Rediscovery of the Type Specimens of the Sarcopterygian Fishes Onychodus sigmoides and Onychodus hopkinsi from the Devonian of Ohio, USA

Abstract

John Strong Newberry described three species of the lobe-fin fish Onychodus (Osteichthyes, Sarcopterygii, Onychodontida) based on parasymphysial teeth, or tusks. Two species, Onychodus sigmoides Newberry, 1857 (type species of the genus) and Onychodus hopkinsi Newberry, 1857, were described from the “fish beds” in the Delaware Limestone (Middle Devonian, Eifelian) of Delaware, Ohio, USA; and one species, Onychodus ortoni Newberry, 1889, was described from the Ohio Shale, Huron Member (Upper Devonian, Famennian) of Perry Township, Franklin County, Ohio. In 1873, Newberry replaced the original species-group definition of O. hopkinsi with a definition based on teeth of different morphology from the West Falls Group (Upper Devonian, Frasnian) of Franklin, New York. Specimens of Newberry’s original Onychodus material, including the primary types, which were long assumed to be lost, have been rediscovered in a 19th-century collection. They show O. hopkinsi to be a junior synonym of O. sigmoides and clarify the species definition of O. sigmoides. Onychodus sigmoides, which is recognized from Middle Devonian strata of the Appalachian Basin in the United States and Canada, shows two end-member shapes of teeth on the parasymphysial whorl: procurved (arcuate) or nearly so proximally and recurved distally (anteriorly). Small teeth are commonly more slender than large teeth, which are robust. Parasymphysial teeth from the Upper Devonian of Ohio and New York are referred to O. ortoni.

1. Introduction

Sarcopterygian, or lobe-fin, fishes are of considerable interest to paleontologists and neontologists alike, as they are the group within which limbed tetrapodomorphs evolved [1,2,3]. Sarcopterygians are osteichthyans, or bony fishes, which are characterized by monobasal articulation with fleshy paired fins extending in a basal lobe [3]. Both the pectoral and pelvic fins resemble tetrapodomorph limbs in their manner of articulation, and these fins were modified into the legs of early tetrapods. Another character that is unique to sarcopterygians is the presence of enamel covering the entire surface of the tooth [3]. In modern times, sarcopterygian fishes are confined to four known species of lungfish and two species of coelacanths. In the middle to late Paleozoic, however, sarcopterygians were much more diverse [1,3,4], and some, such as the Devonian-age Onychodus, were apex marine predators [5,6].

Onychodus Newberry, 1857 [7] is one of the earliest described sarcopterygians. It was named by John Strong Newberry, who authored numerous Devonian, Carboniferous, and Jurassic fish taxa [8]. He named three species of Onychodus from the large teeth or tusks that emanate from the parasymphysial plate (Figure 1, Figure 2, Figure 3 and Figure 4), the anteriormost bone of the lower jaw. When found attached to the parasymphysial, the teeth form a characteristic whorl (Figure 2 and Figure 3), but it is much more common to find shed or detached teeth (Figure 1). In 1857, Newberry [7] described the genus Onychodus and included in it two new species, O. sigmoides Newberry, 1857 (the type species; Figure 1) and O. hopkinsi Newberry, 1857 (Figure 2) and, in 1889, he described another species, O. ortoni Newberry, 1889 [9] (Figure 3B,C). All three species were described from Devonian strata of Ohio, USA [7,9,10,11,12,13].

Onychodus sigmoides and O. hopkinsi have been the subjects of uncertainty and ambiguity from the time of their initial description [7]. These issues have been as follows:

• Their specific geographic provenance: whether from Milford, Delaware, or Sandusky, Ohio.
• Their stratigraphic provenance: Columbus Limestone or Delaware Limestone.
• Their age: Early Devonian or Middle Devonian.
• Their nomenclature: two distinct species or one species having variability in the shape of the parasymphysial teeth.
• Which specimens constitute the name-bearers: whether these are the specimens examined, described, and illustrated by Newberry in the 1850s and 1860s [7,11,12], or the ones illustrated or alluded to in the 1870s and 1880s [9,10,13].

Newberry did not illustrate the type specimens in his 1857 paper, and subsequent attempts to characterize O. sigmoides and O. hopkinsi have been without the benefit of the primary type material. Subsequent work has relied to a large extent on the illustrations of specimens and statements published by Newberry in 1873 [10], and some of the textual information contradicts the information he published earlier [7,8]. Subsequent publications by other authors commonly have emphasized specimens from Middle or Upper Devonian strata of New York or elsewhere [6,14,15,16], rather than ones collected from localities cited in Newberry’s original work.

For more than 100 years, Newberry’s type specimens of O. sigmoides and O. hopkinsi, which were in a collection of Devonian fossil fishes originally assembled in part by R.P. Mann, have been assumed to be lost [6] despite the existence of some published clues to their whereabouts. Newberry [7] (p. 120) stated that “Prof. F. Merrick has had the kindness to send me for examination a collection of exceptionally fine ichthyolites, from the Cliff limestone of Delaware, Ohio …” At that time, Rev. Dr. Frederick Merrick was President and a Professor of Natural History at Ohio Wesleyan University. Additionally, Newberry stated that he was “indebted for similar favors to … Dr. R.P. Mann, of Milford, Ohio”. In 1873, Newberry [10] (p. 265) stated that the specimens he described in a paper that was read before the National Institute for the Promotion of Science, Washington, D.C., from Delaware County, Ohio, and collected by Dr. R.P. Mann of New Milford, Ohio, were “in the museum of the Wesleyan University at Delaware, Ohio”. Additionally, Stauffer [17] (pp. 87–88) and Crowl [18] (p. 16) both made passing references to type Devonian fishes in the geological collection at Ohio Wesleyan University (OWU), Delaware, Ohio. However, in an extensive review of Devonian fishes from North America, Eastman [14] did not illustrate or indicate the location of the specimens. Despite illustrating a skull of the placoderm Macropetalichthyes rapheidolabis in the OWU collection [14] (pl. 11), he did not report whether any Onychodus were in this collection and instead illustrated specimens that he referred to as O. sigmoides and O. hopkinsi from New York.

