Oryzias, Jordan & Snyder, 1906

Benton, MJ, Donoghue, PCJ, Vinther, J, Asher, RJ, Friedman, M & Near, TJ, 2015, Constraints on the timescale of animal evolutionary history, Palaeontologia Electronica (Florence, Italy) 15 (1), pp. 1-107 : 40-42

ORYZIAS , CICHLIDAE – GASTEROSTEUS, TAKIFUGU, TETRAODON (35)

Node Calibrated. Divergence between Ovalentaria and Tetraodontiformes (Wainwright et al., 2012) , corresponding to an unnamed subclade of Percomorpha.

Fossil Taxon and Specimen. Cretatriacanthus guidottii from the ‘Calcari Melissano’ (of historical usage) at Canale near Nardò, Italy (holotype MCSNV 1377 , Museo Civico di Storia Naturale , Verona, Italy) .

Phylogenetic Justification. Molecular hypotheses of teleost relationships indicate that the clade defined by the divergence between Ovalentaria and Tetraodontiformes comprises much of Percomorpha (Near et al., 2013; Betancur-R. et al., 2013). Many generalized percomorphs and ‘perciforms’ are known from Late Cretaceous deposits (e.g., Patterson, 1993a; López-Arbarello et al., 2003; Arratia et al., 2004; Taverne, 2010). However, none of these has been placed with any precision within percomorph phylogeny, and as such are not appropriate calibrations.

The oldest fossils that can be placed with any confidence within Ovalentaria are all late Paleocene-Eocene in age (Patterson, 1993b; Bellwood and Sorbini, 1996; see node 37). However, there are several reports of Cretaceous tetraodontiforms (Patterson, 1993b; Tyler and Sorbini, 1996; Santini and Tyler, 2003; Gallo et al., 2009; Friedman, 2012b; Tyler and Križnar, 2013). We select Cretatriacanthus , from Nardò, Italy as a conservative minimum constraint for the divergence between Tetraodontiformes and Ovalentaria. Cretatriacanthus represents the youngest putative tetraodontiform of Cretaceous age represented by body-fossil remains. The absence of an elaborate dermal carapace in Cretatriacanthus means that key tetraodontiform features like the absence of pleural ribs and geometry of the pelvic girdle are readily apparent in this genus, whereas they are generally obscured by large bony plates in other Cretaceous examples.

Minimum Age. 69.71 Ma

Soft Maximum Age. 130.8 Ma

Age Justification. The fish-bearing limestones at Nardò are generally referred to the ‘Calcari di Melissano’, a name historically applied to the Late Cretaceous platform carbonates of Salento (Martinis, 1967). However, usage has since been restricted to one part of the Late Cretaceous carbonate succession in the Apulian platform (e.g., Bosellini and Parente, 1994; Schlüter et al., 2008). Medizza and Sorbini (1980) provide a list of calcareous nannofossil species recovered from the fish-bearing layers, the most biostratigraphically relevant of which is Uniplanarus trifidus (reported as Quadrum trifidum ). The first appearance of this species marks the beginning of Calcareous Nannoplankton Zone CC23, and it makes its last appearance in the middle of CC24. The top of CC24 is roughly equivalent to the top of the Baculites clinolobatus Ammonite Zone of the Western Interior Seaway, which contains a bentonite horizon dated as 70.08 Ma ± 0.37 Myr (Ogg et al., 2012b). It is from this value that we derive our minimum age estimate of 69.71 Ma.

Percomorphs, and acanthomorphs more generally, are unknown from a series of fish faunas of mid-late Early Cretaceous age that represent a range of depositional settings from fully marine to lacustrine: the Gault Clay of England (Albian; Gale and Owen, 2010; Forey and Longbottom, 2010; Nolf, 2010), Helgoland in Germany (Aptian; Taverne, 1981b), the Crato Formation of Brazil (Martill, 1993; early interpretations of Araripichthys as a lampridiform have been decisively rejected by Patterson, 1993a and Maisey and Moody, 2001), and the Coquiero Seco Formation of Brazil (Gallo and Cohelo, 2008). The oldest of these deposits, the Coquiero Seco Formation of Brazil, yields the oldest putative representative of Eurypterygii (the clade containing Aulopiformes, Myctophiformes, and Acanthomorpha), and provides the basis for our estimated maximum age divergence between Tetraodontiformes and Ovalentaria. The Barremian is dated to approximately 130.8-126.3 Ma (Ogg et al., 2012b); we derive our soft maximum from the oldest limit.

