A large Late Miocene cetotheriid (Cetacea, Mysticeti) from the Netherlands clarifies the status of Tranatocetidae

Collection, preparation, body size and phylogenetic analysis

The two specimens described here were trawled from the bottom of the Western Scheldt (the Netherlands) during NMR expeditions 2014-3 and 2015-1 (Fig. 1). Both fossils were embedded in a matrix of hard glauconitic sandstone, and prepared mechanically. For the description, morphological terminology follows Mead & Fordyce (2009), unless indicated. Photographs of the specimens were digitally stacked in Photoshop CS6.

Total body length was estimated based on bizygomatic width, using the stem balaenopteroid equation of Pyenson & Sponberg (2011), and the general mysticete equation of Lambert et al. (2010). To establish evolutionary relationships, we coded our new material, as well as Tranatocetus argillarius and the recently described Middle Miocene cetotheriid Ciuciulea davidi (Gol’din & Steeman, 2015; Gol’din, 2018), into a slightly modified version of the phylogenetic matrix of Fordyce & Marx (2018). The analysis was run in MrBayes 3.2.6, on the Cyberinfrastructure for Phylogenetic Research (CIPRES) Science Gateway (Miller, Pfeiffer & Schwartz, 2010), using the same settings as in Fordyce & Marx (2018). The full cladistic matrix is available as online supplementary material.

Age determination

Matrix samples from the more complete specimen (NMR 9991-16680) were prepared at Palynological Laboratory Services (PLS, UK) using standard sample processing procedures, which involve HCl and HF treatment, heavy liquid separation, and sieving over a 15 µm mesh sieve. The organic residue was mounted with glycerine-gelatine on microscopic slides. Two microscopic slides were made: one carrying non-oxidized kerogen, and one on which the organic residue was slightly oxidized with HNO₃ to concentrate the palynomorphs and reduce ‘Structureless Organic Matter’.

Palynological analysis was carried out at the Geological Survey of the Netherlands (TNO). We counted the first 200 sporomorphs (pollen and spores) and dinoflagellate cysts, and thereafter scanned for any rarer dinocyst species. Major miscellaneous categories (e.g., marine acritarchs, test linings of foraminifers, and the brackish alga Botryococcus) were calculated separately. Age interpretations were based on the first and last occurrences of dinoflagellate cysts, using the dinozones of Munsterman & Brinkhuis (2004) recalibrated to Ogg, Ogg & Gradstein (2016). Dinoflagellate cyst taxonomy follows the ‘Lentin and Williams index’ (Williams, Fensome & MacRae, 2017).

Nomenclatural acts

The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: urn:lsid:zoobank.org:pub:D39ACC32-687F-4C95-9CD8-A2B17B2DBAFC. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.

Systematic palaeontology

Type species. Tranatocetus argillarius (Roth, 1978)

Emended diagnosis. Large cetotheriid sharing with other members of the family the presence of an enlarged compound posterior process of the tympanoperiotic [hereafter, compound posterior process], an enlarged paroccipital concavity extending on to the compound posterior process, elongate, medially convergent ascending processes of the maxillae, a supraoccipital shield whose tip does not extend beyond the apex of the zygomatic process of the squamosal, and a posteriorly projected angular process of the mandible bearing a well-defined fossa for the medial pterygoid muscle. Further shares with all cetotheriids except Cephalotropis and Joumocetus the posterior telescoping of the ascending process of the maxilla up to, or beyond, the anterior border of the parietal. Differs from all described cetotheriids in its larger size, in having a flattened platform located inside the posterodorsal corner of the mandibular fossa, and in having a mandibular condyle that does not markedly rise above the level of the mandibular neck. Further differs from all cetotheriids except Herentalia in having a notably elongate anterior process of the periotic; from all cetotheriids except Metopocetus in having a reduced lateral tuberosity; from all cetotheriids except Herentalia, Metopocetus and Piscobalaena in having a sharp cranial rim surrounding the proximal opening of the facial canal; from Brandtocetus, Cetotherium, Kurdalagonus and Mithridatocetus in having a larger posteroventral flange of the compound posterior process that completely floors the facial sulcus, a tympanic bulla that is less squared and narrower anteriorly than posteriorly (in ventral view), a more elongate ascending process of the maxilla, and a more gracile zygomatic process of the squamosal; from Cephalotropis, Ciuciulea and Joumocetus in lacking a long exposure of the parietal on the vertex; from Cephalotropis in having a proportionally larger bulla, and a better-developed posteroventral flange of the compound posterior process; from Herpetocetus and Piscobalaena in having a squamosal cleft; from Piscobalaena in having a sharper vomerine crest; from Herpetocetus in lacking a postparietal foramen; from ‘Cetotherium’ megalophysum in having a distally larger compound posterior process with a better-developed posteroventral flange, and a more concave supraoccipital shield; from Metopocetus in having a more elongate ascending process of the maxilla, a narrower posterior portion of the nasal, and a smaller tympanohyal; and from Herentalia in having a more pointed apex of the supraoccipital shield and in lacking a well-developed external occipital crest.

