Introduction

The Jurassic was a turning point in the evolution of Foraminifera. According to Fuchs (1975) and Tappan and Loeblich (1988), this geological period records the origination of calcitic Rotaliida and “Buliminida” (sensu Loeblich and Tappan 1992) from aragonitic Robertinida. The first foraminiferal adaptation to a planktonic, meroplanktonic (BouDagher-Fadel et al. 1997; Hart et al. 2003, 2010; Hudson et al. 2005) or tychopelagic (Rigaud et al. 2015b) lifestyle also took place during the Jurassic, through the robertinid family Favusellidae. Despite their great potential to reconstruct and understand the large-scale evolution of foraminifers, aragonitic multichambered forms (order Robertinida) have been so far overlooked in the fossil record, most likely because their tests are rapidly and often severely affected by diagenesis, rendering their study very sensitive.

Representatives of the family Oberhauserellidae have long been believed to be the only Robertinida survivors of the Triassic-Jurassic major extinction event. In consequence, they have been regarded as the most probable direct ancestors of all calcitic Rotaliida, “Buliminida”, and “Globigerinida” and all aragonitic Ceratobuliminoidea and Conorboidoidea (e.g., Fuchs 1973, 1975; Tappan and Loeblich 1988; BouDagher-Fadel et al. 1997; Hillebrandt 2010, 2012). This monophyletic hypothesis, however, has been contested by molecular studies (e.g., Darling et al. 1997, 2009; de Vargas et al. 1997; Bowser et al. 2006; Schweizer et al. 2008; Ujiié et al. 2008; Pawlowski et al. 2013) and challenged by the reconstruction of the Robertinida early evolution (Rigaud et al. 2015b).

By the description of new Early Jurassic aragonitic multichambered foraminifers descending from two distinct Triassic non-oberhauserellid lineages, this paper aims at definitely dismissing a potential derivation of all Jurassic Robertinida from a unique oberhauserellid ancestor. This discovery opens new perspectives for the long term evolution of the group and its potential phylogenetic links with calcitic multichambered foraminifers (i.e., order Rotaliida sensu lato; i.e., sensu Pawlowski et al. 2013).

Fig. 1.

Studied locality and sampled level. A. Map of Austria showing the location of the Northern Calcareous Alps. B. Enlargement of the area of Salzburg showing the location of Adnet (Austrian Gauss-Krüger coordinate grid). C. Location of the sampled quarries (quarries XXX, XXXI, roman quarry numbering follows Kieslinger 1964). D. Picture of the heavily mineralized, ammonite-rich marmorea crust, as exposed in October 2012 in quarry XXXI. E. Quarry wall in the Rot-Grau Schnöllbruch (quarry XXXI, situation in 1994) exposing the upper part of the Schnöll Formation, the marmorea crust (M.C.) and the lower part of the Adnet Formation. F. Lithologic section of the quarry XXXI and sampling location.

Institutional abbreviations.—MHNG, The Museum of Natu ral History of Geneva, Switzerland.

Geological setting

The studied material comes from Hettangian-Sinemurian limestone rocks of Adnet (Northern Calcareous Alps, Austria). Located approximately twelve kilometers SSE of the city of Salzburg, the village of Adnet is well-known for its Early Jurassic fossil-rich rocks, which are exposed in neighboring quarries (Fig. 1). Our material comes from the Eisenmann Quarry and the Schnöll Quarry (see Fig. 1C: quarries XXX and XXXI according to Kieslinger 1964, respectively). There, the marmorea crust (= “Brandschicht” in the terminology of local quarrymen), an uppermost Hettangian to lowermost Sinemurian guide horizon (Schlotheimia angulata Zone; see Dommergues et al. 1995; Böhm et al. 1999; Böhm 2003), represents a multiphased ferromanganese hardground, which rests directly on Upper Triassic reefal limestone (Oberrhätkalk) in quarry XXX and separates the Schnöll Formation from the Adnet Formation in quarry XXXI (Fig. 1D–F). This condensed, heavily mineralized horizon is particularly well-known for its richness in ammonites (dominated by Schlotheimia marmorea) and exceptionally diversified foraminiferal assemblages (Blau 1987; Blau and Grün 1997; Ebli 1997; Böhm et al. 1999; Rigaud et al. 2013, 2015a).

We sampled the marmorea crust and the uppermost part of the Schnöll Formation, near the marmorea crust. The red to grey limestone rocks of the marmorea crust are rich in echinoderms, ammonites, gastropods, and foraminifers dominated by Involutina (Rigaud et al. 2015a) and other aragonitic forms (i.e., diverse Involutinida and Robertinida). Although these rocks have been a subject of extensive foraminiferal studies in the past (see references above), the forms introduced in the present paper have neither been mentioned nor figured by previous authors.

