ABSTRACT. We describe a new genus of symmoriiform chondrichthyan from the Late Mississippian Fayetteville Shale of Arkansas, USA and include this fossil in a phylogenetic analysis of chondrichthyans. This taxon possesses elongate cartilaginous rays extending from the gill arches, forming an operculate structure that covers at least two of the branchial arches farther posteriorly. Although presence of a ‘hyoid operculum’ has been postulated in at least two unrelated Paleozoic sharks (e.g., Triodus, Tristychius), subsequent investigations failed to corroborate those claims. The new fossil therefore provides the first evidence of an endoskeletal operculum formed by elongate, fused pharyngeal arch rays in a chondrichthyan.
KEYWORDS: Chondrichthyes, Symmoriiformes, operculum, CT scanning, phylogeny, new genus, new species.
RÉSUMÉ. Nous décrivons un nouveau genre de chondrichthyen symmoriiforme des schistes Fayetteville du Mississippien supérieur de l’Arkansas, aux États-Unis, et incluons ce fossile dans une analyse phylogénétique des chondrichthyens. Ce poisson possède des rayons cartilagineux allongés s’étendant des arcs branchiaux, formant une structure operculaire qui couvre au moins deux des arcs branchiaux plus en arrière. Bien que la présence d’un «opercule hyoïde» ait été inféré chez au moins deux requins paléozoïques non apparentés (par exemple, Triodus, Tristychius), des études ultérieures n’ont pas permis de corroborer ces affirmations. Le nouveau fossile fournit donc la première preuve d’un opercule endosquelettique formé par des rayons pharyngiens allongés et fusionnés chez un chondrichthyen.
MOTS CLÉS: Chondrichthyes, Symmoriiformes, opercule, CT scanner, phylogénie, genre nouveau, espèce nouvelle.
Modern crown chondrichthyans include two distinct lineages (et al. 2013; Giles et al. 2015; Dearden et al. 2019). The divergence of elasmobranchs and holocephalans can be traced back to the Late Devonian (Coates et al. 2017, 2018), but the evolutionary relationships of many Paleozoic shark-like fishes to the chondrichthyan crown remain uncertain (Grogan et al. 2012; Coates et al. 2017; Maisey et al. 2019). One group whose position is especially controversial is the order Maghriboselache Klug, Coates, Frey, Greif, Jobbins, Pohle, Lagnaoui, Bel Haouz & Ginter, 2023, and , Falcatus Lund, 1985, Damocles Lund, 1986, and Akmonistion Coates & Sequeira, 2001) (Lund 1985, 1986; Coates & Sequeira 1998, 2001; Pradel et al. 2011, 2014; Coates et al. 2017; Klug et al. 2023). Some analyses have placed symmoriiforms below the divergence of holocephalans and elasmobranchs (i.e., excluding them from the chondrichthyan crown) (Pradel et al. 2011), whereas other studies have allied them with holocephalans, thus placing symmoriiforms within the chondrichthyan crown (Janvier 1996; Coates & Sequeira 2001; Coates et al. 2017). Renewed support for a holocephalan-symmoriiform relationship emerged several years ago from a study of cranial morphology in et al. 2017). This relationship was also recently supported as part of a study of the gill skeleton, jaws, cranium, scales, and teeth of et al. 2018), and from an analysis including the Late Devonian et al. 2020).
Here, we introduce a previously undescribed symmoriiform from the Late Mississippian Fayetteville Shale of Arkansas, which has a unique operculum-like structure, but also has a symmoriiform-like basicranium, jaws, and teeth. The basicranium is only partially preserved, and the gill arches are disarticulated, but these gill arches are associated with elongate endoskeletal rays forming a gill cover.
Chondrichthyan paleontology is continually challenged by the rarity of cartilage in the fossil record. However, when cartilage is preserved, it provides a great deal of information about early chondrichthyan anatomy, evolution (Brazeau 2009; Maisey et al. 2019), life history strategies and reproductive structures (Lund 1982; Capetta 1987; Stumpf et al. 2021), and phylogenetics (Brazeau 2009; Davis et al. 2012; Coates et al. 2017). The Fayetteville Shale of northwestern Arkansas is beginning to yield remarkably preserved fossil cartilage, with implications for our understanding of Late Mississippian chondrichthyan evolution (Lund & Mapes 1984; Maisey 2007; Pradel et al. 2014; Bronson et al. 2018).
The exceptionally preserved chondrichthyans of the Fayetteville Shale represent just one aspect of its biodiversity. The formation is perhaps best known for its oil-bearing black shales, which have been well known to geologists for many years (Meeks 1997; Murthy et al. 2004; Alase 2012). However, the formation is also famous for its fossil plants (Dunn 2002; Dunn et al. 2002, 2003, 2006), whose abundance and good preservation enabled a reconstruction of the formation’s wetland ecosystem (Dunn 2002, 2003; Dunn et al. 2006), and for its diverse cephalopod assemblages (Girty 1910; Gordon Jr et al. 1969), which provide evidence of the formation’s marine ecosystems. The preservation of these cephalopods, including evidence of mass mortality events, is informative for reconstructing the depositional environment of the formation (Meeks 1997), which was likely characterized by rapid burial events and transgression-regression cycles which resulted in preferential preservation of cartilage, and simultaneous degradation of shell and bone (Bronson et al. unpublished data).
AMNH FF 20509, represented by nearly complete lower jaws, dentition in situ in lower jaw, partial upper jaw, partial basicranium, branchial arch elements, partial pectoral girdle cartilage and fins, fin impressions, and branchial arch rays
The specimen was collected along a series of outcrops of the Fayetteville Shale, in the bed of Cove Creek, Searcy County, Arkansas (Mapes Locality ARC-08) (Mapes 1979) beginning at the junction of Highway 65 and Cove Creek Bridge and going upstream for approximately 1 mile. The precise outcrop at which this specimen was collected, however, is unknown.