LSID. urn:lsid:zoobank.org:act:499F1C5C-3C3F-48A9-AD97-AF19F99DE886

Holotype. NMR9991-16680, partial cranium comprising the braincase and ear bones, posterior portions of both mandibles, a fragmentary atlas, and two thoracic vertebrae.

Paratype. NMR9991-16681, basicranium, right periotic, atlas, and seventh cervical vertebra.

Locality and horizon. Both fossils were recovered from the Breda Formation, exposed at site 6d (N51°21′56.9″, E3°54′25.1″) of Post & Reumer (2016). Assemblages from the associated matrix are relatively rich in marine dinoflagellate cysts, but include few (ca 10%) sporomorphs (Table S1). The latter mostly consist of bisaccate pollen (71%), which are relatively buoyant and, along with the abundance of dinocysts, indicate a distal position from the coast. The most abundant dinocyst genus is Spiniferites (39% of the total dinocyst sum), which preferentially occurs in open marine conditions. However, the temperate–tropical, inner neritic Barssidinium graminosum and the coastal Lingulodinium machaerophorum are also well-represented (24% and 6%, respectively), suggesting overall neritic conditions.

Enneadocysta pectiniformis and Glaphyrocysta spp. are reworked from the Oligocene or older intervals. Among the age-diagnostic taxa, Hystrichosphaeropsis obscura last occurs in Zone SNSM14 (ca. 7.5 Ma), and defines both the DN9 Zone of de Verteuil & Norris (1996) and the Hystrichosphaeropsis obscura Zone of Dybkjær & Piasecki (2010). The presence of Selenopemphix armageddonensis confirms a date no older than late Tortonian. The first occurrence of this species has variously been placed at either 7.5 Ma (Zone DN10; De Verteuil & Norris, 1996; Dybkjær & Piasecki, 2010), or at 9 Ma in equatorial areas (Williams et al., 2004). In Belgium, it has been recorded from the Kasterlee Formation (Louwye & de Schepper, 2010), whereas in the Netherlands it extends to the top of Zone SNSM13, which dates to approximately 8.1 Ma (well Groote Heide). Together, these observations constrain the current assemblage to Zone SNSM14, Late Miocene, ca. 8.1–7.5 Ma (Munsterman & Brinkhuis, 2004, recalibrated to Ogg, Ogg & Gradstein, 2016).

Diagnosis. Shares with Tranatocetus argillarius its large overall size, slender ascending processes of the maxillae that are situated centrally on a triangular platform formed by the frontals, a narrow but continuous exposure of the nasals between the ascending processes of the maxillae, the lack of an external occipital crest, a bulbous exoccipital, an elongate anterior process of the periotic lacking a lateral tuberosity, a large paroccipital concavity, a large mandibular fossa housing a flattened platform in its posterodorsal corner, a deep subcondylar furrow, and a mandibular condyle that does not markedly rise above the level of the mandibular neck. Differs from T. argillarius in having a more anterolaterally directed base of the supraorbital process of the frontal, a more robust zygomatic process of the squamosal, relatively larger occipital condyles projecting posteriorly beyond the posterior apex of the nuchal crest, a vomerine crest extending posteriorly far beyond the level of the subtemporal crest, and a sharper, dorsally convex ventral border of the mandibular foramen.

Etymology. After the Latin name of the North Sea, Mare Germanicum, which Tranatocetus once inhabited.

Identification of the new material as Tranatocetus

Tranatocetus maregermanicum closely resembles T. argillarius in its (i) overall size, (ii) slender ascending process of the maxilla, (iii) lack of an external occipital crest, (iv) elongate anterior process of the periotic, (v) lack of a lateral tuberosity on the periotic, (vi) low mandibular condyle, (vii) large mandibular fossa, (viii) presence of a posterodorsal platform inside the mandibular fossa, and (ix) deep subcondylar furrow (Fig. 9).

Nevertheless, the original description of T. argillarius also lists several features that appear to differentiate the two species. Of these, the most obvious include a smaller angular process of the mandible which seemingly does not extend beyond the level of the mandibular condyle (Gol’din & Steeman, 2015, p. 15–12); an even larger mandibular fossa (Gol’din & Steeman, 2015, p. 7); a smaller distal exposure of the compound posterior process (Gol’din & Steeman, 2015, p. 4); and a somewhat rounded, rather than straight, lateral border of the supraoccipital (Gol’din & Steeman, 2015, Fig. 1).

