiders that have laterally flattened WL with ventral setae and a first leg pair with a reduced number of podomeres, that probably served a raptorial function (leg type 2; see Table 2, Fig. 36). This group included P. problematicus and P. vogteli, to which we can also add the Silurian species H. dasos. On the other hand are P. maucheri, F. blocki, and perhaps P. devonica, that have cylindrical legs (leg type 3; see Table 2, Fig. 36) and a broad, segmented abdomen.

Whether these two groups represent clades, or whether one or both are grades, remains uncertain. The group of P. problematicus has a number of characters that have the potential to be synapomorphic. The paddle-like distal podomeres of legs are potentially one: there is a trend in extant chelicerates to have flattened legs, but in these sea spiders this condition affects the proximal podomeres rather than the distal ones. However, the condition of fossils is often unknown. Another candidate is found in the modified, potentially raptorial first legs, but this is not unique to these sea spiders (e.g., the eurypterid families Megalograptidae and Mixopteridae have such limbs; Schmidt et al., 2022). Similarly, the metatibia has the potential to be a synapomorphy, but the podomere composition and homology of many ancient chelicerates remains poorly understood. The group of P. maucheri is unambiguously characterised only by cylindrical legs and the annulated coxae 1, which are characters found in the two other Pycnogonida groups. The super-division of the propodus can only be identified in P. maucheri, and the division of the femur into two immobile podomeres is putative.