Discussion. Many Cretaceous fossils have been interpreted as tetraodontiforms. Fragmentary examples from the Maastrichtian represent dermal scutes of batoids (e.g., materials summarized by Patterson, 1993b and Friedman, 2012b) or are of questionable stratigraphic provenance ( Gallo et al., 2009); such remains are clearly not appropriate as calibrations. Articulated remains identified as tetraodontiforms require more careful consideration. Tyler and Sorbini (1996) interpreted three remarkable genera of Cretaceous fishes as tetraodontiforms based on detailed description of such material: Plectocretacicus , from the early Cenomanian of Hakel, Lebanon; Protriacanthus , from the late Cenomanian-early Turonian of Comen, Slovenia; and Cretatriacanthus , from the late Campanian-early Maastrichtian of Nardò, Italy. To these, Tyler and Križnar (2013) have recently added a fourth genus: Slovenitriacanthus , from the late Santonian-early Campanian Lipica Formation of Slovenia. These are morphologically heterogeneous, but are all no more than 50 mm in total length and share modified scales or plates at the base of the pelvic-fin spine and an absence of teeth (Tyler and Sorbini, 1996). Tyler and Sorbini (1996) presented a manual cladogram uniting Plectocretacicus , Protriacanthus , and Cretatriacanthus as a clade named Plectocretacicoidea on the tetraodontiform stem. Santini and Tyler (2003) corroborated this result using a numerical cladistic analysis, which has since been replicated by re-study of their morphological characters in concert with molecular data ( Arcila et al., 2015).

Since the initial description and placement of plectocretacicoids, genetic data have substantially revised past morphological interpretations of the living sister group of Tetradontiformes (e.g., Rosen, 1984). Molecular analyses consistently place lophiiforms as the closest extant relatives of tetraodontiforms (Near et al., 2012, 2013; Betancur-R. et al., 2013). Many of the features used to argue for the tetraodontiform affinity of plectocretacicoids (e.g., restricted opercular opening, absence of a beryciform foramen, absence of ribs, absence of an anal-fin spine, reduced vertebral count, reduced caudal-fin ray count) also characterize many or all lophiiforms (Regan, 1912; Pietsch, 1981, 1984; Chanet et al., 2013), suggesting that these traits are candidate synapomorphies of a more extensive clade including lophiiforms and tetraodontiforms. Unfortunately, published analyses incorporating plectocretacicoids do not sample lophiiforms (see comments in Santini et al., 2013, p. 178-179), meaning that available solutions cannot test placement of these fossils within the tetraodontiform total group. Perhaps more critically, tetraodontiforms and lophiiforms share several derived features not apparent in plectocretacicoids, the most notable of which include: absence of infraorbital bones (present in plectocretacicoids); sutural relationship between the posttemporal and cranium (plectocretacicoids have a non-structural connection comparable to that of generalised acanthomorphs); six or fewer branchiostegal rays (a feature also shared with immediate outgroups to [ Tetraodontiformes + Lophiiformes ] including Caproidae , Priacanthidae , Cepolidae , and Siganidae ; Plectocretacicus is the only plectocretacicoid where branchiostegals are visible, and has seven); lateral-line system unenclosed in bony canals (pores for bony canals are present in plectocretacicoids); and absence of procurrent caudal rays (present in plectocretacicoids; loss of procurrent rays is interpreted as crown tetraodontiform synapomorphy by Santini and Tyler, 2003, with inferred reversals within the group) (McAllister, 1968; Pietsch, 1981; Fujita, 1990; Nakae and Sasaki, 2010; Carnevale and Pietsch, 2012). These observations indicate either considerable parallelism between tetraodontiforms and lophiiforms, numerous reversals in plectocretacioids, or that plectocretacicoids are not tetraodontiforms. Selection between these contrasting alternatives requires a formal analysis, which we are not in a position to provide here.