In 2024–2025, much of the geological collection of Ohio Wesleyan University was transferred to the Orton Geological Museum at The Ohio State University (OSU) in Columbus, Ohio, to safeguard its future and conserve as much of its history as possible. In the process of moving the collection, fossil fish specimens studied by J.S. Newberry in the 1850s through the 1880s were rediscovered [19]. The collection includes the type material of O. sigmoides (Figure 1A,B), O. hopkinsi (Figure 2A,B), and other Devonian species from Ohio, plus a specimen of Onychodus, apparently an intended type, from Franklin, New York (Figure 3A).

2. Scientific and Historical Importance

The Devonian fossil fish material examined and described by J.S. Newberry, which was part of the Mann cabinet (collection) at Ohio Wesleyan University, represents some of the earliest-reported fossil fishes from North America. Their published documentation [7,9,10,11,12,13] greatly expanded the scientific understanding of early vertebrate biodiversity and evolution during the Devonian, or so-called “Age of Fishes” [14,20], a critical time in the evolutionary history of vertebrate animals. The descriptions of Devonian fossil fishes represent one of the most important scientific legacies of John Strong Newberry, who was one of the leading figures in American science during the mid-to-late 19th century [8,19,21].

Rediscovery of the Onychodus types allows a reevaluation of the morphology and nomenclature of the two first-named species in the genus, extending our understanding of a clade that occupies an uncertain phylogenetic relationship among the early sarcopterygians [6,22,23,24,25,26,27,28,29,30,31,32]. The specimens prompt reevaluation of previous interpretations of O. sigmoides and O. hopkinsi. The types and other specimens examined by Newberry in his earliest work on Onychodus open the way for continuing studies that are focused on directly comparing O. sigmoides and O. hopkinsi to other species that are referred to the genus from North America and elsewhere. Onychodus had a pan-global distribution in the Devonian [31], and was an apex predator in marine ecosystems [5,6]; thus, the rediscovery of this material takes on added importance from the standpoint of understanding biodiversity during the Devonian Period.

Another reason the type and other examined specimens are important is that they represent the physical documentation for one of the earliest described sedimentary deposits that is unusually rich in paleontological information from North America [7,10,33]. Fossil fish material was so abundant in the Delaware Limestone that some intervals were known to quarry workers as “fish beds” [10,33]. From the 20th century, such deposits have been celebrated by paleontologists and recognized as Konzentrat–Lagerstätten (concentration deposits) [34], one of the end-member types of Fossil–Lagerstätten [35]. The fish beds from which the specimens illustrated here were collected [33] can no longer be collected at their place of discovery, as the quarry has been flooded (Figure 5).

This story of rediscovery underscores the continued need to conserve natural history collections, as they are an invaluable and irreplaceable repository of biodiversity data. A future, more detailed study of Newberry’s type and examined specimens, which were fortunately retained at Ohio Wesleyan University, is expected to shed further light on the affinities of onychodontiformes and on the taphonomic conditions leading to deposition of the rich Delaware fish beds.

3. Materials and Methods

Most examined specimens were originally part of the geological collection of Ohio Wesleyan University (OWU), which included the R.P. Mann cabinet of minerals, rocks, and fossils. Those specimens, and others illustrated here, are now reposited in the Orton Geological Museum at The Ohio State University (OSU). Additional Devonian fish specimens collected by J.S. Newberry, E. Orton, H. Herzer, L.G. Westgate, E.T. Nelson, and perhaps others between about 1850 and 1925, originally in the Ohio Wesleyan collection, are also now in the Orton Museum.

The specimens were photographed with a Canon EOS R6 Mark II camera(Canon Inc., Tokyo, Japan). The images were assembled in Adobe Photoshop(Canon Inc., Tokyo, Japan).

4. Nomenclature of Early-Named Onychodus Species

4.1. Onychodus sigmoides and O. hopkinsi from Ohio

In 1853, J.S. Newberry illustrated the remains of two genera of fishes collected from the “Cliff limestone” of Ohio (Devonian) [11]. One woodcut was a composite illustration of the cranial roof (part of the dermal armor) of a placoderm later named Agassizichthyes manni [11] (fig. 1); today, it is classified as Macropetalichthyes rapheidolabis. The other woodcut, also a composite figure, illustrated two teeth [11] (fig. 2; herein, Figure 1C), and it is clear from the 1857 description that these teeth belong to a sarcopterygian Newberry named Onychodus sigmoides [7]. Newberry noted in words the “doubly curved” morphology of the teeth (Figure 4A) but did not illustrate O. sigmoides in 1857. As he did not publish any other specimens under that name until 1873 [10], the two teeth must be considered as the syntypes; they are the only specimens that Newberry assuredly examined in preparing his 1857 paper. These two specimens, preserved on separate limestone slabs (Figure 1A,B), are among the rediscovered types in the OWU collection. The larger and better preserved parasymphysial tooth, 55 mm in length (Figure 1A), is here designated the lectotype for the purpose of stabilizing the species definition, and the smaller specimen (Figure 1B) is designated as the paralectotype.

Newberry also described O. hopkinsi, without an illustration, in his 1857 paper [7]. He stated that the teeth were simply but strongly curved, conical, and acute at their tips, with seven or more teeth on the parasymphysial tooth whorl. In 1862, Newberry [12] illustrated a parasymphysial whorl with seven teeth under the name O. hopkinsi [12] (caption to fig. 3, p. 77; herein, Figure 2C). This is the only specimen Newberry ever illustrated using this species-group name and, as such, it is considered the holotype by monotypy. The woodcut matches the description: most notably, it has seven teeth, which is highly unusual for Onychodus. The holotype of O. hopkinsi, both part and counterpart slabs, are among the rediscovered types in the OWU collection (Figure 2A,B). There is little doubt that this is the specimen illustrated and described by Newberry [12], as it is the only parasymphysial whorl known retaining seven teeth (contra Lesley [36]; see also Newberry [10] (p. 297)). The original figure and the specimen both show recurvature in some of the teeth, which Newberry [7] cited as the principal defining characteristic of O. sigmoides. Newberry [10] (p. 145) apparently referred to this distinctive specimen again in 1873, stating “The mandibles … embraced between their anterior extremities an arch of bone from which sprang a crest of seven hooked or sigmoidally curved conical teeth”. A study of the specimen (Figure 2A,B) shows that not all the teeth are distinctly recurved; the proximal ones are more arcuate (procurved).

