Abstract: Pachytraga Paquier, 1900, the stratigraphically oldest genus of caprinine caprinid rudist, was previously known from only two chronospecies from a single lineage, that is the Hauterivian P. tubiconcha Astre, 1961 and the early Aptian P. paradoxa (Pictet and Campiche, 1869). Here, a new species, Pachytraga? tanakahitoshii, is erected on the basis of isolated left valves recovered from the Osaka and Sanchu areas, south-west Japan. This species has a moderate shell size (antero-posterior commissural diameter c. 30 mm), and its left valve is characterized by at least one possibly autapomorphic character (narrow anterior myophore, inclined inwards), as well as a mosaic of primitive (single longitudinal carina developed on the anterior side) and derived (simple marginal canals in the antero-dorsal valve margin) characters of Mediterranean and Middle East Tethyan Pachytraga. The Japanese Pachytraga? represents the first probable record of this genus outside the Mediterranean/Middle Eastern Tethyan province, and its early Barremian age partly fills the ‘gap’ in its previously known stratigraphical record, although the evolutionary relationship of the Japanese form with Mediterranean and Middle East Tethyan Pachytraga remains unsolved. However, the discovery of early Barremian Pachytraga? in Japan indicates that the evolutionary history of the genus is more complex than previously thought and should thus be discussed in a broader palaeogeographical context that must now include the Pacific.
The family Caprinidae d’Orbigny, 1847 constitutes the most notable component of mid-Cretaceous rudist evolution (the ‘Caprinid Phase’ of Skelton 2003) and is divided into two subfamilies, viz. the Caprininae d’Orbigny, 1847 and the Caprinuloideinae Damestoy, 1971 (Skelton and Smith 2000; Carter et al. 2011). Pachytraga Paquier, 1900 is the stratigraphically oldest genus of the former subfamily, flourishing in the Mediterranean Tethys and Middle East during the Hauterivian and in the north-western part of the former region during the early Aptian (Skelton and Masse 1998, 2000; Masse and Fenerci-Masse 2008). Because Hauterivian P. tubiconchaAstre, 1961 and early Aptian P. paradoxa (Pictet & Campiche, 1869) have been considered as successive chronospecies (Skelton and Masse 1998), the absence of this genus from the known record for most of the Barremian is an enigma, and possible controls on its restriction to some unknown refugium have been discussed by Masse and Fenerci-Masse (2008). Other species have been assigned to the genus, but hitherto, none of these attributions has survived revision (Skelton and Masse 1998). In particular, Okubo and Matsushima (1959) recorded Pachytraga japonica Okubo inOkubo and Matsushima, 1959 from the Lower Cretaceous Shimanto Group of the Akaishi Mountains in central Honshu, Outer Zone (Pacific side) of south-west Japan. However, Skelton and Masse (1998) restudied the holotype and reassigned it to the polyconitid genus PraecaprotinaYabe and Nagao, 1926, which had been previously established on the basis of material from the Miyako Group in north-east Honshu Island and the Yezo Group in central Hokkaido, both north-east Japan (Yabe and Nagao 1926). In addition, material referred to Pachytraga japonica from the putative Barremian–Aptian Osaka Formation in the Osaka area, eastern Kyushu, south-west Japan (Tanaka 1989; Tanaka et al. 1996) has also been reassigned to Praecaprotina, and consequently, doubt has been cast on the existence of Pachytraga in Japan (Sano et al. 2008). Nevertheless, investigation of a specimen referred to ‘Pachytraga japonica’ by Ichise (2008), from the Kanto Mountains near Tokyo, and newly collected specimens from the Osaka area reveal that these specimens show a close similarity to Pachytraga and represent the first probable record of the genus outside the Mediterranean Tethys and Middle East. Interestingly, reinvestigation of age-diagnostic ammonites previously known from these localities indicates their early Barremian age, falling within the ‘gap’ in the known stratigraphical record of this genus in the Mediterranean/Middle Eastern Tethyan province (Masse and Fenerci-Masse 2008).