Material and methods

Aragonite is metastable and particularly susceptible to dissolution. Aragonitic fossils are therefore very rarely recovered in a preserved state. Only few robertinid fossil tests have been directly analyzed by mineralogical determination methods (Wiewióra 1964; Fuchs 1969; Gorbatchik and Kuznetsova 1986; di Bari and Rettori 1996; BouDagher-Fadel et al. 1997; di Bari and Laghi 1998; di Bari 1999). However, since all fossil types do not respond equally to diagenesis, a detailed textural analysis of diagenetic results can allow the successful identification of originally aragonitic components (e.g., Wernli 1987; Rigaud et al. 2013, 2015b).

In our thin sections, gastropods, ammonites, involutinid and robertinid foraminifers show a very distinctive and atypical state of preservation. Despite their complete recrystallization to sparry calcite, these originally aragonitic fossils may display, in the vicinity of the marmorea crust, ghost relics of the primary structure of their shells (e.g., Figs. 2A, B, 3M, 4A, H). This conspicuous state of preservation is exceptional for originally aragonitic fossil components, which primary architectural and structural features are usually obliterated by diagenetic processes. It is related to a more or less pervasive impregnation by Fe-Mn fluids, which occurred prior to the test (?full) recrystallization. Unique to aragonitic tests and shells in Adnet, this preservation state permits the unequivocal identification of originally aragonitic components (Rigaud et al. 2013, 2015a).

In both involutinid and robertinid foraminifers, impregnated parts are strongly dependent on the test primary microporosity. Perforations and interstices between two lamellae or two whorls are the most impregnated parts of the tests. They appear as dark reddish-brownish thin lines under natural light (Figs. 2–4). This preservation of fine structures is uneven, localized in some specimens, prevalent in others. Intensely perforated specimens display the best preservation (Fig. 4). However, when specimens are too heavily mineralized (e.g., in the most impregnated parts of the marmorea crust), tests are too dark and structures are barely recognizable. Non-impregnated specimens are uncolored and lack structural details. Similar preservation states have been described in Cenomanian trocholinids of the Altamira and Bielba formations of Spain (Schlagintweit et al. 2015).

Thanks to the early impregnation of foraminiferal tests by Fe-Mn fluids, important structural details are indirectly preserved, allowing in-depth description. However, the original aragonitic wall of our specimens is not preserved and information concerning the primary arrangement of aragonite needles cannot be provided. In the following section, we thus assume that the wall of the examined foraminifers was originally fibrous and radially arranged, as it has been observed and defined in Robertinida (e.g., Reiss and Schneidermann 1969; di Bari and Laghi 1998).

The following classification has been developed after Cavalier-Smith (2003) and recent classifications proposed by Pawlowski et al. (2013) and Rigaud et al. (2015b).

Systematic palaeontology

Fig. 2.

Robertinid foraminifer Velleditsiella felicitaszae gen. et sp. nov., uppermost Hettangian-lowermost Sinemurian, Adnet, Austria. Holotype (B), paratypes (A, C–P). A. MHNG-75631-1, subaxial section. B. MHNG-75631-1, subaxial section. C. MHNG-75631-1, oblique centered section. D. 4a1(2), subaxial, slighty oblique section. E. 1620(2), oblique section. F. 1620(1), oblique section. G. 4a2(8), transverse section. H. 4a2(5), transverse section. I. 1620(2), transverse section. J. AU1P, transverse section. K. AU1E, transverse section. L. 4a2(6), oblique section. M. 4a1(2), oblique section. N. 4a2(8), subaxial section. O. 4a2(7), centered, slightly oblique section. P. 4a2(5), oblique section. Abbreviations: Ch, chamber; F, foramen; iH, inner hook; L, lamellae; NP, narrowest part of the chambers (open on the siphon); P, perforations; Pr, proloculus; S, siphon; SO, secondary opening.

Fig. 3.