In 1873, Newberry [10] re-illustrated his woodcut of a parasymphysial whorl with seven teeth but labeled the specimen as O. sigmoides (Figure 2D), thus making O. hopkinsi a junior subjective synonym of O. sigmoides. By the time of this publication, the limestone from which his Onychodus specimens were collected was referred to as “Corniferous limestone”. Newberry stated that his inclusion of O. hopkinsi in the “Corniferous limestone” fauna was in error, and recognized O. hopkinsi as occurring only in the “Chemung” (Upper Devonian) of New York. His paper was republished in German in 1874 [13], and he repeated that information in 1889 [9].

4.2. Onychodus hopkinsi from New York

Newberry redescribed O. hopkinsi in 1889 [9], stating, “In the Chemung rocks at Franklin, Delaware County, N.Y., Mr. Andrew Way collected many detached teeth to which I have given the above name. These teeth are generally about one inch in length, conical, acute, and simply curved. Occasionally, however, the point is slightly turned forward, giving a hint of the sigmoidal curve which is so conspicuous a feature in the great species of the Corniferous limestone (O. sigmoides)”. He stated further that the genus was represented to the time of publication by only three species, O. sigmoides from the “Corniferous limestone”, O. ortoni from the Huron Shale, and O. hopkinsi from the “Chemung”. Newberry never illustrated a specimen under the name O. hopkinsi from New York, but one specimen inferred to have been collected by Andrew Way (referred to as “Mr. J.M. Way” by Newberry in 1889 [9] (p. 270)) from the Chemung of New York, and passed to J.S. Newberry by way of Edward Orton, was discovered in the OWU collection (Figure 3A). This specimen is interpreted as an intended type pertaining to Newberry’s revised definition of O. hopkinsi, but which was never published as such. In the 1800s, there were no formal nomenclatural rules, and transposing species definitions did occasionally take place, as evinced by this example. Eastman [14] illustrated three specimens from Franklin, New York, under the name O. hopkinsi: two isolated parasymphysial teeth and a parasymphysial whorl having five teeth. These specimens also may have been examined by Newberry around 1889 before being reposited in the New York State Museum (NYSM). Newberry [9] (p. 270) indicated that after Way’s death in the American Civil War, “his collection has been disposed of, perhaps scattered”.

4.3. Onychodus ortoni from Ohio

In 1889, Newberry [9] described the third North American species of Onychodus, O. ortoni. This species, based on a single, moderately large parasymphysial whorl with six teeth (Figure 3C), was collected from black shale in the lower part of the Huron Member of the Ohio Shale (Upper Devonian, Famennian), in Perry Township, Franklin County, Ohio; the locality is inferred to be Flint Run, which is now on the property of Camp Mary Orton [37]. This specimen was reposited in the Orton Geological Museum (OSU) in about 1889. Another, somewhat smaller parasymphysial whorl with six teeth, evidently a topotype, is illustrated here (Figure 3D); it is from a carbonate concretion in the lower Huron Member. These two specimens show that the parasymphysial teeth reach moderately large size, and are simply and evenly curved (procurved) except for a slight deflection or recurvature at the distal tips.

Newberry [9], in a differential comparison of the parasymphysial whorls of O. ortoni and O. hopkinsi, recorded that “In size and general aspect this arch with its row of teeth resembles the corresponding organ in Onychodus Hopkinsi of the Chemung, but may be distinguished from that at a glance by the very different manner in which the teeth are attached to the bony arch. In O. sigmoides and O. Hopkinsi teeth of the intermandibular crest are expanded at the base with root-like projections on either side, which clasp the summit of the arched bone. … As their attachment was only ligamentous, they were deciduous …” In O. ortoni, “… they are implanted in the substance of the arch …” Re-examination of the specimens of all three species from Ohio, and the intended type of O. hopkinsi from New York, indicates that the apparent difference in attachment mode of the teeth in the holotype of O. ortoni (Figure 3C) is a taphonomic artifact: this specimen, preserved in black shale, shows the parasymphysial plate enclosed within a small concretion that closely approximates the shape of the bone but is somewhat larger than the bone within. Growth lines indicating the surface of the parasymphysial plate are not exposed. As a result of the concretionary growth, the teeth appear to be deeply embedded in the structure. Teeth in the topotype specimen illustrated here, mechanically prepared from a carbonate concretion, show attachment that is indistinguishable from that of the other species (Figure 3D).

The Ohio Shale, which yielded the holotype of O. ortoni, has been discussed as a Fossil–Lagerstätte [34,37]. Seilacher and others [34] described the black shale as a Konservat–Lagerstätte (conservation deposit) of Holzmaden-type.

5. Geographic and Stratigraphic Provenance of Onychodus sigmoides and O. hopkinsi

The type specimens of Newberry’s fishes from Ohio, at the time they were rediscovered in the OWU collection, were without labels and were unnumbered; a catalog of the collection was not found. As a result, it has been necessary to recover information about where, when, and by whom they were collected, primarily from published information. That information is scant, and some is ambiguous. Some of the types and other examined fishes have indications on the matrix that presumably reflect their connection to Newberry’s early studies and publications, such as white or yellow paint spots, or glued paper stars or diamonds. The lectotype of O. sigmoides has a white paint spot, and the intended type of O. hopkinsi from Franklin, New York, has a yellow paint spot; both are on the reverse side of the slab.