Specimen number AMNH FF 20509 was photographed using a Canon EOS 70D SLR camera. The entire specimen except for the partial pectoral girdle and fins was then scanned at the American Museum of Natural History on a 2010 GE phoenix v|tome|x s240 high-resolution microfocus computed tomography system (General Electric, Fairfield, CT, United States), using a diamond target and 0.5 mm copper filter. This specimen was too large to be imaged as a single scan, and so the anterior and posterior ends were scanned one at a time, and each half was reconstructed and segmented as a separate image stack, producing 1 836 images for each portion, both with a voxel size of 0.13 mm.
Image stacks were reconstructed using GE phoenix datos|x and FIJI software (Schindelin et al. 2012). Anatomical features were segmented using Mimics software (Materialise, Leuven, Belgium).
This study built upon an existing character matrix focused on fossil chondrichthyans and bony fishes (Coates et al. 2018; Frey et al. 2020). A data set of 222 morphological characters was compiled for 48 genera. Characters, taxa, and character scorings are listed in the Supplementary Material (Appendix 1). The matrix is available as a NEXUS file on MorphoBank, as project 3996, matrix 28783 (Appendix 2).
We added 32 characters to the matrix described in Frey et al. 2020, largely focused on the chondrichthyan chondrocranium (see Supplementary Material in Appendix 1) and removed 40 characters that were primarily focused on early osteichthyans or acanthodians and thus became phylogenetically uninformative for our set of taxa. One notable character added to the matrix is Character 39: Endoskeletal operculum absent (0) or present (1). We defined an endoskeletal operculum as a flap or cover, supported by endoskeletal hyoid rays, such as that which exists in extant holocephalans, and unlike the dermal bone opercle of bony fishes or the palatomandibular operculum of some acanthodians or the short, septate branchial arch rays of extant elasmobranchs. This character was thus marked ‘present’ in both this new taxon and in holocephalans.
The scoring of several characters from Coates et al. (2018) and Frey et al. (2020) was modified: In terms of coding character states in the dorsal aorta and occipital region, Coates et al. (2018) characterized the position of the dorsal aorta similarly between holocephalans and symmoriiforms, likely to maintain character independence. However, we chose to code this character (our Character 166) as follows: occipital region overlies paired prebranchial dorsal aortae (0); overlies median prebranchial dorsal aorta (1); overlies paired postbranchial aortae (2). This character portrays the three conditions observed across total group median prebranchial dorsal aorta. In xenacanths, euselachians, pair of prebranchial aortae that divide posterior to the occipital region. The modern holocephalan arrangement is completely different: the occipital region overlies the postbranchial (not prebranchial) part of the aortic circuit. Extinct holocephalans with subcranial branchial arches likely had the same aortic arrangement (Pradel et al. 2021). Other modifications to character scorings contraCoates et al. 2018 and Frey et al. 2020 are noted in the Supplementary Material (Appendix 1).
To explore the effect of parameterization on the inferred topology for this dataset, unweighted parsimony, implied weights parsimony (Goloboff et al. 2008a), and the Lewis Mkv model (Lewis 2001) were compared under the Akaike Information Criterion (Akaike 1974) using likelihood approximations of parsimony methods within TNT software using the “morphotestall” and “morphodoaic” scripts (Goloboff & Arias 2019). The best AIC score was obtained for the Lewis/Mkv model, which we show here alongside unweighted parsimony (the method used by Frey et al. 2020, to compare our results more directly to their analysis). For all analyses, we used
For unweighted parsimony analysis, the data set was subjected to a new technology search in TNT software (Goloboff et al. 2008b) through ratchet, fusing, and drifting. Bootstrap values were generated using 1 000 replicates on this strict consensus tree, as were Bremer support values, a consistency index, and a retention index (CI and RI). Search parameters and the TNT script used in this search can be found in the Supplementary Material (Appendix 1). We used Driven search (“xmult”) as the source of initial trees, the best score was found 10 times and we used 10 replications. We then used default parameters (without constraint) for Sectorial searches, 10 Ratchet iterations, one round of Tree fusing, and five Drift iterations. Finally, a round of TBR (tree bisection reconnection), with trees from RAM, was performed after the initial search.
Maximum likelihood analysis was conducted using the Mkv model in RAxML v8.2.9 (Stamatakis 2014), with five independent searches. One thousand bootstrap replicates were conducted and mapped onto the inferred topology. Likelihood analysis yielded a slightly different topology from the parsimony topology, likely at least partially because polymorphisms must be transformed to the character state “unknown” to use the Mkv in RAxML. Results of the likelihood analysis, parameters, and AIC scores used to determine model rankings, are included in the Supplementary Material (Appendix 1).
Institutional abbreviations
Anatomical abbreviations
add depression potentially for adductor musculature;
c cranium;
cb ceratobranchials;
ch ceratohyals;
d denticles;
da? possible opening for the dorsal aorta;
mc Meckel’s cartilage;
occ occiput;
op opercular cartilage;
ot cap otic capsule;
pa f pharyngeal arch fragments;
pq palatoquadrate fragment;
t teeth.
-
-
-
As for type and only species.
(Figs 1-6)
Partial cranium, jaws, gill arches, partial pectoral girdle and fins, and lower teeth, AMNH FF 20509.
Fayetteville Shale (Upper Mississippian, Middle Chesterian) of northwest Arkansas, United States.
River bed of Cove Creek, near the town of Leslie, Searcy County, Arkansas, United States, Middle Chesterian (Upper Mississippian), Carboniferous.