In this study, the new Japanese Pachytraga? is formally named as P?. tanakahitoshii sp. nov. Putative records of other caprinines from the Pacific are also briefly reviewed from the viewpoint of their similarity with those in the Mediterranean Tethyan province. The discovery of early Barremian Pachytraga? in Japan indicates that the evolutionary history of Pachytraga and also of caprinine rudists in general is more complicated than previously thought and should be discussed in a broader palaeogeographical context that must now include the Pacific.
Geological setting and age
The rudist specimens described here were recovered from two localities: Osaka (eastern Kyushu Island) and Sanchu (central Honshu Island; Fig. 1). The precise age of the rudist-bearing strata in the Osaka and Sanchu areas is still controversial, because of complex geological structures and also the scarcity of well-preserved age-diagnostic fossils. In this study, the geological age of these rudist-bearing strata is discussed chiefly on the basis of ammonites collected from areas adjacent to the rudist localities.
A single specimen of Pachytraga? was discovered in a sandstone float from locality OS03 (32°59′N, 131°38′10″E) of Tanaka (1989) and Tanaka et al. (1996) in the Osaka area, Mie Town (Bungoono City, Oita Prefecture). The upper part of the Osaka Formation, containing abundant fossils of shallow-marine origin, crops out in this area (Tanaka 1989; Yokomizo et al. 1990; Tanaka et al. 1996). Yokomizo et al. (1990) reported the presence of two ammonite taxa, Eodouvilleiceras(?) sp. aff. E. horridum (Riedel, 1938) and Hamites sp. aff. H. attenuatus (Sowerby, 1814) from their ‘Northern Belt of the Osaka Formation’ and broadly assigned a late Barremian–Albian age to the whole Osaka Formation, based on these ammonites and other fossils. Tanaka et al. (1996) recorded Cheloniceras sp. and Dufrenoyia aff. justinae (Hill, 1893) from their ‘Upper Member of the Osaka Formation’, where the Pachytraga? specimen here described was recovered and considered its age to be late Aptian, although Iba and Sano (2007) later mentioned that these ammonites probably indicated an early Aptian age. However, restudy of the ammonites recorded by Yokomizo et al. (1990) and newly collected specimens from the Osaka area reveals the presence of Macroscaphites? sp., Crioceratites sp., Phyllopachyceras sp. and Shasticrioceras sp. in the Osaka Formation (Fig. 2), suggesting an early Barremian age, in comparison with the Early Cretaceous ammonite fauna of south-west Japan (e.g. Obata et al. 1982; Matsukawa and Obata 1993; Toshimitsu and Hirano 2000; Matsukawa et al. 2007). Unfortunately, the possible early Aptian ammonite specimens reported by Tanaka et al. (1996) were broken after the preliminary investigation, and these records cannot now be confirmed. Thus, the age of the rudist-bearing part of the Osaka Formation is here considered as early Barremian, although it could be extended to the early Aptian. However, it should be noted that the finding of Praecaprotina (= Pachytraga japonica in Tanaka et al. 1996) and a relatively advanced caprinuloideine rudist from the Osaka area (Sano et al. 2008) do suggest a younger age (possibly Aptian), while complex geological structures in this area are also suspected.