Today, what was once called the “Cliff limestone” and later the “Corniferous limestone” is classified as the Columbus Limestone–Delaware Limestone succession [37,38]. Prosser [38] provided a detailed history of the nomenclature of the Delaware Limestone, attributing its origin to Orton [33], who distinguished it from the Columbus Limestone [39] (pp. 143–144), and supplanting all previous names, including “Sandusky limestone” [39] (p. 143). Orton [33] (pp. 606–607) named the Delaware Limestone for “its occurrence at Delaware”, and Westgate [40] (p. 27) later specified the type section of the formation as “the quarry in the south bank of Delaware Run, just east of the Hocking Valley Railway”. This quarry was called “Campbell’s quarry” by Stauffer [17]. By 1955, the abandoned quarry became a public park, Blue Limestone Park (Figure 5), and the deep pit is now water-filled (Figure 5C).

During some intervals of the Early and Middle Devonian (Emsian–Eifelian ages), carbonate strata extended over expansive areas of North America, including the Appalachian Basin region, and the Columbus and Delaware limestones are part of this set of strata [41,42]. In Ohio, the ages of the units are constrained biostratigraphically [17,43,44,45,46,47,48] and by a volcanic airfall tephra. The Tioga A-G cluster of K-bentonites has been correlated into central Ohio [41,42], where it occurs in the upper part of the Columbus Limestone and the lower part of the Delaware Limestone. The Tioga B K-bentonite bed [44], which is at the Columbus Limestone–Delaware Limestone contact, approximately 2 m stratigraphically below the base of the Delaware Limestone fish beds at Delaware, Ohio, has yielded an age of 390 ± 0.05 Ma [45]. On this basis, the Columbus Limestone is inferred to be Emsian (upper Lower Devonian) to Eifelian (lower Middle Devonian), and the Delaware Limestone is Eifelian (lower Middle Devonian) [41].

Newberry, in 1857 [7] (p. 124), stated that the teeth of O. sigmoides were “not uncommon in the Cliff limestone of Ohio … at Milford, Delaware, and Sandusky, Ohio”. On the same page, he indicated that the teeth of O. hopkinsi are from Milford, Ohio, and were collected by Dr. Mann. This initial description has caused uncertainty about the geographic and stratigraphic provenance of the type specimens.

Milford, Ohio, which has also been published as New Milford [19] and Milford Centre [49], is a place name known from at least three counties in Ohio. The historical records indicate that Dr. Reuben P. Mann, the collector, resided in Milford, Union County, and it is reasonable to infer that Newberry mistakenly referred to Mann’s residence as a locality.

In 1873, Newberry [10] (p. 265) wrote “A very fine collection of the ichthyolites of the Corniferous limestone of Delaware county, Ohio, was made … by Dr. Mann, of New Milford. A part of these were described by me in the paper read before the National Institute”. Later, Stauffer [17] (p. 87) indicated that Campbell’s Quarry, in Delaware, Ohio, “is the locality from which a number of types have been collected”, and on p. 88, he clarified which types by stating “fragments of plants and fish bones are commonly found. It was from this locality that many of Newberry’s types were taken”. Crowl [18] (p. 16), in reviewing the history of the Department of Geology and Geography (as it was known in 1979) at Ohio Wesleyan University, noted that “By 1871 Ohio Wesleyan boasted a considerable museum …”, including “… the Mann cabinet of minerals, rocks, and fossils, including type specimens of local Devonian fossil fish”. All of these indicators point to Delaware, Ohio, and specifically the now-abandoned Campbell’s Quarry, as the source of the types of O. sigmoides and O. hopkinsi.

Orton [33] (p. 607) noted that fossil fish pieces are so common in one interval of the Delaware Limestone that it was known to quarry workers as the “fish beds”. Most of Newberry’s examined specimens are probably from this zone of high vertebrate fossil abundance. Newberry [10] referred to this zone, together with bone beds of the “Corniferous” as “fish-beds”. Intervals of the Columbus Limestone and Delaware Limestone that are rich in vertebrate remains, including bone beds [10,46,50] and the Delaware fish beds [33], are zones of fossil concentration (Konzentrat–Lagerstätten), one end-member type of Fossil–Lagerstätten [34,35].

Stauffer [17] (p. 87, 88) provided a written stratigraphic section of Campbell’s Quarry. In 1909, the quarry exposed only Delaware Limestone, and the quarry floor was inferred to be “not far above the “bone-bed” or top of the Columbus limestone”. Onychodus sigmoides was recorded from Horizon 3, a 1.6-m-interval of “Rather massive blue limestone in layers” 15 to 35 cm thick … “and containing little or no chert”. The interval began about 1.35 m above the quarry floor at the time of Stauffer’s writing. By 1926, quarrying had evidently extended downward into the uppermost Columbus Limestone [40], and the quarry was no longer referred to as Campbell’s Quarry. Westgate [40] (fig. 7) illustrated a measured stratigraphic section from this quarry (possibly incorporating additional neighboring quarry pits), showing that the total thickness of the Delaware Limestone is 13 to 16.5 m thick. The Onychodus-bearing interval is labeled “Heavy-bedded, blue-gray, finely crystalline limestone”, a description which matches both Stauffer’s [17] earlier description and the lithology of the rock matrix containing the type Onychodus remains. Westgate also published a contemporary photograph [40] (pl. V B) showing a normal fault cutting through and displacing Delaware Limestone on the south wall of Campbell’s Quarry [40] (fig. 14).

Biostratigraphic evidence supporting the conclusion that the type Onychodus specimens were collected from the Delaware Limestone is the presence of brachiopod fossils, Mucrosprifer consobrinus, on the surface of the slab containing the lectotype of O. sigmoides (Figure 1A). Mucrospirifer consobrinus, a Middle Devonian species, is common in the Delaware Limestone [43] (referred to Delthyrus consobrina, pl. 12, figs. 10, 11).

6. Revised Nomenclature, Biostratigraphy, and Lithofacies Distribution of Appalachian Basin Onychodus Species

6.1. Newberry’s Onychodus Species and Variation in Tooth Morphology

Based on an examination of Newberry’s type materials of O. sigmoides, O. hopkinsi, and O. ortoni, the following systematic revisions of Devonian species from the Appalachian Basin are indicated. Two species are recognizable from parasymphysial teeth: O. sigmoides, which has slender to robust teeth showing a curvature ranging from nearly fully procurved to sigmoidal (procurved in the lower part to recurved distally); and O. ortoni, which has slender and much more arcuate (procurved) teeth that have a minor deflection, if any, at the tips. The parasymphysial teeth of O. sigmoides reach much greater lengths than those of O. ortoni, as known at present.