Robertinid foraminifer Velleditsiella spinaferra gen. et sp. nov., uppermost Hettangian-lowermost Sinemurian, Adnet, Austria. Holotype (E), paratypes (A–D, F–P). A. MHNG-75631-1, subaxial section. B. 1620(2), oblique section. C. 1620(2), oblique section. D. 1620(2), subaxial section. E. MHNG-75631-2, subaxial section. F. 1620(8), subaxial section. G. MHNG-75631-2, oblique section. H. 1620(3), subaxial, slightly oblique section. I. 4a2(8), transverse, slightly oblique section. J. 1620(2), oblique section. K. MHNG-75631-2, oblique section. L. 1620(2), centered, subaxial section. M. 1620(9), oblique section. N. 1620(8), oblique section. O. 1620(3), transverse, slightly oblique section. P. 1620(2), transverse section. Abbreviations: A, aperture; Ch, chamber; F, foramen; iH, inner hook; NP, narrowest part of the chambers; P, perforations; Pr, proloculus; S, siphon; SO, secondary opening; Sp, spine.

Fig. 4.

Robertinid foraminifer Rossanella martinii gen. et sp. nov., uppermost Hettangian-lowermost Sinemurian, Adnet, Austria. Holotype (B), paratypes (A, C–M). Note the shape modifications from juvenile (e.g., A, B, F, J, L) to adult (e.g., I1, M) and senile (E) forms. A. 4a2(8), axial section. B. Holotype, MHNG-2011-1-716a, subaxial section, displaying the distinctive scroll-shaped, inner chamber structure. C, D, J. MHNG-2011-1-716b-e, axial and subaxial sections. E. AU1L, slightly oblique, centered section. F, G. 1620(2), axial and subaxial sections. H. 4a1, slightly oblique, centered section. I. AU1K, subaxial section. K. 1620(2), sub-centered, slightly oblique section. L. AU1C, axial section. M. 1620(2), oblique section. Abbreviations: A, aperture, F, foramen, L, lamellae, P, perforations, Pr, proloculus, Scr, scroll-shaped, inner chamber structure.

The phylogeny of Robertinida

Until recently, our knowledge of the evolution of the aragonitic order Robertinida was mostly founded on the pioneer works by Fuchs (1973, 1975). The family Oberhauserellidae was therefore regarded as the only survivor of the Triassic- Jurassic major extinction event and the unique ancestor of all post-Triassic Robertinida. Yet, Oberhauserellidae possess entirely evolute tests, which contrast with the rather sturdily laminar tests of post-Triassic robertinid, non-oberhauserellid lineages (superfamily Duostominoidea).

Rigaud et al. (2015b) entirely revised the taxonomy and early Triassic evolution of Robertinida, separating the oberhauserelloids and duostominoids at the superfamily level, and proposing other alternatives for the post-Triassic evolution of the group. In contrast to Oberhauserelloidea, Duostominoidea have showed a rapid diversification during the Triassic, giving rise to a large range of morpholotypes. Their evolution is notably marked by (i) the development of a planispiral coiling in Asymmetrina, Involvina, and Plagiostomella, (ii) the materialization of a siphon in Pragsoconulus and Trochosiphonia, (iii) the increasing folding of umbilical lamellae (papillose lamellae) in Cassianopapillaria and Diplotremina, and (iv) the formation of a double aperture (in Variostoma), possibly separated by a tenon in Cassianopapillaria, Diplotremina, and Pillerita. Based on the increasing complexity of the architecture in Triassic Robertinida, here considered as evolutionary acquisitions, and on account of their stratigraphic distribution, we propose here a new phyletic tree for Triassic-Early Jurassic Robertinida (Fig. 5). This tree combines observations and updates made by Rigaud et al. (2015b) with our new, Early Jurassic data. Robertinida are especially well-documented in the Carnian and, probably for that reason, show their greater generic diversity during this stage. In other Triassic stages, their record is lacunar. Part of their early evolution is most likely undisclosed. On account of their aragonitic, finely perforated and primarily bilamellar wall, high variability in chamber shape, and high trochospiral coiling, the genera Pragsoconulus, Trochosiphonia, and Robertonella are here interpreted as probable descendant of Variostoma. An alternative phyletic option would be to consider Trochosiphonia as a descendant of a Triassic “Verneuilinoides” and Pragsoconulus and Robertonella as descendants of Triassic Tetrataxidae but this polyphyletic hypothesis would only be founded on superficial morphological similarities (no structural evidence).