Three specimens of Pachytraga? were obtained from granule conglomerate and sandstone beds of the Tozawa Formation at localities Tz-03 (36°8′14″N, 138°39′4″E) and Tz-19 (36°7′45″N, 138°39′26″E) of Ichise (2008), north of the Jikkoku Pass, in the Sanchu area (Kanto Mountains, central Honshu Island). Ichise (2008) defined the Tozawa Formation (Nanmoku Group) in this area, although the same strata have been assigned to the upper Hauterivian–Barremian Ishido Formation of the Sanchu Cretaceous in other studies (e.g. Matsukawa 1983; Takei 1985; Matsukawa and Tomishima 2009). Ichise (2008) reported Paracrioceras cf. asiaticum (Matsumoto, 1947) from float at locality Tz-03 and also mentioned the report by Hisada et al. (1992) of Paracrioceras aff. elegans (von Koenen, 1902), Shasticrioceras nipponicumMatsumoto, 1947 and Shastricrioceras(?) sp. from the nearby area of Tz-03, concluding that these ammonites indicated a Barremian age, by reference to Matsukawa and Obata (1993). However, Ichise (2008) assumed the age of the Tozawa Formation to be Barremian(?) to Aptian, probably Aptian, from the similarity of in situ bivalve fossils to those of the putative Aptian Kesado and Osaka formations in Kyushu in preference to the age given by ammonites recovered from float. Recent findings of ammonites in the Kesado and Osaka formations suggest, rather, that at least some (possibly most) parts of them are, after all, Barremian in age (Tanaka et al. 2009, and see paragraph on the Osaka Formation above). Thus, it is not necessary to consider the age of the Tozawa Formation as Aptian by comparison with the Kesado and Osaka formations, and the age of the Pachytraga?-bearing stratum in the Sanchu area is again taken to be early Barremian in this study, applying the ammonite age assignment in Matsukawa and Obata (1993) more strictly.
Institutional abbreviations. KSG, Department of Earth Science, Faculty of Science, Kochi University, Kochi; UMUT, University Museum, University of Tokyo, Tokyo, Japan.
Type species. Sphaerulites paradoxaPictet and Campiche, 1869, from erratic blocks of Urgonian limestone in the ‘Plaine des Rocailles’, between Regny and La Roche, Switzerland, by the subsequent designation of Paquier (1905).
non 1959 Pachytraga japonica Okubo in Okubo and Matsushima, p. 1, figs 1–7.
non 1989 Pachytraga sp. Tanaka, p. 28, pl. 4, fig. 15.
non 1996 Pachytraga japonica Okubo and Matsushima [sic]; Tanaka et al., p. 37, pl. 6, figs 4, 5.
non 1999 Pachytraga sp. cf. P. japonica Okubo and Matsushima [sic]; Tanaka et al., p. 71, pl. 7, figs 5, 6.
2008 ‘Pachytraga japonica’ Okubo and Matsushima [sic]; Ichise, p. 51.
Derivation of name. In honour of Dr. Hitoshi Tanaka, Professor of Kumamoto University, for his contributions to the studies of Cretaceous bivalves and stratigraphy of south-west Japan for over thirty years.
Holotype. KSG-ss010 (Fig. 3A–D), internal and partial external mould of incomplete left valve, collected by H. Tanaka, Y. Iba and S. Sano from the Osaka area (see previous section), eastern Kyushu Island, south-west Japan.
Other material examined. Three moulds of incomplete left valves; KSG-ss011-1, both external and internal mould (Fig. 3G–I) from Tz-19; KSG-ss011-2, internal mould (Fig. 3E, F) from the same rock specimen with KSG-ss011-1, and KSG-ss012, internal mould (Fig. 3J, K) from Tz-03, collected and reported by M. Ichise (maiden name of M. Watarai) from the Sanchu area (see previous section), Kanto Mountains, central Japan (Ichise, 2008).
Diagnosis. Moderate-sized shell (antero-posterior commissural diameter c. 30 mm) with Pachytraga-like cardinal and, apparently, posterior myophoral arrangement; commissural outline almost oval, somewhat compressed dorso-ventrally; left valve domed to conically capuloid, with a longitudinal carina developed on the anterior side; anterior myophore of the left valve relatively recessed with only a narrow connection to the base of the anterior tooth and inclined into the valve interior; small marginal canals may be developed on the antero-dorsal margin of the left valve.
Description. The left valve is low-domed in shape and the umbo little extended (Fig. 3C, D, H, K) with a single longitudinal carina on the anterior side (Fig. 3G, I); the commissural outline may be almost oval in shape, with the maximum diameter-oriented antero-posteriorly (Fig. 3A, B, G). The two teeth are unequal, with the bluntly conical anterior tooth significantly larger than the posterior one in the left valve (Fig. 3B, D, F). A large endomyophoral cavity adjoins the central tooth socket and is separated from the body cavity by a thin vertical wall extending from the anterior tooth to the postero-ventral valve margin (Fig. 3B, D, F, G, J). The erect posterior myophoral ridge is situated on the posterior valve wall and connects dorsally to the posterior tooth (Fig. 3B, D, F); in none of the specimens, unfortunately, is its ventral termination preserved, but the visible remains suggest a salient ridge fully seated on the posterior valve wall (Fig. 3A, B, E, F). The anterior myophore of the left valve is relatively recessed along the shell wall with connection only to the base of the anterior tooth and inclined into the interior of the valve (Fig. 3B). Small marginal canals are developed on the antero-dorsal margin of the left valve of the holotype (Fig. 3A–B). The right valve of this species has not been recovered yet.