• Systematic Paleontology (following [37])
• Sarcopterygii Romer, 1955 [51]
• Onychodontida Andrews, 1973 [22] sensu Ciudad Real, Mondéjar Fernández, Vidal, and Botella, 2022 [32]
• Onychodontidae Woodward, 1891 [52]

Genus Onychodus Newberry, 1857

• 1857 Onychodus Newberry, p. 124 (original description).
• 1862 Onychodus Newberry–Newberry, p. 77
• 1873 Onychodus Newberry–Newberry, pp. 296–299.
• 1874 Onychodus Newberry–Newberry, pp. 294–297.
• 1889 Onychodus Newberry–Newberry, pp. 53–56.
• 1889 Onychodus Newberry–Miller, pp. 602–603.
• 1907 Onychodus Newberry–Eastman, pp. 168–169.
• 1918 Onychodus Newberry–Hussakof and Bryant, p. 178.
• 1944 Onychodus Newberry–Wells, p. 43
• 1996 Onychodus Newberry–Hansen, pp. 292–293.
• 1996 Onychodus Newberry–Janvier, pp. 198–199, fig. 4.74C–F.
• 2020 Onychodus Newberry–Mondéjar-Fernández, p. 574.
• 2022 Onychodus Newberry–Ciudad Real, Mondéjar Fernández, Vidal and Botella, p. 96.
• 2024 Onychodus Newberry–Babcock, 2024, pp. e2308621-7–e2308621-9.
• Type species: Onychodus sigmoides Newberry, 1847 (by original designation [7]).
• Diagnosis: Onychodontiform fish reaching a large size and having a single row of large, procurved to sigmoidal teeth on the paired parasymphysial plates; the parasymphysial plates rest on either side of the lower jaw symphysis. The teeth are simple, with plicidentine infolding only at the base, and are covered with finely striated enamel. Premaxilla is slender without a sensory-line canal; the maxilla is large, reaching the preopercular. The caudal fin is epicercal.
• Remarks: Janvier [3] recently provided a concise discussion of the characteristics that diagnose the genera included in the Onychodontiformes (or Onychodontida), and Ciudad Real et al. [32] recently discussed the data matrices used to reconstruct the phylogenetic relationships of the onychodonts among sarcopterygians. These sources and references therein provide further background to the interpretation of Onychodus and its phylogenetic history.

Onychodus is known mostly from separated teeth and bones from North America. A species, O. jandemarrai Andrews, Long, Ahlberg, Barwick and Campbell (2006), from the Gogo Formation (Devonian) of Australia [27], which is known from complete material, is the sole exception. From many localities, Onychodus is best known from its teeth, especially the distinctive parasymphysial teeth. Although the diagnosis of species from the parasymphysial teeth alone is subideal [6], it is practical, and for most species in the genus, it is the primary, if not only, means of distinguishing species. Where complete parasymphysial tooth whorls are known, the parasymphysial teeth show sufficient differences among described species as to be useful for diagnosing species.

The inference that Newberry intended Onychodus sigmoides to be the type species of Onychodus was discussed by Babcock [37]. Mondéjar-Fernández [31] recently discussed species that were referred to as Onychodus and the geographic and stratigraphic distribution of the genus. The genus is restricted to the Devonian [3,31,53]. Onychodontiform fishes became extinct at the end of the Devonian Period [54].

