The material under consideration was discovered by Mr. Bronislav Novosad at the “Kelč-Strážné” locality in 2008. The fossil was partially prepared using small scalpels or needles. Examination was made of the fossil specimen, as well as plaster casts or latex peels (in some cases, coated with ammonium chloride for better visibility of features). Photographs were taken using a Canon EOS 1000D camera attached to a Leica MZ6 stereomicroscope. The drawings were prepared using a camera lucida drawing tube and measurements of the specimens were based on the photos. The comparison with fossil and recent species of Moridae is based on the literature.

The fossil is housed in the collection of the National Museum in Prague (NMP).

Anatomical abbreviations: boc: basioccipital; br: branchiostegal rays; cl: cleithrum; co: circumorbitals; den: dentary; ect: ectopterygoid; end: endopterygoid; epio: epiotic; epn: epineural; exo: exoccipital; fr: frontal; hy: hyomandibula; le: lateral ethmoid; lpr: lower process of hyomandibula; mes: mesethmoid; mtp: metapterygoid; mx: maxilla; na: nasal; op: opercle; P: pectoral fin; pa: palatine; par: parietal; pcl: postcleithrum; pmx: premaxilla; pop: preopercle; prar: articular process of premaxilla; pras: ascending process of premaxilla; prop: opercular process of hyomanibula; prpm: postmaxillary process of premaxilla; pst: posttemporal; pter: pterotic; q: quadrate; r: ribs; rad: radials; scl: supracleithrum; soc: supraoccipital; sph: sphenotic; uh: urohyal; V: pelvic fin; vo: vomer; vert: vertebrae.

The Kelč locality (Moravia, Western Carpathians; Subsilesian Unit) is historically well known for providing a rich fossil record (e.g., Kalabis, 1975, Knobloch, 1969 and Prokop et al., 2007). Fossils were originally collected at “Kelč-Zámek” (Kalabis, 1975), but this area is now built up and as such is no longer accessible. All recently collected materials referred to the Kelč locality originate from the “Kelč-Strážné” area, situated about 1.5 km northwest of the city centre. Fragments of Dynów marlstones (one of the litho-stratigraphic members of the Menilitic Formation; for details regarding stratigraphy, see Bubík et al., 2016) are found in ploughed fields or meadows, and although no larger slabs of the sediments are available, the locality has provided relatively plentiful material (ongoing research).

Subdivision Teleostei Müller, 1845 sensu Arratia, 1999

Order Gadiformes Goodrich, 1909 sensu Endo, 2002

Family Moridae Goode and Bean, 1896

Genus et species unidentified

(Fig. 1, Fig. 2, Fig. 3 and Fig. 4a)

Referred specimens: NMP Pv 10054a and 10054b (part and counterpart).

Description: The fossil preserves the anterior part of the body and the dorso-ventrally flattened skull of a small fish. The length of the preserved portion is approximately 38 mm (measured on specimen NMP Pv 10054a), with an estimated standard length of not more than 80 mm.