Teeth small and cladodont type, with unfused bases and three lingually gently recurved cusps, set in elongate, scalloped Meckel’s cartilages that house widely spaced tooth families. Denticles present between tooth families. Hyoid and branchial arches possess elongate rays, which are fused together to form a corrugated structure covering the gill openings.
The presence of ten upper and lower families of small cladodont teeth, widely spaced and housed in large alveoli, which, combined with phylogenetic analysis places the taxon within the family et al. 2010).
The presence of buccopharyngeal denticles that may be compound or single (insufficient resolution in CT scan to determine their type), scalloped jaw margins, and tentatively a single canal for the dorsal aorta entering the basicranium, combined with symplesiomorphic characters (such as possessing 15-18 distal fin radials and cladodont teeth, the shape of the basicranium [see Description], and elongate jaws), as well as the results of a phylogenetic analysis, place this taxon within Order
Our description is based on the holotype AMNH FF 20509, which has been part of the Mapes Fossil Collection since the 1970s. The specimen was recognized as a chondrichthyan by Royal Mapes and brought to the attention of Rainer Zangerl in a photograph dated 1979. Since this early photograph (Supplementary Material, Fig. S1; Appendix 1) was taken, the specimen was damaged in shipping, and later experienced significant pyrite decay; it was previously a continuous slab of matrix, but now the pectoral girdle is separated from the pharyngeal region (Fig. 1). Much of the morphology is recognizable on the surface: An acuminate rostral end of elongate lower jaws, posterior to which are a collapsed set of gill arches that support elongate cartilaginous rays which extend posteriorly for most of the length of the pharynx, as well as an exceptionally preserved pair of pectoral fins with three-dimensional fin radials and impressions of the ceratotrichia and fin margins. Computed tomography reveals teeth, pharyngeal denticles, a partial basicranium, and fragments of additional branchial arches within the largely phosphatic matrix.
Teeth
AMNH FF 20509 (Fig. 1) has an apparently complete lower dentition (Figs 2, 3), with cladodont teeth that have a lingually recurved central cusp and a reniform tooth base (Fig. 3) (Ginter et al. 2010). No teeth from the upper jaw are identifiable in the specimen, and all teeth are hidden within the matrix and made visible only by CT scanning. Like et al. 2014), the teeth are small and housed in widely-spaced “pockets,” with tooth families an average of 4.4 mm apart (ranging from 1.6 to 6.42 mm apart, excluding areas where the fossil is broken). Tooth family spacing is most regular at the middle of Meckel’s cartilages; tooth families are slightly closer together at the anterior of the jaw, and spacing is difficult to determine more posteriorly, where teeth are clumped and disorganized, possibly because of taphonomic processes. The fossil has uniformly sized teeth within tooth families, indicating relatively rapid tooth replacement in a linguo-labial sequence. There are generally seven teeth per family, though this can be hard to see in rows that are less organized, and there are at least 25 tooth families (12 clear families on one side, and 13 on the other, plus a jumble of teeth at the back of that row), as well as a single symphysial tooth family (Fig. 2).
Based on the shape of the teeth, with a large central cusp and subtle lateral cusps, they were likely used for clutching prey (Ginter et al. 2010; Maisey et al. 2014). Their appearance is somewhat like that of et al. 2010).
Denticles
Denticles cover the space between tooth families (Fig. 3), and pharyngeal denticles are visible both as bumps just barely under the matrix surface, and in more detail through CT imaging (Fig. 4). The location of these denticles in the specimen, between the posterior ends of the Meckel’s cartilages, indicates they are buccopharyngeal denticles. The pharyngeal denticles are difficult to see in detail in the CT scan, due to low resolution, but they are shaped like grains of rice. The denticles between tooth rows are less regular in shape. Higher resolution scanning would confirm whether or not the pharyngeal denticles are stellate (Williams 1979; Ivanov 2005), as well as determine whether they are simple or compound (as in Akmonistion [Coates & Sequeira 2001]). No external dermal denticles have been observed in this specimen, though they may have been lost to taphonomy or preparation early in the specimen’s history.
Cartilage structure
Scan resolution was insufficient to discern many details of cartilage structure, apart from there being evidence of only single-monolayered tessellated calcified cartilage (Maisey et al. 2021) in much of the skeleton (i.e., it seems to lack multiple-monolayer tesserae like that of forms such as et al. 2016]).
Visceral arches
The jaws are long and gracile, extending nearly 5 centimeters behind the posterior margin of the cranium. The mandibular joint is located posterior to the braincase, as is the case in symmoriiform taxa such as et al. 2014). This is in contrast to the condition in , Chondrenchelys Traquair, 1888, et al. 2017). The Meckel’s cartilages have a moderate depression along the ventral surface, running along the concave curvature of the jaw from roughly five to ten centimeters anterior to the jaw joint, which does not appear to be an artifact of taphonomy and may have housed an attachment for the adductor musculature in life (Fig. 2). A similar ventral depression is present in the jaws of et al. 2014).
The visceral arches are partially preserved, but the branchial elements have collapsed. The flattened condition of the branchial arches makes it difficult to identify individual elements with certainty, though components of at least three pharyngeal arches are recognizable (Fig. 2). Each of these three pharyngeal arches is represented by paired ventral (ceratal) elements (Fig. 2). The dorsal (epal) elements of the pharyngeal arches are mostly absent, though the dorsal portion of the mandibular arch is preserved as a small remnant that is likely part of the palatoquadrate. A single basihyal may be present based on an irregular chunk of cartilage present in the correct location; however, its identity as a basihyal is uncertain because it is poorly preserved, and as such is not rendered in the reconstruction in Fig. 2. Two pairs of ceratobranchials are preserved, presumably representing the first and second branchial arches, with a fragment of a third arch just posterior to them. The ceratohyals (Fig. 2B) have a slight flange or ridge on their ventromedial edge; however, the ceratobranchials do not have a similar flange and are more rounded in cross section than the flattened ceratohyals.