Onychodus sigmoides Newberry, 1857

Figure 1, Figure 2, and Figure 4A

• 1853 Ganoid fish–Newberry, 1853, p. 13, fig. 2.
• 1857 Onychodus sigmoides Newberry, p. 124 (original description).
• 1857 Onychodus hopkinsi Newberry, p. 124 (original description).
• 1862 Onychodus hopkinsi Newberry, fig. 3.
• 1873 Onychodus sigmoides Newberry–Newberry, pp. 299–302, fig. p. 301, pl. 26, figs. 1–5; pl. 27, figs. 1–2.
• 1874 Onychodus sigmoides Newberry–Newberry, pp. 297–299, fig. p. 299, pl. 26, figs. 1–5; pl. 27, figs. 1–2.
• 1878 Onychodus sigmoides Newberry–Orton, p. 625.
• 1889 Onychodus sigmoides Newberry–Newberry, pp. 56–57, pl. 36, figs. 1–4; pl. 37, figs. 1–11.
• 1889 Onychodus sigmoides Newberry–Lesley, pp. 497–498, 8 figs.
• 1889 Onychodus sigmoides Newberry–Miller, p. 603, figs. 1150, 1151.
• 1889 Onychodus hopkinsi Newberry–Miller, p. 603 (in part).
• 1892 Onychodus sigmoides Newberry–Lesley, p. 1160, 5 figs.
• 1902 Onychodus sigmoides Newberry–Hay, p. 363 (see for additional synonymy).
• 1907 Onychodus sigmoides Newberry–Eastman, p. 169 (in part).
• 1907 Onychodus hopkinsi Newberry–Eastman, p. 169 (in part).
• ? 1907 Onychodus sp. indes. cf. O. sigmoides Newberry–Eastman, p. 169, pl. 1, fig. 3.
• 1907 Onychodus (sigmoides ?)–Eastman, p. 169, pl. 3, fig. 2.
• 1908 Onychodus sigmoides Newberry–Hussakof, pp. 57–58.
• 1918 Onychodus sigmoides Newberry–Hussakof and Bryant, pp. 178–181, text fig. 59, pl. 58, figs. 1–3.
• 1933 Onychodus Newberry–Moodie, fig. 20, 2–3.
• 1944 Onychodus sigmoides Newberry–Wells, pp. 43–45, pl. 3, figs. 28–38; pl. 4, figs. 19, 20,; pl. 8, figs. 7–11; text fig. 9.
• 1966 Onychodus sigmoides Newberry–Gardiner, pp. 91–93 (see for additional synonymy).
• 1996 Onychodus sigmoides Newberry–Hansen, pp. 292-293, figs. 21-10.2–21-10.4 (not fig. 21-10.1).
• 2014 Onychodus sp.–Wilson, p. 236, 3 figures p. 237 (not figure at left).
• 2017 Onychodus eriensis Mann, Rudkin, Evans, and Laflamme, pp. 234–241, fig. 2, tab. 1 (original description).
• 2017 Onychodus cf. eriensis Mann, Rudkin, Evans, and Laflamme, pp. 234–241, fig. 3.
• 2017 Onychodus sigmoides Newberry–Mann, Rudkin, Evans, and Laflamme, p. 240, fig. 5.
• Types: Lectotype (designated here), parasymphysial tooth, OSU 54751; and paralectotype, parasymphysial tooth, OSU 54752; both are from the OWU collection; from the Delaware Limestone (Lower Devonian, Eifelian), Campbell’s Quarry (now Blue Limestone Park), Delaware, Delaware County, Ohio, USA (Figure 4). Holotype of Onychodus hopkinsi Newberry, 1857, a subjective junior synonym, parasymphysial whorl, OSU 54753A (part) and counterpart 54753B (counterpart), from the OWU collection; from the Delaware Limestone (Lower Devonian, Eifelian), Campbell’s Quarry, Delaware, Delaware County, OH, USA.
• Diagnosis: Onychodus having four? to seven (usually five or six) slender to robust parasymphysial teeth, reaching about 60 mm in length; they are simple at the tips, procurved to recurved, and if recurved, recurvature occurs in the distal 10–35% of the tooth length above the basal flare. The teeth are morphologically dissimilar along the length of the parasymphysial plate; teeth in the proximal location show little recurvature, and recurvature in the teeth increases distally (anteriorly) along the parasymphysial. The lower jaw is distinctly curved, especially in large specimens; the jaw curves anteriorly from the middle of the dentary to the mandibular ramus; the anterior surface of the dentary ramus is flattened. Jaws are large, with up to about 50 teeth that decrease in size both anteriorly and posteriorly; the teeth are short, conical, and straight to slightly curved; the dermal bone is ornamented with large tubercles.
• Remarks: Prior to the rediscovery of Newberry’s original suite of type and examined specimens of Onychodus from the Delaware Limestone of Delaware, Ohio, and his examined specimen from Franklin, New York, considerable uncertainty about how to classify the North American species of Onychodus was introduced to the literature. Uncertainty arose, in part, from the publication of a distinctive seven-tooth parasymphysial whorl from the “Corniferous limestone” of Ohio under the names O. hopkinsi in 1862 [12] and O. sigmoides in 1873 [10] (reprinted in German, 1874 [13]) using the same woodcut illustration (herein, Figure 2C,D). Further complicating the uncertainty, Newberry [9,10,13] stated that his report [7] of O. hopkinsi from Ohio was erroneous despite his erection of the species-group name using a specimen (Figure 2A–C) from Ohio. Newberry [9,10,13] stated that teeth collected from the “Chemung” of New York instead fulfilled his species definition, but he never illustrated or specifically referred to any specimen as a substitute name-bearer (or neotype) for the species.

Among the specimens in Newberry’s suite of specimens from the OWU collection is one parasymphysial whorl of Onychodus having five teeth (Figure 3A) from the “Chemung” of Franklin, New York, and it is reasonable to presume that this was an intended, substitute name-bearing type of O. hopkinsi. The presence of the holotype of O. hopkinsi in the same collection, however, would invalidate the use of this or any other specimen as a substitute name-bearer according to Articles 73 and 75.8 of the International Code of Zoological Nomenclature [55].

Some of Newberry’s 1873 figures and textual matter [10] were reproduced by him [9] and others in later publications [36,56,57], leading to continued uncertainty about the classification of the parasymphysial teeth of Onychodus from North America. The published opinions of leading specialists also changed, resulting in continued nomenclatural uncertainty. Eastman in Hay [58], for example, reported that O. hopkinsi is a junior synonym of O. sigmoides, but he later treated O. hopkinsi as valid for some onychodontiform material from New York [14].

Gardiner [59] (p. 92) mistakenly referred to four specimens in the AMNH collection, which were illustrated by Newberry in 1873 [10], as “syntypes” of O. sigmoides. Similarly, Mann et al. [6] referred to these and other specimens illustrated by Newberry [10] as the type series of O. sigmoides. Hussakof [60] listed these specimens in a catalog of type fishes in the AMNH. They cannot be unambiguously linked to the original description of O. sigmoides in 1857 [7] and should be considered figured specimens instead of types [60].

A study of the lectotype of O. sigmoides, the holotype of O. hopkinsi, and additional specimens presumably examined by Newberry in the 1850s–1860s, suggests that the parasymphysial teeth in the holotype of O. hopkinsi fall within the range of variation that is observed in both parasymphysial whorls and the isolated teeth of O. sigmoides. Onychodus hopkinsi is thus considered a junior subjective synonym of O. sigmoides. Specimens in the collection show variations in the tooth shape, notably the proportion of procurvature to distal recurvature and relative width of the tooth, and these characteristics track with the size and location on the whorl. Long teeth of O. sigmoides tend to be more sigmoidal (more recurved) than shorter teeth. Recurvature increases not only according to the length of the tooth, but also according to the position on the parasymphysial whorl, with the teeth situated distally (toward the anterior) tending to show more recurvature than the teeth situated proximally. In addition, the parasymphysial teeth of O. sigmoides have relatively slender and robust end-member morphologies. This seems to be at least partly a function of length. The short teeth (<20 mm) tend to be slender above the basal flare and have sharp distal tips, whereas the longer teeth tend to be wider and more robust with somewhat blunter distal tips. Teeth > 35 mm long are consistently robust. Among teeth of intermediate lengths, there is much variation in the width.

The large lectotype tooth of O. sigmoides (Figure 1A) is more robust and more recurved than that of the shorter paralectotype (Figure 1B). However, the distal tip of the paralectotype is still buried in a matrix. The teeth in the holotype of O. hopkinsi (Figure 2A,B) are almost the same length as the paralectotype and show a gradational range of morphologies that includes a recurvature similar to that in the paralectotype. The holotype of O. hopkinsi shows slightly recurved teeth distally and nearly fully procurved teeth proximally, in keeping with the general pattern observed in O. sigmoides.