Of the plectocretacicoids, Cretatriacanthus represents the most conservative candidate for establishing a minimum age for Tetraodontiformes , or at least the extended tetraodontiform-lophiiform clade, because it lacks traits that suggest alternative phylogenetic placements. Protriacanthus shares derived features of the skull (e.g., an extended rostrum, subterminal mouth, absence of an ascending process of the premaxilla, absence of teeth), dorsal and anal fins (greatly reduced finray count, with approximately five rays in each), and caudal-fin skeleton (e.g., extensive fusion between hypurals, parhypural, and associated centra) with pegasids (Pietsch, 1978), and many of the tetraodontiform features preserved in the genus are also characteristic of this other group (e.g., absence of an anal-fin spine, reduced caudal-fin ray count, reduced vertebral count). Molecular phylogenies place pegasids as close relatives of syngnathiforms, which branch outside the most restrictive clade containing Ovalentaria and Tetraodontiformes . Plectocretacicus , the oldest of the candidate tetraodontiforms, broadly resembles three new genera of small, armoured acanthomorphs of mid-Cretaceous age that González-Rodríguez et al. (2013) identified as beryciforms or, more doubtfully, as non-tetraodontiform percomorphs. These Mexican taxa share a series of clear similarities with Plectocretacicus , including: small size; a bony carapace formed of sutured ( Plectocretacicus ) or overlapping ( Handuichthys , Pseudomonocentris , Dalgoichthys ) enlarged plates often bearing central bosses and radiating tuberculation; absence of teeth; absence of dorsal- and anal-fin spines (bases of supposed anal-fin spines in the Mexican taxa are always preserved on the lateral face of specimens, suggesting they are in fact pelvic spines); posteriorly located pelvic fins (following the reinterpretation provided for the Mexican taxa); embrace of pelvic-fin base by dermal shield; and a very narrow caudal peduncle. Non-tetraodontiform interpretations of morphologically similar taxa of comparable age suggest some ambiguity in the placement of Plectocretacicus .

Our proposed calibration for this node is revised relative to previous reviews of calibrations for major divergences in animal phylogeny (Benton and Donoghue, 2007; Benton et al., 2009), which advocated the use of Plectocretacicus as a fossil-based minimum. Tetraodontiformes represent one of the most highly nested clades within acanthomorph phylogeny, and are resolved in an apical position within the percomorph tree (Smith and Wheeler, 2006; Near et al., 2012, 2013; Betancur- R. et al., 2013). In addition to the reservations we have expressed about the placement of Plectocretacicus with tetraodontiforms on anatomical grounds, it is remarkable that an early representative of such a nested clade would appear in the record without contemporary examples of several additional, identifiable percomorph lineages. Other acanthomorphs known from the early Cenomanian include a diversity of beryciforms, polymixiids, paracanthopterygians, and representatives of various extinct groups (e.g., aipichthyids, pharmacichthyids) recently algined with lampridiforms (Davesne et al., 2014). Assignment of Plectocretacicus to Percomorpha is therefore unique among acanthomorphs of early Cenomanian age (Patterson, 1993a, b; Forey et al., 2003). The contrasting placements of superficially similar fossils offered by González-Rodríguez et al. (2013) suggest that this genus might alternatively be interpreted as an armored beryciform. If the Cenomanian-Turonian Protriacanthus is a percomorph, its affinities might lie with pegasids rather than tetraodontiforms. Apart from enjoying the anatomical support discussed above, this alternative placement is more in line with stratigraphy; pegasids belong to one of the earliest diverging percomorph radiations (Smith and Wheeler, 2006; Near et al., 2013; Betancur-R. et al., 2013). The Campanian-Maastrichtian Cretatriacanthus is coeval with representatives of a series of percomorph lineages, including primitive syngnathiforms (Sorbini, 1981; Patterson, 1993a, b) and incertae sedis ‘percoids’ (Taverne, 2010). This lends further circumstantial support to our assertion that Cretatriacanthus is the most appropriate minimum marker for the origin of the percomorph clade including Oryzias and Tetraodontiformes . However, our choice of a generous soft maximum age allows for the possibility that subsequent analyses with suitable taxon sampling might demonstrate that Plectocretacicus is indeed a member of the tetraodontiform total group, and thus the appropriate calibration for this node (cf. Benton and Donoghue, 2007; Benton et al., 2009).