Gill cover
One unique feature of AMNH FF 20509 is a structure interpreted as an operculum, which extends posteriorly from the level of the hyoid and first branchial arch. The hyoid rays extend to almost the same point as the rays from the first branchial arch. Though the gill arches are collapsed so we cannot confirm that the hyoid and first branchial arch rays covered all the posterior successive gill slits in life, the rays are certainly long enough to close off all the gill openings. The cartilaginous opercular rays attached to the hyoid arch appear to have a ventral and at least partly dorsal component (Fig. 5). Possibly because of weathering, or events post-preparation, the rays lift off from the rest of the specimen in cohesive layers. The flaps formed by the cartilaginous rays have a corrugated appearance, indicating the rays are more adhered to one another than they are to any other part of the specimen, unlike the hyoid rays of other chondrichthyans. A similar adherence of other small structures (denticles, cartilaginous fin radials) is not observed in the specimen, so we interpret the adherence of these hyoid rays as a true feature of the fossil rather than purely a taphonomic artifact. This close adhesion of adjacent rays is visible directly on the surface of the fossil and is consistent in tomographic sections. At least two other branchial arches possess thin and elongate rays, which do not seem to be fused together quite so closely as the rays from the hyoid arch, but they do appear to be fused as they also produce a corrugated appearance. It is difficult to determine the length of the branchial rays from the oblique sectional view provided by CT scans, but the hyoid arch rays are over 10 cm long, covering much of branchial opercular rays and therefore likely covering the branchial openings in life.
Fins
Two pectoral fins and partial pectoral girdle cartilages are preserved with the specimen (Supplementary Material, Fig. S1; Appendix 1); however, they are no longer articulated with the rest of the specimen and were not CT scanned alongside the cranial skeleton. The connection between the fins and cartilages is degraded due to pyrite decay, leaving little preserved of the pectoral cartilages, however the preservation of the fin radials is exceptional; they are three dimensionally preserved without any evidence of becoming crushed taphonomically. Joints appear to be present between the proximal and distal radials (Supplementary Material, Fig. S2; Appendix 1). Impressions of the fins are also preserved, including impressions of ceratotrichia (Supplementary Material, Fig. S1; Appendix 1). Length of the fin impressions indicates the fins were at least 24 cm in length. In this specimen, there are 15-18 distal fin radials (differing from right to left, due to preservation), with no interradials (smaller, sometimes distally forked elements found between fin radials). Among Paleozoic chondrichthyans, only
Partial basicranium
The most posterior portion of the cranium is partially preserved, and it is not dramatically crushed by taphonomic processes (Figs 2; 6). However, little information can be inferred about the structure and size of the orbits, or about the internal structure of the braincase. Compared with the braincase of ‘Maiseyacanthus Bronson, 2021 (Bronson 2021), the dorsal surface of the cranium is not domed; unfortunately, this region is so degraded that no precerebral fontanelle can be discerned. There is no evidence of subcranial ridges or a spiracular groove, the lateral commissure appears chondrified, and there is no evidence of a subotic occipital fossa. A crack runs through the right ventral side of the cranium just anterolateral to the hypotic lamina; however, this opening is an artifact of taphonomy. There appears to be a single opening or canal for the dorsal aorta (Fig. 6), but due to the resolution of the scan, it is challenging to reconstruct any smaller foramina.
We recovered eight most parsimonious trees from the parsimony analysis, with a tree score of 547. The Consistency Index for these trees was 0.444 and the Retention Index was 0.721.
Symmoriiform chondrichthyans grouped together in all eight most parsimonious trees, and
Acanthodians were consistently recovered as paraphyletic to conventionally defined chondrichthyans, consistent with other recent analyses. In all most parsimonious trees, et al. 2017).
Symmoriiform teeth are distinctive, having a biconvex median cusp and a shallow basolabial projection that is generally the same width or less than the width of the median cusp (Ginter et al. 2010). Some symmoriiform teeth have a basolabial depression, and an orolingual button is present, though sometimes divided by a large foramen to the basal nutritive canal. In symmoriiforms, the lateral and medial cusps are structurally separated with no histological connection between the dentine of adjacent cusps (Sequeira & Coates 2000).
Scanning resolution and the density of the specimen make it difficult to discern detailed morphology of the teeth, but, based on the morphology visible from the CT scan (Figs 3, 4), tiny lateral cusps are present and are not deeply separated from the central cusp. The small height of these lateral cusps distinguishes Akmonistion (Coates & Sequeira 2001). The teeth are also unlike those of et al. 2010), while
The teeth are convex lingually, similar to the condition in et al. 2010). However,
The wide spacing of tooth families in et al. 2014). Multiple teeth in each tooth row were probably functional simultaneously, as was the case in many Paleozoic sharks.
The presence of oropharyngeal denticles between tooth rows in et al. 2009). As they mature, the teeth migrate forward to the biting edge of the jaws from the mucosal fold, where they function in feeding and then are typically shed (with some exceptions, e.g. et al. 2012; Hodnett & Lucas 2015]). et al. 2012). It is unlikely that the denticles between tooth rows in
The presence of oropharyngeal denticles between tooth batteries is perhaps most similar to the condition in
The gill arches of many Paleozoic sharks were probably poorly calcified and thus rarely preserved, though the mandibular arch cartilages are generally better preserved but crushed flat. The long, gracile jaws of et al. 2021).