Eastman’s [14] compilation of Devonian fish has been an influential source for the identification of Onychodus, particularly from New York State, where remains are locally abundant. Uncertainty about the classification of Onychodus teeth, stemming from confusing earlier work and prior to the rediscovery of Newberry’s original examined specimens, however, has evidently led to nomenclatural decisions that can now be clarified and improved. It is reasonable to assume that the teeth referred to by Eastman [14] from the Columbus and Delaware limestones of Ohio, and the Onondaga Limestone of Leroy, New York, belong to O. sigmoides, as now understood. Illustrated specimens from the “Chemung” of Delaware County, New York [14] (p. 196, pl. 1, fig. 14; pl. 3, fig. 3), are similar to Newberry’s specimen illustrated here (Figure 3A), and are now referred to O. ortoni. In contrast, specimens referred by Eastman [14] to O. hopkinsi from the “basal bituminous layer of the Marcellus shale” of upstate New York, all seem referrable to O. sigmoides based on study of numerous specimens in museum and private collections from the contact between the Seneca Member of the Onondaga Limestone and the Union Springs Formation (Marcellus Group). Although Eastman [14] indicated that a “considerable stratigraphic interval” separated the specimens collected from the basal Marcellus in New York from the material of O. sigmoides in the Delaware Limestone of Ohio, the ages of both occurrences are now interpreted to be nearly coeval [41,44]. The Tioga A-G cluster of K-bentonites occurs in this interval, both in the upper Columbus Limestone–lower Delaware Limestone of Ohio and the upper Onondaga–Union Springs Formation [44]. By implication, a scale from the Marcellus Shale at Oran, Onondaga County, New York [14] (p. 196, pl. 3, fig. 2), belongs to O. sigmoides. A broken parasymphysial tooth from the Hamilton Group of Wisconsin, USA [14] (p. 196, pl. 1, fig. 3) may be O. sigmoides, but is too incomplete to be identified with certainty. A broken tooth from the Middle Devonian of the Eifel district, Germany [14] (p. 196, pl. 1, fig. 4), which was reproduced by Wilson [61] (p. 237, left), does not belong to any species of Onychodus known from North America. Some of Eastman’s [14] illustrations were reproduced in subsequent literature [16,58].

Mann et al. [6] described a species of Onychodus, O. eriensis, from the Dundee Formation (Devonian, Eifelian) of Pelee Island, Ontario, Canada. In the Lake Erie Islands region, the Dundee Formation, a limestone unit that is temporally equivalent to the Delaware Limestone [6], is recognized in Ontario. Onychodus sigmoides is present in the Columbus and Delaware limestones in nearby areas of Ohio, including Kelleys Island, Sandusky, and the Firelands area, and it is a reasonable expectation that it occurs in the equivalent limestone of Ontario. The holotype of O. eriensis is a dentary lacking parasymphysial teeth, so a direct comparison with teeth of O. sigmoides from coeval carbonate strata is not possible. However, one small, isolated tooth, referred to as Onychodus cf. eriensis [6] (fig. 3B), is indistinguishable from some small teeth of O. sigmoides (compare Figure 1B, herein). Several jaw bones with teeth, including dentaries, are among Newberry’s original examined specimens of O. sigmoides from the Delaware Limestone fish beds of Ohio. A study of these indicates that the holotype of O. eriensis, with its curved and expanded anterior dentary, and its flattened anterior surface of the mandibular rami, is characteristic of the O. sigmoides specimens studied by Newberry, as well other specimens, including a dentary illustrated by Hansen [62] (fig. 21-10.4). Whether there are other, more diagnostic differences is uncertain, and testing must await further work. Unless consistent differences between the characteristic O. sigmoides and specimens described as O. eriensis can be demonstrated, O. eriensis should be treated as a junior subjective synonym.

Onychodus sigmoides can be distinguished from all other known species in the genus on the basis of tooth morphology. Onychodus sigmoides has a variable number of teeth on the parasymphysial plate, usually five or six, but possibly as few as four in large individuals, and rarely as many as seven. The parasymphysial teeth range in morphology from slender in small (<20 mm) to mid-sized specimens to large and robust in the largest specimens (typically >35 mm). Larger specimens have more robust teeth and tend to have fewer teeth, which suggests that the number of parasymphysial teeth decreased through ontogeny to accommodate the increased width of robust parasymphysial teeth in later life. The parasymphysial teeth are simple in outline and procurved to sigmoidal. Teeth having distal recurvature exhibit a sigmoidal outline. Teeth can range in shape from nearly fully procurved with only a slight recurvature at the tip of the tooth on the proximal part of the parasymphysial plate to significantly sigmoidal with a distinct recurvature along 10 to 35% of the tooth length distally. Although all species of Onychodus show some recurvature at the tips of the parasymphysial teeth, the amount of recurvature expressed in the most sigmoidal teeth of O. sigmoides is greater than in any other species. In addition, O. sigmoides appears to be unique among species in the genus in the amount of variation in tooth shape along the parasymphysial. Some species, such as O. ortoni and O. jandemarri, have little variation in shape along the parasymphysial [27], whereas it can be inferred from isolated parasymphysial teeth that O. yassensis Lindley, 2002 had teeth of variable shape [63], but that variation is less than in O. sigmoides. The largest dentaries of O. sigmoides have up to 50 short, conical to slightly curved teeth, and this is the largest number known in the genus.