Our re-examination of T. argillarius suggests that all of these differences can be explained by the poor state of preservation of the holotype (GMUC VP2319). Thus, the angular process of the latter is not preserved, and has been entirely reconstructed in resin, leaving its shape and size in doubt (Roth, 1978) (Fig. 10). A morphology similar to that of T. maregermanicum therefore cannot be excluded, and perhaps might even be indicted by the similarly pronounced subcondylar furrow in both species.

The size and shape of the mandibular fossa are similarly problematic, as its ventral portion in T. argillarius has broken off, and no longer makes direct contact with the remainder of the mandible. During reconstruction, it was fixed into its inferred position with resin, but likely somewhat out of place, and at the wrong angle (Fig. 10). In anterior view, the outer wall of the mandibular fossa curves inwards and becomes markedly thinner towards its ventral border. By contrast, the now juxtaposed ventral portion of the mandible (including the ventral border of the mandibular fossa) is notably less concave and relatively thick, implying that it is not in its original position. Extrapolating the curvature of the outer wall suggests that the cross section of the ramus, and therefore also the mandibular fossa, would originally have been smaller, and thus more similar to that of T. maregermanicum.

The compound posterior process of T. argillarius is extremely poorly preserved on both sides of the skull (Fig. 10). Despite repeated attempts by two of the authors (FGM and KP), we were unable to trace its outline, thus invalidating it as a diagnostic feature. We note, however, that the space between the external acoustic meatus and the exoccipital is large, and thus consistent with a broadly exposed compound posterior process as seen in T. maregermanicum. A large compound posterior process would furthermore match the sizeable paroccipital concavity.

Finally, the supraoccipital is highly fragmentary, which makes its shape difficult to assess. The tip is fixed in place by a large amount of resin, giving rise to an artificially rounded left lateral outline in dorsal view. The right lateral border is comparatively straight, although the supraoccipital shield as a whole still seems somewhat broader and blunter than in T. maregermanicum. On the whole, the detailed morphology of this feature likely differs between T. argillarius and T. maregermanicum, but not as much as the reconstructed skull of the former might suggest.

Based on these observations, we conclude that the features distinguishing the two species of Tranatocetus are relatively mild, with the most pronounced differences being attributable to artefacts of preservation. In keeping with the results of our phylogenetic analysis (see below), we therefore reaffirm their placement in the same genus.

Phylogeny and status of Tranatocetidae

For more than a decade, there has been broad agreement on the basic concept of Cetotheriidae (Bouetel & de Muizon, 2006; El Adli, Deméré & Boessenecker, 2014; Gol’din & Startsev, 2017; Marx & Fordyce, 2015; Steeman, 2007; Whitmore Jr & Barnes, 2008). Nevertheless, the scope of the family has been thrown in doubt by the inclusion of the pygmy right whale, Caperea marginata (Fordyce & Marx, 2013; Marx & Fordyce, 2016; Park et al., 2017), and the proposed grouping of several species usually regarded as cetotheriids into the separate family Tranatocetidae (Gol’din & Steeman, 2015). The latter is thought to be more closely related to balaenopterids than to cetotheriids, and comprises the eponymous Tranatocetus, as well as Mesocetus longirostris, Mixocetus elysius, ‘Aulocetus latus’, ‘Cetotherium’ megalophysum, and ‘Metopocetus’ vandelli (Gol’din & Steeman, 2015). Of these, the last three frequently cluster in phylogenetic analyses, and may represent the same genus or even species (El Adli, Deméré & Boessenecker, 2014; Gol’din, 2018; Marx, Bosselaers & Louwye, 2016).

Our phylogenetic analysis contradicts the status of Tranatocetidae as a separate family by recovering Tranatocetus deeply nested within Cetotheriidae, as sister to Metopocetus (Fig. 11). The same applies to other presumed tranatocetids, including ‘Aulocetus’ latus, ‘Cetotherium’ megalophysum, and ‘Metopocetus’ vandelli (Gol’din & Steeman, 2015). The monophyly of Cetotheriidae is primarily supported by the presence of a posteroventral flange on the compound posterior process (char. 184), and the parabolic outline of the postglenoid process (char. 118, in posterior view). Additional characters shared by all cetotheriids except Tiucetus and Joumocetus include ascending processes of the maxillae that directly contact the nasals (char. 67, reversed in cetotheriines), parietals that extend no further forward than the level of the postorbital process (char. 81), a distal surface of the compound posterior process that is firmly integrated into the lateral skull wall (char. 188), a squared anterior border of the tympanic bulla (char. 191), deep transverse creases on the dorsal surface of the involucrum (char. 207), and a mandibular body that increases in height towards the symphyseal area (char. 222, in lateral view).