As in modern elasmobranchs, gill rays in the new fossil also extend from some (perhaps all) arches behind the hyoid (in analogous fashion to those in the modern frilled shark
Modern sharks and rays do not have endoskeletal opercula. Instead, the cartilaginous rays that may extend from elasmobranch gill arches are ‘septate’ – short, but coming from each gill arch, to support the interbranchial septa between gill slits. Holocephalans, on the other hand, have elongate hyoid rays (with no rays on other gill arches)(Didier 1995) that form the endoskeletal support for the operculum (Gillis et al. 2011). However these rays are only proximally joined (Didier 1995; Pradel et al. 2013).
Lacking an operculum, living sharks and rays pump water over the gills by depressing and raising the floor of the buccal cavity (buccal pumping), or by opening the mouth and swimming forward (ram ventilation). Extant holocephalans have a much lower pressure system of gill ventilation, that incorporates little cranial movement, and they utilize their gill cover to prevent backflow of water into the pharyngeal cavity through the gill slits (Dean et al. 2012).
Among Paleozoic sharks, et al. 2019). These taxa do have longer branchial rays than most modern elasmobranchs, which could indicate that longer branchial rays were a condition common in the chondrichthyan stem. However, no branchial rays have been reported in et al. 2014), as well as in the iniopterygians .
A palatomandibular, dermal operculum is present in some acanthodians (Gillis et al. 2011; Dearden et al. 2019) covering septate gill arches. Acanthodians are stem chondrichthyans (Miles 1973; Hanke & Wilson 2006; Brazeau 2009), so therefore septate gill arch appendages are likely plesiomorphic for conventionally defined chondrichthyans, and holocephalans have undergone a reduction in the number of branchial rays (Gillis et al. 2011). As discussed by Dearden et al. (2019), a bony operculum was probably lost somewhere along the chondrichthyan stem, but it is impossible to rule out the presence of an endoskeletal operculum composed of cartilaginous rays in other species of fossil chondrichthyans – it could be that additional taxa had an operculum, but this delicate cartilaginous structure simply was not preserved.
Crown chondrichthyans represent a highly specialized monophyletic lineage that evolved from ‘bony’ Paleozoic ancestors (specifically, the ‘acanthodian’ fishes, named for their spine-bearing fins) (Brazeau 2009). Holocephalans have traditionally been considered one of the two lineages that define crown chondrichthyans. However, in several recent analyses, symmoriiform sharks were hypothesized to share a common ancestor with holocephalans (Coates et al. 2017, 2018; Frey et al. 2020), a relationship that was first alluded to two decades ago (Janvier 1996). Despite these findings, this relationship is not supported by other phylogenetic analyses (Pradel et al. 2011; Maisey et al. 2017; Brazeau et al. 2020). Our phylogenetic analyses specifically include the character “endoskeletal operculum”, which might theoretically unite these groups, and yet AMNH FF 20509 is recovered within symmoriiforms and the symmoriiform-holocephalan relationship is not supported.
Symmoriiform pectoral fins tend to have long metapterygial axes, which may have allowed for increased fin mobility (Maisey 2008). Preservation makes it impossible to confirm the condition of the metapterygium, though it may be visible as a projection from the pectoral girdle (Supplementary Material, Fig. S2; Appendix 1). Jointing is apparent in the pectoral fins of
The phenomenal preservation of the fin radials speaks to the particular preservational environment of the Fayetteville Shale, which has yielded numerous well-preserved chondrichthyans in recent years, such as et al. 2014; Bronson et al. 2018). The reducing depositional environment and anoxic sediments likely contributed to the preservation of these fin radials and branchial rays in three-dimensional condition. A number of symmoriiform caudal fins have been recovered from the Fayetteville Shale (Bronson et al. unpubl. data), though none were found in association with AMNH FF 20509.
The structure interpreted here as a partial basicranium has a smaller postorbital process than many symmoriiforms, including et al. 2017)Maiseyacanthus (Bronson 2021). More information from a higher resolution scan could be useful, however pending the discovery of a more complete cranium, only the most general shape of the basicranium can be compared with previously described taxa.
Recent anatomical studies of et al. 2017)et al. 2019), et al. 2017)et al. 2018)et al. 2020)et al. 2018; Frey et al. 2020). However, contrary to our findings, these authors found that symmoriiform chondrichthyans (including et al.’s (2018) description of
However, the result of our maximum likelihood analysis was a topology that did recover this symmoriiform-holocephalan clade. This could be the result of algorithmic differences between parsimony and likelihood models. An examination of the signal in the matrix using a p-distance NeighborNet in SplitsTree4 suggested that phylogenetic signal in the data more closely recalled the parsimony output than it did the likelihood output. Notably, to estimate a tree in RAxML using a Mkv analysis, the four polymorphisms in the character matrix had to be transformed to unknowns, effectively resulting in a different data matrix from the one used for our unweighted parsimony analysis.
To better understand why these results were in conflict, we used Mesquite (Maddison & Maddison 2021) to identify which traits were grouping particular clades together in both the parsimony and likelihood trees. In the likelihood tree,
Character development and scoring, though a process that aims for objectivity, is perpetually evolving to accommodate new data, particularly as hypotheses of homology are confirmed or debunked. In fishes, characters such as presence/absence of claspers or, indeed, presence/absence of an operculum are often, perhaps erroneously, treated as homologous traits by virtue of being scored the same way. Sonic hedgehog (shh) signaling maintains elongation of the branchial rays in all modern cartilaginous fishes (Gillis et al. 2011), however the operculum character could be scored differently: a holocephalan hyoid-only state, a chondrichthyan septate state, and a
Given the combination of characters present in
Features of the dentition and elongate Meckel’s cartilages, as well as phylogenetic analysis by both parsimony and likelihood, place et al. 2011; Giles et al. 2015; Coates et al. 2018; Frey et al. 2020). Our results do not provide strong support for a symmoriiform-holocephalan clade, despite inclusion of an endoskeletal operculum character coded identically in both
Chondrichthyans represent a specialized group of fishes, at the base of which is a paraphyletic stem group of ‘acanthodians’ (Davis et al. 2012; Maisey et al. 2017). Some acanthodians had an operculum formed from a series of bony plates, supposedly of dermal origin (Eaton Jr. 1939), and thus these are non-homologous with the endoskeletal cartilaginous supports of the chimaeroid operculum. In developmental studies, holocephalan embryos have shown slightly elongated branchial rays, posterior to the hyoid arch (Gillis et al. 2011). If the structure identified here is functionally an operculum, then a hyoid operculum may have been repeatedly acquired and lost among chondrichthyans, or the operculate structure may be plesiomorphic here, due to the deep divergence between elasmobranchs and holocephalans, in which case the structure was lost in all non-operculate chondrichthyan lineages.