Onychodus ortoni Newberry, 1889

Figure 3 and Figure 4B

• 1889 Onychodus ortoni Newberry, pp. 71–72, pl. 19, fig. 1, 1a (original description).
• 1889 Onychodus hopkinsi Newberry–Newberry, pp. 99–100.
• 1889 Onychodus hopkinsi Newberry–Miller, p. 603 (in part).
• 1902 Onychodus ortoni–Hay, p. 363.
• 1907 Onychodus hopkinsi Newberry–Eastman, p. 169 (in part), pl. 3, fig. 3; pl. 9, figs. 1, 2.
• 2024 Onychodus ortoni–Babcock, p. e2308621-9, fig. 1C.
• Holotype: Parasymphysial whorl, OSU 14154 (by monotypy); from the Huron Member of the Ohio Shale (Upper Devonian, Famennian), Perry Township (probably Flint Run [37]), Franklin County, Ohio, USA.
• Diagnosis: Onychodus having five to six slender parasymphysial teeth, reaching about 25 mm in length; they are simple at the tips and almost completely procurved except for a slight recurvature in the distal 1 to 2 mm. They are uniformly similar in morphology along the length of the parasymphysial plate.
• Remarks: Onychodus ortoni, which was previously known from only the holotype parasymphysial whorl (Figure 3B,C), was discussed recently [37]. The holotype is moderately large and shows relatively long but slender parasymphysial teeth; they are comparatively more slender than most of those in O. sigmoides at similar length, and they are much more uniform in shape along the parasymphysial. The arcuate, slender shape of the parasymphysial teeth, with a subtle deflection at the distal tips, is a useful characteristic for distinguishing teeth of O. ortoni from those of O. sigmoides. The differences between the teeth of the two species are most pronounced in moderately large specimens. A second parasymphysial whorl of O. ortoni (Figure 3D), apparently a topotype, is smaller than the holotype, but has teeth similar in shape to those in the holotype.

Both O. ortoni and O. jandemarrai [27] have parasymphysial teeth that vary little along the length of the parasymphysial plate. O. ortoni can be distinguished, however, by its more strongly procurved teeth, which are recurved subtly at the tips.

A moderately large, studied parasymphysial whorl from the “Chemung” of New York (Figure 3A) has teeth that differ from all specimens studied from the Delaware Limestone of Ohio. The specimen has arcuate (procurved) teeth, with subtle recurvature at the distal tips. The teeth are slender, and the tips are sharp. The specimen closely resembles both known specimens of O. ortoni from the Huron Shale Member of the Ohio Shale in Ohio (Figure 3C,D).

6.2. Biostratigraphy and Lithofacies Distribution

Based on an examination of the material from across the Appalachian Basin, O. sigmoides and O. ortoni have distinct stratigraphic ranges: O. sigmoides ranges from the Columbus Limestone/Onondaga Limestone and equivalents (Emsian?–Eifelian) through the Delaware Limestone and equivalents (Eifelian) into the Hamilton Group and equivalents (Givetian), whereas O. ortoni ranges from the West Falls Group and equivalents (Frasnian) into the Huron Member of the Ohio Shale and equivalents (Famennian) in the Appalachian Basin. Specimens that were described or intended to be referred to as O. hopkinsi seem to be O. sigmoides if they derive from the Delaware Limestone (Eifelian), and O. ortoni if they derive from the “Chemung” (Frasnian) or Ohio Shale (Famennian). The illustrated specimens, referred by Eastman [14] to O. hopkinsi from the contact of the Seneca Member of the Onondaga Limestone and the Union Springs Formation in New York, all seem referable to O. sigmoides. All specimens that I have studied from the Seneca–Union Springs contact likewise appear to be referable to O. sigmoides. However, considerable morphological variation is present in the material from this horizon [15], much of which is relatively small in size. The age of this lag deposit is close to that of the lower Delaware Limestone of Ohio, and the type horizon of O. sigmoides in the Delaware fish beds Lagerstätte.

Onychodus sigmoides, as known at present, has a lithofacies distribution primarily in carbonate strata and mixed carbonate-fine siliciclastics, all representing relatively shallow shelf seas of the Appalachian Basin area. Phosphatic remains are common in time-averaged beds, including the Delaware fish beds [7,10,11,12,13,33], bone beds [47,50], and other lag deposits [15]. Onychodus eriensis is treated here as a junior synonym of O. sigmoides. The specimens described as O. eriensis are known from only carbonate lithofacies in the Dundee Formation of Ontario, where they occur in or are close to the equivalent stratigraphic position in the Delaware Limestone.

Onychodus ortoni has a known lithofacies distribution in black shale (the Huron Member of the Ohio Shale) of Ohio, presumably representing deeper portions of the Appalachian foreland basin; and in fine-grained siliciclastics (siltstone beds) of the West Falls Group of New York, representing shallow, nearshore marine sedimentation of the westwardly prograding Catskill Delta complex. As such, it seems to have had a wider environmental tolerance than O. sigmoides.

7. Summary

A restudy of J.S. Newberry’s original type and examined specimens of the sarcopterygian fish Onychodus from Ohio and New York permits clarification of the major uncertainties and ambiguities that have surrounded the species reported from the Devonian of the Appalachian Basin. The following points answer the questions outlined in the Introduction (Section 1):

• The type specimens of both O. sigmoides and O. hopkinsi were collected from a limestone quarry in Delaware, Delaware County, Ohio. This quarry is inferred to be the former Campbell’s Quarry, which is now water-filled and on the site of Blue Limestone Park in Delaware.
• The lectotype and paralectotype parasymphysial teeth of O. sigmoides, as well as the holotype parasymphysial whorl of O. hopkinsi, were collected from the “fish beds” of the Delaware Limestone [33].
• The age of the fish bed interval in the Delaware Limestone is Middle Devonian (Eifelian); it has a numerical age estimate of approximately 390 Ma [41,44,45].
• All studied Onychodus remains from the Delaware Limestone at Delaware, Ohio, are referable to a single species, O. sigmoides, which shows two end-member morphologies of parasymphysial teeth.
• The original name-bearing specimens of O. sigmoides are the two parasymphysial teeth illustrated by Newberry [7], from which the larger tooth (Figure 1A) is selected as the lectotype. The holotype of O. hopkinsi is the parasymphyial whorl described by Newberry [7] and illustrated in 1862 [12] (Figure 2A,B). A parasymphysial whorl from the West Falls Group (“Chemung”; Upper Devonian, Frasnian) of Franklin, Delaware County, New York, examined by Newberry and evidently referred to in text [9,10,13], is now referred to as O. ortoni.