The neurocranium is preserved in dorsal view. The frontal is narrow in the interorbital area, slightly wider in the ethmoid region and considerably wider in the postorbital region, making the lateral margins of the frontal broadly concave. Some parts of the supraorbital margin have a lobular, segmented appearance. The anterior-most tips of the frontal extend antero-laterally. The skull roof bears an X-shaped skull ridge (Fig. 3 and Fig. 4). In the midline, anterior to the frontals, the remains of the mesethmoid are recognisable; no traces of the vomer are identifiable. The lateral ethmoids are well developed, expanded laterally and have concave anterior margins (the lateral extremities reach half of the orbit diameter). The small, oval sphenotic is identifiable in the antero-lateral portion of the postorbital region. The supraoccipital with a well-developed crest is identifiable. The supraoccipital separates the left and right parietals and the posteriorly located epiotics, while the parietals and the epiotics are in contact with each other. No remains of the exoccipitals, nor a basioccipital, are recognizable. The pterotic is large (medially it contacts the epiotic, parietal and frontal, and anteriorly it contacts the sphenotic), bearing two horn-like processes (the ventro-lateral and posterior; white and black arrows respectively in Fig. 3A, B). The remains of the oval-shaped nasals are recognisable between the mesethmoid and the lateral ethmoids. Specimen NMP Pv 10054b shows well-preserved remains of four small ossifications in the orbital region, tentatively interpreted as circumorbitals. The hyomandibula has a single articulation with the neurocranium (the articular facet of the neurocranium is located just anterior to the ventro-lateral process of the pterotic). The hyomandibula bears well-developed opercular and (possibly also) anterior processes (Fig. 3C). The metapterygoid bears two dorsally oriented processes, the posterior one probably serving for articulation with the main hyomandibular body, while the anterior process serves for articulation with the presumed lower process of the hyomandibula. The details of the quadrato-pterygo-palatal complex are not clear, but it seems to be relatively strongly built and generally in the shape of an elongated, posteriorly expanded strap. The palatine seems to be well developed, with a long, antero-laterally oriented articular process. The lower jaw joint is located at the level of the posterior margin of the orbit. The lower jaw is insufficiently preserved, but seems to be relatively straight. The premaxilla is slightly bent, with broadly separated and well-developed articular, ascending and postmaxillary processes, and a postmaxillary notch (Fig. 3E). The dentary and premaxillae bear numerous small teeth (the latter all along the ventro-lateral margins; the distribution of the teeth on the dentary is not clear); some of the teeth seem to be somewhat larger than others, but these are randomly placed. The maxillae (Fig. 3E) show a well-developed articular groove within the articular head; the posterior part of the maxilla is spear-shaped. The precise shape of the preopercle is not clearly recognisable, but preserved fragments suggest it was a broad sickle-shaped bone. The opercle is more or less triangular, with a concave postero-ventral margin. The urohyal shows a typical gadoid shape (Fig. 3D). Fragments of branchiostegal rays and remains of the branchial arches (and associated toothed patches) are recognisable, but their detailed morphology and numbers are not clear.

The pectoral girdle is represented by remains of a relatively robust crescent-shaped cleithrum, articulating dorsally with the stick-like supracleithrum. The whole pectoral girdle articulates with the neurocranium via the V-shaped posttemporal. The postcleithrum is straight, needle-shaped, with a slightly enlarged proximal part. Remains of the scapula and coracoid are not recognizable. There are recognisable fragments of three proximal radials, but this area of the fossil is partly covered with sediment and their actual number may thus be higher. The pectoral fin has a more or less rounded margin and is composed of about 15 rays. The pelvic girdle is not preserved, but the preserved remains of the pelvic fin rays suggest the fins were relatively short, and composed of about five rays each.

The axial skeleton is far from complete and preserves only the remains of eight vertebral centra, followed by remains of four parapophyses (the parapophyses are easily recognisable from the seventh preserved vertebra). The first preserved vertebra is expected to be in fact the second one of the column, while the first vertebra is coalesced with the occipital part of the skull (as described in morids, e.g., in Lotella cf. tosaensis by Prokofiev, 2008 or in Eophycis jamnensis by Přikryl, 2015). The neural arches of the anteriorly preserved vertebrae are well developed, broad, and relatively robust. From the second preserved vertebra, the epipleural bones are attached to the ventro-lateral margins of the centra (and more posteriorly on the parapophyses). The first six epipleurals are straight, while those more posteriorly located are slightly bent. Ribs are clearly recognisable, attaching to the epipleurals or (in two penultimate cases) the parapophyses. The last preserved parapophysis bears an epipleural, but no rib. Neither scales nor otoliths are recognisable. Soft-tissue remains are partly preserved as a dark film in the eye and abdominal areas.

The preservation of the specimen in dorsal view and its incomplete condition impedes adequate interpretation and comparisons. Nevertheless, the fossil under consideration shows one feature considered to be diagnostic for the Paracanthopterygii and Anacanthini, namely the absence of the parapophyses on the anteriorly located centra (Patterson and Rosen, 1989). The inclusion of the fossil in the Gadiformes is supported by its overall appearance, the presence of a single articulation between the hyomandibula and neurocranium (Wiley and Johnson, 2010), and by the absence of the epipleural bones on the first and second centra (Ford, 1937, Murray and Wilson, 1999 and Patterson and Rosen, 1989).