in situ dentition, and preservation of thin cartilaginous structures such as the branchial rays are examples of the remarkable information available from Fayetteville Shale fishes. Continued study of fossils from this formation will likely yield more data that can improve understanding of chondrichthyan morphology and evolution.
R. Mapes and G. Mapes collected the holotype specimen and donated it to the AMNH. A. Gishlick facilitated access to AMNH fossil fishes. V. Lee repaired and rehoused AMNH FF 20509. L. Meeker photographed the specimen. H. Towbin and M. Chase assisted with CT imaging at the AMNH Microscopy and Imaging Facility. Use of CT segmentation software at Cal Poly Humboldt was supported by NSF grant MRI 1920204. C. Burrow and P. Godoy contributed to discussions regarding the phylogenetic analysis. We thank several reviewers including M. Brazeau, R. Dearden, and J. P. Hodnett for their thoughtful comments on this manuscript.
Surface files (.stl) of teeth, denticles, cranial fragments, jaws, and gill arch elements have been deposited in MorphoSource under the project title “
The Nexus file (Appendix 2) of the character matrix is available on MorphoBank under matrix 28783 as part of Project 3996 (“A new operculate symmoriiform chondrichthyan from the Late Mississippian Fayetteville Shale [Arkansas, United States]”), alongside input files for the parsimony and likelihood analyses, as well as tree file output from our analyses and a tree file of the strict consensus tree. All of these can be found at http://morphobank.org/permalink/?P3996.
Supplementary material on
Chondrichthyan phylogeny including
A, entire specimen, ventral view. Abbreviations:
mc, Meckel’s cartilage;
op, opercular flap;
B, detail of opercular flap. Scale bars: A, 5 cm; B, 2 cm. Photographs by Lorraine Meeker.
A, ventral view of posterior half of specimen, anterior to top;
B, ventral view of anterior part of the specimen, anterior to right, with opercular elements removed;
C, dorsal view of anterior half of specimen, anterior to right. Note loss of anterior portion of cranium due to taphonomy. Abbreviations:
add, depression potentially for adductor musculature;
c, cranium;
cb, ceratobranchials;
ch, ceratohyals;
d, denticles;
mc, Meckel’s cartilage;
op, opercular cartilage;
pa f, pharyngeal arch fragments;
pq, palatoquadrate fragment;
t, teeth. Scale bars: 5 cm.
A, Tooth families of
B, detail of tooth batteries in axial section, anterior to right, showing denticles between each tooth family, on the scalloped Meckel’s cartilage. Abbreviations:
d, denticles;
mc, Meckel’s cartilage; t, teeth. Scale bars: A, B, 4 cm; C, 2 mm.
arrowhead) revealed by CT scanning. Anterior to top, ventral view in frontal section. Scale bar: 2 cm.
A, cross section through the posterior half of the specimen, at the point indicated by the inset 3D reconstruction, showing multilayered opercular structure made of cartilaginous branchial rays, fused to one another. Layer of rays from the first branchial arch (behind the hyoid arch) indicated by the
second-from-left arrowhead. The
right-most arrowheadmarks the second row of pharyngeal rays (first branchial arch); another layer (second branchial arch) can be found just beneath it (
second arrowhead from the right). Hyoid operculum appears to fold at the arrowhead furthest to the left;
B, longitudinal section of posterior half of AMNH FF 20509 showing hyoid operculum (
right-most arrowhead) and additional elongate branchial rays (left two arrowheads). Anterior to left, slice location within the specimen indicated by 3D reconstruction in the upper right corner. Scale bars: 5 cm.
A, dorsal view;
B, ventral view. Anterior to top. Abbreviations:
da?, possible opening for the dorsal aorta;
occ, occiput;
ot cap, otic capsule. Scale bar: 3 cm.
IEEE Transactions on Automated Control19: 716-723. https://doi.org/10.1109/TAC.1974.1100705
Gamma Spectrometry and Geochemical Investigation of the Mississippian (Chesterian) Fayetteville Shale and Imo Shale, Arkoma Basin, Arkansas. Master of Science of the Oklahoma State University, 138 p. https://api.core.ac.uk/oai/oai:shareok.org:11244/7989
Chlamydoselachus anguineus.
Acta ZoologicaIV: 1-221. https://doi.org/10.1111/j.1463-6395.1923.tb00161.x
Nature457 (7227): 305-308. https://doi.org/10.1038/nature07436
Nature Ecology & Evolution. https://doi.org/10.1101/2020.06.09.132027
inPradel A., Denton J. S. S. & Janvier P. (eds),
Ancient Fishes and their Living Relatives: a Tribute to John G. Maisey. Verlag Dr. Friedrich Pfeil, München: 93-100.
Papers in Palaeontology4 (3): 349-362. https://doi.org/10.1002/spp2.1110
Chondrichthyes II: Mesozoic and Cenozoic Elasmobranchii,
inSchultze H.-P. (ed.),
Handbook of Paleoichthyology. Gustav Fischer Verlag, New York: 1-193.