Remarks. The new species shares diagnostic characters of cardinal and posterior myophoral arrangement with Pachytraga: two unequal teeth with the bluntly conical anterior tooth significantly larger than the posterior one; posterior tooth and erect posterior myophoral ridge connecting dorsally to the posterior tooth both situated on the valve wall; a large endomyophoral cavity adjoining the central tooth socket, separated from the body cavity by a thin vertical wall extending from the anterior tooth to the postero-ventral valve margin.
The two already known species of Pachytraga, P. tubiconcha and P. paradoxa, generally have similar morphological features and have been considered as successive chronospecies showing phyletic size increase (Skelton and Masse 1998). The size of the Japanese Pachytraga? is just above the size range of P. tubiconcha shown in Skelton and Masse (1998). However, some differences between the Japanese form and either, or both, of the two previously known species of Pachytraga are evident. Thus, in the Japanese specimens, an external posterior longitudinal ridge is not developed as it is in P. tubiconcha. On the other hand, although small marginal canals in the antero-dorsal side of the left valve are not known in P. tubiconcha, they are in some specimens of P. paradoxa (Skelton and Masse 1998, 2000). Yet, left valves of the latter species, by contrast, usually show a small, but distinct mid-ventral ridge extending down inside the body cavity (Skelton and Masse 1998), a feature not seen in any of the Japanese specimens.
Moreover, all specimens of the Japanese Pachytraga? have a low-domed left valve and do not show the projecting, prosogyrally enrolled umbo often observed in both P. tubiconcha and P. paradoxa (Fig. 3). Furthermore, the commissural outline is somewhat compressed dorso-ventrally in the Japanese form, in contrast to those of P. tubiconcha and P. paradoxa, which are slightly compressed antero-posteriorly. Finally, and perhaps most significantly in terms of taxonomic characterization, the anterior myophore of the left valve in the Japanese form is relatively narrow and inclined into the valve interior, whereas those in both P. tubiconcha and P. paradoxa are broader and flat or even tilted upwards to face out onto the depressed anterior myophore on the anterior wall of the right valve (Skelton and Masse 1998). Thus, the Japanese Pachytraga? has a puzzling mosaic of characters of both P. tubiconcha and P. paradoxa, as well as some possible autapomorphic characters, so can be distinguished from the two known species of Pachytraga.
Mention should also be made of a single left valve reported from the lower Aptian of central Greece by Baron-Szabo and Steuber (1996) under the name of P. paradoxa. This latter form does have a low-domed shape with dorso-ventrally compressed commissural outline and shows little umbonal protrusion, similar to Pachytraga? tanakahitoshii sp. nov. Indeed, Skelton and Masse (1998, p. 347) already regarded these same characters as casting doubt on the assignment of the Greek specimen to P. paradoxa. However, the larger size and acute projection of the anterior part of the valve in the latter, and, particularly, its broad anterior myophoral shelf (like that in P. paradoxa) are distinct from the Japanese form. The similarities in external shape may thus be merely convergence. In fact, the taxonomic status of the Greek specimen itself remains problematical, as its relatively broadly inward-sloping posterior myophore might alternatively suggest kinship with the Himeraelites/Sellaea group (family Caprinulidae Yanin, 1990 in Carter et al. 2011) rather than with Pachytraga– a possibility that remains to be tested by reference to other, as yet unpublished material from the same locality, including diagnostically important right valves.