Several features were previously noted as distinguishing tranatocetids from cetotheriids, including a smaller distal exposure of the compound posterior process, a narrower anterior portion of the tympanic bulla, a small angular process of the mandible, a low mandibular condyle, and a shallow glenoid fossa (Gol’din & Steeman, 2015).

The exposure of the compound posterior process is variable among cetotheriids, with the process being broadly exposed in herpetocetines, cetotheriines (sensu Gol’din & Startsev, 2017), Herentalia, Metopocetus and Piscobalaena, but less so in Joumocetus and Tiucetus. This range presumably is the result of a trend, with the basalmost taxa also having the smallest exposures (Kimura & Hasegawa, 2010; Marx, Lambert & Muizon, 2017). Tranatocetids fit different parts of this spectrum, with the exposed surface of Tranatocetus being comparable to that of Herentalia and Metopocetus, and far larger than in ‘Aulocetus’ latus, ‘Cetotherium’ megalophysum and ‘Metopocetus’ vandelli. All tranatocetids furthermore share with cetotheriids a common structure of the compound posterior process, with the latter being expanded distally, and bearing a posteroventral flange which partially floors the facial sulcus (Figs. 7 and 12) (Marx, Bosselaers & Louwye, 2016; Marx & Fordyce, 2016). Tranatocetids therefore do not systematically differ in the morphology of their compound posterior process, but form part of morphological continuum encompassing all of Cetotheriidae.

Like the compound posterior process, the shape of the tympanic bulla is variable among cetotheriids. In ventral view, the anterior portion of the bulla is equal to or wider than the posterior half in Herpetocetus and cetotheriines, slightly narrower in Ciuciulea, and notably narrower in Metopocetus and Piscobalaena (Fig. 12). In most cetotheriids—in particular, cetotheriines and Piscobalaena—the anterior border of the tympanic bulla is furthermore notably squared. Tranatocetids generally conform to the narrow morphotype, including the squared anterior border, and thus tend to resemble Piscobalaena. (Fig. 12). We agree that narrowing of the anterior bulla may be a derived state setting apart certain species in the broader context of Cetotheriidae. Nevertheless, as shown by the striking resemblance of ‘C.’ megalophysum and Piscobalaena (Fig. 12), there is no clear division between this morphology and that of several undoubted cetotheriids.

The angular process of the mandible tends to be enlarged in Cetotheriidae, either dorsoventrally as in cetotheriines (Gol’din, Startsev & Krakhmalnaya, 2014; Gol’din, 2018), or anteroposteriorly as in Piscobalaena and herpetocetines (Bouetel & de Muizon, 2006; El Adli, Deméré & Boessenecker, 2014). In the latter two, the process notably projects beyond the level of the mandibular condyle, and bears a well-developed fossa for the medial pterygoid muscle. Our new observations show that Tranatocetus precisely fits this elongate morphotype (Fig. 8), nesting it deep within Cetotheriidae. The mandibular morphology of other ‘tranatocetids’ is poorly known, with no lower jaws having been described for ‘A.’ latus, ‘C.’ megalophysum and ‘M.’ vandelli.

A comparatively small angular process, as previously suggested for tranatocetids (Gol’din & Steeman, 2015), appears to be the plesiomorphic state, based on its occurrence in both balaenopterids and a variety of Miocene non-cetotheriids (Kellogg, 1934; Kellogg, 1968b; Steeman, 2009; Tanaka, Ando & Sawamura, 2018a). In light of this observation, we question its usefulness as a distinguishing characteristic of tranatocetids, and predict that basal cetotheriids, such as Joumocetus and Tiucetus, may ultimately be revealed to have a markedly smaller angular process than other members of the family.

Like a small angular process, a low mandibular condyle appears to be a plesiomorphic feature characterising both balaenopterids and a broad range of Miocene non-cetotheriid mysticetes (Kellogg, 1934; Kellogg, 1968a; Kellogg, 1968b; Kimura, 2002; Tanaka, Ando & Sawamura, 2018a). By contrast, the condyle of cetotheriids tends to be somewhat elevated above the mandibular neck. The position and orientation of the condyle in turn correlates with that of the glenoid fossa of the squamosal, with the latter reportedly being relatively shallow in cetotheriids (Gol’din, Startsev & Krakhmalnaya, 2014; Gol’din & Steeman, 2015).