Proceedings of the Royal Society B285 (1870): 20172418. https://doi.org/10.1098/RSPB.2017.2418
Nature541 (7636): 208-211. https://doi.org/10.1038/nature20806
Journal of Vertebrate Paleontology21 (3): 438-459. https://www.jstor.org/stable/20061974. https://doi.org/10.1671/0272-4634(2001)021[0438:ANSCFT]2.0.CO;2
Transactions of the Royal Society of Edinburgh: Earth Sciences89 (02): 63-85. https://doi.org/10.1017/S026359330000701X
Science Advances5 (9): 1-8. https://doi.org/10.1126/sciadv.aax2742
Nature486 (7402): 247-250. https://doi.org/10.1038/nature11080
Journal of Morphology273 (5): 461-479. https://doi.org/10.1002/jmor.11035
Journal of Anatomy238 (5): 1082-1105. https://doi.org/10.1111/joa.13362
Nature Communications10 (1): 1-7. https://doi.org/10.1038/s41467-019-10032-3
Transactions of the Royal Society of Edinburgh70: 63-109. https://doi.org/10.1017/S0080456800012898
American Museum Novitates(3119): 1-86. https://www.biodiversitylibrary.org/page/62191374
inAnnual Meeting of the Geological Society of America, Denver, Colorado. Wetlands Paleoecology Through Time Symposium, Abstract #3. https://gsa.confex.com/gsa/2002AM/webprogram/Session2773.html
Floristic Survey and Paleoecology of the Fayetteville Formation (Upper Mississippian) of Northwestern Arkansas, United States. PhD Thesis, Ohio University.
Rhynchosperma quinnii(Medullosaceae): a permineralized ovule from the Chesterian (Upper Mississippian) Fayetteville Formation of Arkansas.
American Journal of Botany89 (11): 1799-1808. https://www.jstor.org/stable/4124106
Medullosa steiniisp. nov., a seed fern vine from the Upper Mississippian.
Review of Palaeobotany and Palynology124: 307-342. https://doi.org/10.1016/S0034-6667(02)00254-3
Geological Society of America Special Paper399: 127-137. https://doi.org/10.1130/2006.2399(06)
Copeia1: 42-46. https://doi.org/10.2307/1436015
Communications Biology3 (681): 1-10. https://doi.org/10.1038/s42003-020-01394-2
Nature520 (7545): 82-85. https://doi.org/10.1038/nature14065
Proceedings of the National Academy of Sciences108 (4): 1507-1512. https://doi.org/10.1073/pnas.1012968108
Chondrichthyes – Paleozoic Elasmobranchii: Teeth,
inSchultze H. P. (ed.),
Handbook of Paleoichthyology.Volume 3D. Verlag Dr. Friedrich Pfeil, München, 168 p.
Palaeontology50 (2): 305-322. https://doi.org/10.1111/j.1475-4983.2006.00633.x
Annals of the New York Academy of Sciences20 (3): 189-238. https://doi.org/10.1111/j.1749-6632.1910.tb55149.x
Cladistics: 1-22. https://doi.org/10.1111/cla.12380
Cladistics24: 758-773. https://doi.org/10.1111/j.1096-0031.2008.00209.x
Cladistics24: 774-786. https://doi.org/10.1111/j.1096-0031.2008.00217.x
Revision of Some of Girty’s Invertebrate Fossils from the Fayetteville Shale (Mississippian) of Arkansas and Oklahoma. US Government Printing Office, Washington, 94 p. (Geological Survey Professional Paper; 606).
Debeerius ellefseni(fam. nov., gen. nov., spec. nov.), an autodiastylic Chondrichthyan from the Mississippian Bear Gulch Limestone of Montana (USA), the relationships of the Chondrichthyes, and comments on gnathostome evolution.
Journal of Morphology243: 219-245. https://doi.org/10.1002/(SICI)1097-4687(200003)243:3<219::AID-JMOR1>3.0.CO;2-1
Acta Zoologica90 (SUPPL. 1): 134-151. https://doi.org/10.1111/j.1463-6395.2008.00371.x
inCarrier J. C., Musick J. A. & Heithaus M. R. (eds),
Biology of Sharks and their Relatives. CRC Press: 3-30. https://doi.org/10.1201/b11867-8
Chlamydoselachus anguineus,
inGudger E. W. (ed.),
The Bashford Dean Memorial Volume: Archaic Fishes. American Museum of Natural History, New York: 247-319.
Journal of Vertebrate Paleontology26 (3): 526-537. https://doi.org/10.1671/0272-4634(2006)26
Historical Biology24 (4): 381-395. https://doi.org/10.1080/08912963.2012.683193
inLucas S. G. & Sullivan R. M. (eds), Fossil Vertebrates in New Mexico.
New Mexico Museum of Natural History and Science Bulletin68: 51-64.
Revista Brasileira de Paleontologia8 (2): 127-138. https://doi.org/10.4072/rbp.2005.2.05
Pleuracanthus sessilisund von
Polyacrodus(
Hybodus)
hauffianus.
Sitzungsberichte der Gesellschaft naturforschender Freunde zu Berlin1906 (6): 155-159.
Early Vertebrates. Clarendon Press Oxford Science Publications, Bath, 408 p. (Oxford Monographs on Geology and Geophysics).
Swiss Journal of Palaeontology142 (1): 2. https://doi.org/10.1186/s13358-023-00266-6
Systematic Biology50 (6): 913-925. https://doi.org/10.1080/106351501753462876
Harpagofututor volsellorhinusnew genus and species (Chondrichthyes, Chondrenchelyiformes) from the Namurian Bear Gulch Limestone,
Chondrenchelys problematicaTraquair (Visean), and their sexual dimorphism.