Systematic position. Although Pachytraga? tanakahitoshii sp. nov. shares its cardinal and posterior myophoral arrangement with Pachytraga, some of its morphological characters, such as the relatively unextended umbo, dorso-ventrally compressed commissural outline, and the narrow and inwardly inclined form of the anterior myophore of the left valve are not observed in its congeners. Thus, if the Japanese form is eventually assigned to the genus Pachytraga, the diagnosis of the latter would need to be amended correspondingly. However, we refrain from doing so at present in view of the limitations of the material so far available, especially the lack of right valves.
The above-mentioned characters of the Japanese Pachytraga?, however, can be observed in at least some specimens of right valve-attached rudists of Late Jurassic age, such as ‘Valletia’antiqua Favre inJoukowsky and Favre (1913) and ‘Valletia’auris Favre inFavre and Richard (1927; SS and PWS, pers. obs.). Such information leaves open two possibilities: either the Japanese Pachytraga? was derived independently of the Mediterranean and Middle East Tethyan Pachytraga lineage, from a more primitive common ancestral form, despite sharing the relatively derived character of antero-dorsal canals with P. paradoxa (which, in that case, would thus be homoplasious); or it diverged from the latter’s lineage, autapomorphically re-assuming a more primitive anterior myophoral form. In the former case, the Japanese form could be placed in a separate, more primitive genus, possibly tracing back to the Late Jurassic; in the latter case, the Japanese form should remain within the genus Pachytraga. This question cannot be settled without further information, especially on the form of the right valve. Thus, the Japanese form is tentatively assigned to Pachytraga here.
Distribution and stratigraphic range. Probably lower Barremian in the Osaka and Sanchu areas, south-west Japan.
Implications for evolutionary history of the genus Pachytraga
The distribution of Pachytraga has recently been summarized by Masse and Fenerci-Masse (2008). The Hauterivian P.tubiconcha is recorded from south-east France, Switzerland, Sardinia, southern Spain, Portugal, Algeria and Oman. On the other hand, its early Aptian descendant, P. paradoxa, is known only from south-east France, Spain, Portugal and Algeria. Although Baron-Szabo and Steuber (1996) noted P. paradoxa also from central Greece, this assignment is questionable, as discussed above. The absence of Pachytraga from the known record for most of the Barremian in the Mediterranean/Middle Eastern Tethyan province is an enigma (Masse and Fenerci-Masse 2008). In addition, Skelton and Masse (1998) mentioned possible early Aptian examples (Pachytraga? sp.) from Cuba in the Caribbean Province. Subsequently, the rudist limestones at the locality of Tumbadero in Cuba were considered more likely to be Barremian, because of the presence there of Offneria simplexChartrousse and Masse, 1998b. The Japanese early Barremian Pachytraga? is, perhaps together with the above-mentioned Cuban material, not only the first possible record of this genus outside the Mediterranean/Middle Eastern Tethyan province, but also partially fills the ‘gap’ in the known stratigraphical record, although the reasons for the disappearance and revival of the genus in the Mediterranean/Middle Eastern Tethyan province remain unsolved.
Furthermore, as the early Barremian P.? tanakahitoshii sp. nov. has a mosaic characters of Hauterivian P. tubiconcha and early Aptian P. paradoxa, and also some possible autapomorphic characters, it could be an endemic north-west Pacific-margin offshoot from the main Pachytraga lineage or it could even have been derived from an older common ancestor. Either way, the evolutionary history of Pachytraga is probably more complex than previously thought and should be discussed in a broader palaeogeographical context that must now include the Pacific and possibly the Caribbean, at least in the Barremian.
Comments on putative caprinine records in the Pacific
Two biotic provinces, Mediterranean and Caribbean, were established in the Tethyan Realm from the early Aptian onwards (Chartrousse and Masse 2004), probably because of separation due to the formation of the Atlantic (e.g. Coates 1973). As the Caprininae flourished in both the Mediterranean (largely) and Caribbean provinces, and the Caprinuloideinae mainly in the Caribbean province, caprinid rudists are considered as key palaeobiogeographic indicators (Chartrousse and Masse 2004). In this context, the presence of caprinuloideine rudists in the Central Pacific has attracted attention in recent palaeobiogeographic studies of the Pacific (e.g. Premoli Silva et al. 1995; Chartrousse and Masse 1998a; Skelton et al. 2011).