The familial status of the specimen differs from that of any other gadiform described earlier from the region (family Merlucciidae with the genera Palaeogadus from the Oligocene and Merluccius from the early Miocene; Gregorová, 2013, Gregorová and Požár, 2003 and Jaroš, 1937), mainly due to the generally different architecture of the skull (compare Fig. 4A, D and E): while the merlucciids are characterised (among others) by a V-shaped neurocranial crest opening anteriorly (Cohen, 1984, Inada, 1981 and Lloris et al., 2005), the fossil studied here clearly shows an X-shaped neurocranial crest (Fig. 4A). Although the inclusion of the specimen in the family Moridae is not corroborated directly by any of the diagnostic characters mentioned above, such a classification is supported by the proportions of the skull, with the orbito-rostral portion being larger than or of a similar size to the postorbital portion (Daniltshenko, 1960) and a generally similar morphology of the dorsicranium. Unfortunately, no detailed comparative osteology of the family Moridae has been published yet, but some of the morid crania figured by Svetovidov (1948; namely Uraleptus Costa, 1846 (= Gadella Lowe, 1843), Lotella Kaup, 1858, and Physiculus Kaup, 1858) share several important characters with this fossil, namely: the lateral ethmoids expanding laterally at the level of half the height of the orbit; the sphenotics being located more laterally than the lateral-most margins of the lateral ethmoids; the sphenotics being significantly smaller than the pterotics; the pterotics being expanded anteriorly; and an X-shaped neurocranial crest present.

The pterotic of the genus Lotella in dorsal view (as figured by Svetovidov, 1948) bears two processes, one ventro-lateral and one posterior, similar to the fossil under consideration (white and black arrows, respectively, in Fig. 3 and Fig. 4). A somewhat similar condition was figured also for the genus Eeyorius, where both processes have the form of slight ridges (see Paulin, 1986; Fig. 4B).

The fossil is therefore most probably a member of the “Pseudophycis group” (including the genera Pseudophycis Günther, 1862, Lotella Kaup, 1858 and Eeyorius Paulin, 1986), which is characterised by “otoliths, with the ostium approximately equal to the cauda and a crista superior more than three-fourths as long as the crista inferior” (Paulin, 1989). Unfortunately, otoliths are not preserved, and such definitive confirmation of this conclusion is not possible.

The fossil is reminiscent of the genus Lotella, based on the features mentioned above and the shape of the postcleithrum, which is simple and straight (see Endo, 2002: fig. 38). The genus Lotella consists of four valid extant species (Cohen et al., 1990), two questionable fossil species based on articulated skeletons (Daniltshenko, 1953 and Daniltshenko, 1960, but see also revision by Świdnicki et al., 1990) and relatively few otolith-based occurrences (e.g., Schwarzhans et al., 2012 and Schwarzhans et al., 2017), but a closer comparison with any of these taxa is not possible due to the preservation of the fossil.

The number of fin rays in the paired fins of the fossil, compared with data published for the “Pseudophycis group”, suggests a strong similarity with the genus Pseudophycis (see Table 1).

The fossil studied here cannot be included in the genus Eophycis, mainly due to the different morphology of the hyomandibula, the shape of the lower jaw and the pterotic, and the number of fin rays in the paired fins (Přikryl, 2015).

Some characters of the fossil are unknown (such as the total number of vertebrae, number of rays in the unpaired fins and morphology of their skeletal supports, otoliths, and soft tissue features) or are difficult to interpret correctly (such as the presumed circumorbitals, which seem to be relatively unusual; compare Fig. 2 and e.g., Endo, 2002: fig. 29). Nevertheless, these problems are related to the fragmentary state of the specimen and poor preservation, as well as to our limited understanding of the osteology of extant morid species. The preservation of the fossil does not allow a full comparison and proper description of a new taxon; such a description must wait until new, better preserved specimens become available. However, it is possible to conclude that the Central Paratethys was inhabited by a morid fish morphologically closely related to Lotella during the Rupelian.