Journal of Paleontology56 (4): 938-958. https://www.jstor.org/stable/1304712
Falcatus falcatus(St. John and Worthen), a Mississippian Stethacanthid Chondrichthyan from the Bear Gulch Limestone of Montana.
Journal of Vertebrate Paleontology5 (1): 1-19. https://www.jstor.org/stable/4523026. https://doi.org/10.1080/02724634.1985.10011842
Damocles serratus, nov. gen. et sp. (Elasmobranchii: Cladodontida) from the Upper Mississippian Bear Gulch Limestone of Montana.
Journal of Vertebrate Paleontology6 (1): 12-19. https://www.jstor.org/stable/4523069
Carcharopsis worthenifrom the Fayetteville Formation (Mississippian) of Arkansas.
Journal of Paleontology58 (3): 709-717. https://www.jstor.org/stable/1304913
Journal of Vertebrate Paleontology9 (2): 174-190. https://www.jstor.org/stable/4523252. https://doi.org/10.1080/02724634.1989.10011751
Bulletin of the American Museum of Natural History307: 1-122. https://doi.org/10.1206/0003-0090(2007)307[1:TBIPSA]2.0.CO;2
Denaea fournieri(Chondrichthyes, Symmoriiformes) from the Lower Carboniferous of Belgium.
Acta Geologica Polonica58 (2): 185-190 [retried from https://geojournals.pgi.gov.pl/agp/article/view/9999].
Journal of Vertebrate Paleontology29 (1): 14-24. https://www.jstor.org/stable/20491065. https://doi.org/10.1671/039.029.0130
Journal of Fish Biology80 (5): 918-951. https://doi.org/10.1111/j.1095-8649.2012.03245.x
Journal of Vertebrate Paleontology21 (4): 702-713. https://www.jstor.org/stable/20062011. https://doi.org/ 10.1671/0272-4634(2001)021[0702:APCBFT]2.0.CO;2
Journal of Morphology275 (5): 586-596. https://doi.org/10.1002/jmor.20242
Doliodus: Bridging the ‘Acanthodian’-Chondrichthyan divide.
American Museum Novitates3875: 1-15. https://doi.org/10.1206/3875.1
Doliodusand Pucapampellids: Contrasting perspectives on stem chondrichthyan morphology,
inJohanson Z. (ed.),
Evolution and Development of Fishes. Cambridge University Press, Cambridge: 87-109.
Journal of Fish Biology98 (4): 919-941. https://doi.org/10.1111/jfb.14376
The University of Kansas Paleontological Contributions64: 1-75. http://hdl.handle.net/1808/3838
Ammonoid Taphonomy, Biostratigraphy, and Lithostratigraphy of the Fayetteville Formation (Mississippian-Chesterian) in its Type Area, Northwest Arkansas. Unpublished PhD thesis, University of Iowa, 141 p.
inGreenwood P. H., Miles R. S. & Patterson C. (eds),
Interrelationships of Fishes. Vol. 53, Supplement no. 1.
Zoological Journal of the Linnean Society.
Environmental Geosciences11 (2): 99-111. https://doi.org/10.1306/eg.01080403018
PLoS ONE6 (9): 1-22. https://doi.org/10.1371/journal.pone.0024938
PLoS ONE8 (6): 4-9. https://doi.org/10.1371/journal.pone.0066988
Nature509 (7502): 608-11. https://doi.org/10.1038/nature13195
inPradel A., Denton J. S. S. & Janvier P. (eds.),
Ancient Fishes and their Living Relatives: a Tribute to John G. Maisey. Verlag Dr. Friedrich Pfeil, München: 183-192.
Pala
̈
ontologische Zeitschrift52 (1/2): 110-128. https://doi.org/10.1007/BF03006733
Zoologica Scripta2: 231-250. https://doi.org/10.1111/j.1463-6409.1974.tb00753.x
Nature Methods9 (7): 676-682. https://doi.org/10.1038/nmeth.2019.Fiji
Journal of Anatomy229 (5): 681-702. https://doi.org/10.1111/joa.12508
Cladodus’
neilsoniTraquair: a primitive shark from the Lower Carboniferous of East Kilbridge, Scotland.
Palaeontology43 (1): 153-172. https://doi.org/10.1111/1475-4983.00122
Chlamydoselachus anguineusGarman,
inGudger E. W. (ed.),
The Bashford Dean Memorial Volume: Archaic Fishes. American Museum of Natural History, New York: 335-519.
Proceedings of the Royal Society B: Biological Sciences276 (1660): 1225-1233. https://doi.org/10.1098/rspb.2008.1526
Bioinformatics30: 1312-1313. https://doi.org/10.1093/bioinformatics/btu033
Durnonovariaodus maiseyigen. et sp. nov., a new hybodontiform shark-like chondrichthyan from the Upper Jurassic Kimmeridge Clay Formation of England.
PeerJ9 (e11362): 1-30. https://doi.org/10.7717/peerj.11362
Biological Journal of the Linnean Society75 (4): 483-502. https://doi.org/10.1046/j.1095-8312.2002.00037.x
Journal of Vertebrate Paleontology21 (2): 214-226. https://doi.org/10.1671/0272-4634(2001)021[0214:TRICSW]2.0.CO;2
The ‘Cladodont Level’ Sharks of the Pennsylvanian Black Shales of Central North America. PhD Thesis, University of Kansas, 203 p.
Chondrichthyes I, Paleozoic Elasmobranchii,
inSchultze H.-P. (ed.),
Handbook of Paleoichthyology. Volume 3A. Gustav Fischer Verlag, Stuttgart: 1-115.
Paleontology18 (2): 333-341. https://www.biodiversitylibrary.org/page/49759838
Nature502 (7470): 188-193. https://doi.org/10.1038/nature12617