Putative caprinine rudists, however, have been recorded from the Pacific, though their systematic assignments have proved controversial. Caprina mulleri and C. mediopacifica were described by Hamilton (1956), based on the dredged materials from Cape Johnson Guyot in the Mid-Pacific Mountains. Later, Swinburne and Masse (1995) suggested that these species were probably synonymous, but questioned their relationship to Caprina or even the known subfamilies of the Caprinidae, while Chartrousse (1998, pp. 190, 191) proposed reassignment to a new genus within the ‘Coalcomaninae Coogan, 1973’ (= Caprinuloideinae). However, reinvestigation of the holotype of C. mulleri deposited at the Smithonian National Museum of Natural History (Washington, DC) and additional Pacific specimens confirms its original generic attribution, indeed emphasizing similarities with the coeval late Albian C. choffatiDouvillé, 1898 from Iberia, for example in having opisthogyral umbones (Skelton et al. 2011; PWS and SS, work in progress with J-P Masse).
Although Swinburne and Masse (1995) reported possible caprinine rudists of early Aptian age from Resolution Guyot in the Mid-Pacific Mountains, Chartrousse and Masse (1998a) formally described well-preserved rudist specimens including those discussed by Swinburne and Masse (1995) as a new genus and species, Conchemipora skeltoni, which they assigned to the Coalcomaninae (= Caprinuloideinae). Thus, the presence of caprinines in the Mid-Pacific Mountains in the early Aptian is, as yet, unsubstantiated.
Shikama and Tanabe (1970) described Caprina uwajimensis from the Uwajima area, western Shikoku Island, south-west Japan. Our reinvestigation of the holotype and newly collected specimens from a nearby locality reveals that the putative pallial canals of ‘C. uwajimensis’ are not located in the inner shell layer, but in the outer shell layer. Thus, it does not belong to the Caprinidae. Tightly pleated infoldings of the outer shell layer indicate its affinity, rather, to certain specialized polyconitids, such as Horiopleura? juxiSteuber, 1999 from Greece (Steuber et al. 2011).
In summary, we can confirm the presence of at least two probably caprinine species in the Pacific: early Barremian Pachytraga? tanakahitoshii sp. nov. from Japan and late Albian Caprina mulleri from the Mid-Pacific Mountains. Both species show strong affinities with Mediterranean species, implying the expansion of caprinine distribution to the Pacific at least in early Barremian and late Albian times, although the connecting routes remain unknown.
Two enigmatic bio-events in the evolutionary history of the caprinine rudists in the Mediterranean/Middle Eastern Tethyan province are the absence of Pachytraga for most of the Barremian, and of the whole of the subfamily in the late Aptian–middle Albian (e.g. Skelton 2000; Masse and Fenerci-Masse 2008). The ‘rediscovery’ of caprinine rudists in the Pacific possibly provides the clue in the search for the refugia of the caprinines during these lacunae in their known stratigraphical record. Further studies of the Pacific and possibly northern Tibet rudist records (e.g. Scott et al. 2010) are necessary to reveal the missing chapters in the evolutionary history of these rudists.
Acknowledgements. We acknowledge helpful discussions with Jean-Pierre Masse (Université de Provence, Marseille) concerning the taxonomy of the rudists discussed herein. We are grateful to Hitoshi Tanaka (Kumamoto University) and Ken-ichiro Hisada (University of Tsukuba) for their kind help in conducting the geological and palaeontological studies of the Osaka Formation and the Sanchu Cretaceous. We would like to thank Thomas Steuber (Petroleum Institute, Abu Dhabi) and an anonymous reviewer for critical comments on an earlier version, and John W. M. Jagt (Natuurhistorisch Museum Maastricht) for his editorial work. Thanks are extended to Yasunari Shigeta and Yuta Shiino (National Museum of Nature and Science, Tokyo) for their help in identification and taking the photographs of ammonite specimens, respectively.