Tranatocetus has a posteriorly oriented, non-elevated condyle, and—in this regard—shows a relatively primitive morphology of the craniomandibular joint (Figs. 8 and 9). This anatomy could plausibly be plesiomorphic but, considering the otherwise clearly cetotheriid shape of the ramus, might also represent a secondary feature, perhaps associated with large body size. The anatomy of ‘A.’ latus, ‘C.’ megalophysum and ‘M.’ vandelli remains unknown. As with a small angular process, we predict that a low mandibular condyle also primitively occurred in some basal cetotheriids, and hence is of limited value in distinguishing the latter from tranatocetids.

Overall, the evidence supporting a separation of Cetotheriidae and Tranatocetidae is thus relatively weak. By contrast, our new observations on Tranatocetus reveal a marked resemblance of this genus with several undoubted cetotheriids, borne out by the results of our phylogenetic analysis. These results cast doubt on the status of Tranatocetidae as a separate clade, and instead imply the existence of a single, extended family Cetotheriidae, including Tiucetus as its basalmost form.

Implications for cetotheriid palaeobiology

At an estimated body length of 7.7 m (based on Lambert et al., 2010) to 8.7 m (based on Pyenson & Sponberg, 2011), Tranatocetus maregermanicum is the largest formally described cetotheriid. In stark contrast, most of the remaining members of the family do not exceed 5 m in length, and thus are relatively small compared to other mysticetes (Bouetel & de Muizon, 2006; El Adli, Deméré & Boessenecker, 2014; Gol’din & Startsev, 2017; Gol’din, 2018; Slater, Goldbogen & Pyenson, 2017). There are, however, notable exceptions, including Herentalia nigra, ‘Cetotherium’ megalophysum, an as yet unnamed species from Peru, and a fragmentary skeleton from northern Belgium (Bisconti, 2015; Bosselaers et al., 2004; Collareta et al., 2015; Slater, Goldbogen & Pyenson, 2017).

In general, larger cetotheriids appear to cluster in the early Late Miocene, whereas smaller forms –in particular, herpetocetines, and Piscobalaena—dominate during the latest Miocene and Pliocene (Bouetel & de Muizon, 2006; El Adli, Deméré & Boessenecker, 2014; Tanaka, Furusawa & Barnes, 2018b; Tanaka & Watanabe, 2018; Whitmore Jr & Barnes, 2008). Larger size plausibly correlated with a different ecological niche, with Tranatocetus perhaps being more pelagic than other cetotheriids, or targeting free-swimming schooling fish instead of benthos. The same may have applied to other large cetotheriids (Collareta et al., 2015), and possibly suggests a Late Miocene shift towards a different diet, habitat or feeding strategy.

The reasons behind this shift, if indeed it occurred, remain obscure, but it seems noteworthy that it coincided with the initial diversification of rorquals. In the Pisco Formation of Peru, for example, rorquals became locally abundant, and represented by two to three different species (Di Celma et al., 2017), at the same time as cetotheriids declined in number and size (Bianucci et al., 2016a; Bianucci et al., 2016b). We suggest that this phenomenon might be explained by niche partitioning between small, suction feeding and possibly benthic/neritic cetotheriids on the one hand (El Adli, Deméré & Boessenecker, 2014; Gol’din, Startsev & Krakhmalnaya, 2014), and large, lunge-feeding, pelagic rorquals on the other. Cetotheriids continued to occupy the ‘small (benthic) filter feeder’ niche for the remainder of the Miocene and Early Pliocene, but then largely disappeared-alongside small balaenids and most small rorquals- with the onset of Northern Hemisphere glaciation around 3 Ma (Marx & Fordyce, 2015; Slater, Goldbogen & Pyenson, 2017).

In the Northern Hemisphere, grey whales currently occupy a niche similar to that inferred for cetotheriids, which may help to explain their convergent skull morphologies (Bisconti, 2008; Steeman, 2007). The oldest described eschrichtiine material dates to the Late Miocene, but is currently restricted to a fragmentary mandible (Bisconti & Varola, 2006). More confidently identified specimens are only known from the Pliocene (Bisconti, 2008; Ichishima et al., 2006; Kimura, Hasegawa & Kohno, 2017; Whitmore Jr & Kaltenbach, 2008), raising the question for how long, and how broadly, cetotheriids and eschrichtiines overlapped. Competition between the two could plausibly have limited the resources available to cetotheriids further, even though they seemingly were more common and had a wider distribution (including in the Southern Hemisphere) (Bouetel & de Muizon, 2006). Further insights into when and where eschrichtiines acquired their modern morphology may help to answer this question.