On account of their morphological and structural characteristics, the newly described Early Jurassic Robertinida constitute evidence that additional Triassic Robertinida lineages (not only the family Oberhauserellidae) survived the Triassic-Jurassic biotic crisis. With their distinctive conic elongate test and median siphon, Early Jurassic representatives of Velleditsiella gen et sp. nov. obviously originated from Trochosiphonia, known from Upper Triassic rocks of Tethyan and Panthalassan domains (Rigaud et al. 2015b). Rossanella gen. et sp. nov. does not have affinities with any Oberhauserellidae but shares its distinctive test shape, laminar structure, and chamber morphology with Middle- Late Triassic Praereinholdellinae. This new genus also possesses the characteristic scroll-shaped, inner-chamber structure of the post-Triassic family Conorboididae and only differs from Conorboides in its simpler apertural morphology. Therefore, Conorboides most likely originated from Rossanella, which itself originated from a Praereinholdellinae. The last known Praereinholdellinae has been documented in the lowermost Hettangian of the Northern Calcareous Alps of Austria (Hillebrandt 2010: fig. 1 as “Reinholdella sp. A” and Hillebrandt 2012: pl. 4: 1, 4, 5; pl. 6: 3a, b as “cf. Reinholdella sp. nov.”). This species shows an intermediary apertural system between Triassic Praereinholdellinae and post-Triassic Robertinina (sensu Rigaud et al. 2015b = Robertinida with an inner-chamber structure attached to the aperture), but does not possess a true internal plate yet. It is a good candidate as ancestor for the first Epistominidae sensu lato and/or Conorboididae. Hillebrandt (2010, 2012) incorrectly described an “arcus” (sensu Fuchs 1969) in this transitional species and regarded it as the Hettangian missing link between Oberhauserella and Reinholdella. The observed structure is in fact a simple thickening of the chamber roof (see Hillebrandt 2012: pl. 4: 5) and not the distinctive wall fold (arcus) characteristic of the family Oberhauserellidae (see Fuchs 1969 for details).

Fig. 5.

Triassic-Early Jurassic evolution of the order Robertinida at the genus level (partly based on Rigaud et al. 2015b, with new data). Dashed lines indicate uncertain stratigraphic distribution, dotted lines indicate uncertain phyletic relationship.

In combining published data (Fuchs 1973, 1975; Rigaud et al. 2015b) with our new observations, we propose a new phyletic tree for Mesozoic-Recent Robertinida families (Fig. 6). Interestingly, this phyletic tree shows that 60% of robertinid families survived the T/J biotic crisis, giving rise to at least two new superfamilies (Conorboididae and Epistominidae), as soon as the Hettangian. The earliest Jurassic records the first appearance of inner-chamber plates in Robertinida, which then rapidly display high morphological variabilities. Following their Early Jurassic development, the position and shape of these “toothplates” remain almost unchanged for several million years in both conorboidoid and ceratobuliminoid lineages. The function of these internal structures is still unknown but their origination is probably linked to environmental constraints, possibly related to the T/J major extinction event. As observed by Scott (1978), toothplate structures further partition the chambers and may impede the passage of cytoplasm. In view of their connection to the aperture, they may have played a role to face hostile environmental conditions, providing a better isolation from the surrounding environment.

Fig. 6.

Mesozoic-Recent phyletic tree of Robertinida, at the family level. Line width indicates the generic diversity for the different families (see bottom-right for associated number of genera). “Epistominidae” include the sub families Epistomininae, Garantellinae, and Reinholdellinae. Dashed lines indicate uncertain stratigraphic distribution, dotted lines indicate uncertain phyletic relationship.

Evolutionary implications

Molecular data have challenged traditional taxonomic and phylogenetic views on Foraminifera. The sequencing of calcitic forms has evidenced that the orders Buliminida, Rotaliida, and Globigerinida, as defined in Loeblich and Tappan (1987, 1992), are not monophyletic (e.g., Darling et al. 1997, 2009; de Vargas et al. 1997; Ujiié et al. 2008; Schweizer et al. 2008). Our understanding of the origination and long term evolution of calcitic foraminifers is limited. Molecular studies have demonstrated that, in these foraminifers, the mode of life, the acquisition of inner-chamber structures, and even the wall structure and coiling mode have a lower taxonomic value than previously widely accepted (see review in Schweizer et al. 2011). The adaptation to a planktonic mode of life occurred several times in the foraminiferal evolution (Ujiié et al. 2008; Darling et al. 2009), inner-chamber structures, such as the toothplates, can appear or disappear in the same lineage (Ujiié et al. 2008), and even when morphologically close, two species are not necessarily phylogenetically linked (see Ujiié et al. 2008; Schweizer et al. 2008). It indicates that multichambered calcitic lineages previously classified in the orders “Buliminida”, “Globigerinida”, and Rotaliida have experienced a parallel evolution.

Such an evolutionary pattern is also characteristic of the robertinid lineage. For example, several robertinid lineages developed inner-chamber structures (Duo sto minidae, Ceratobuliminoidea, and Conorboidoidea) and progressively more complex apertural system (Duo stominidae, Trochosiphoniidae, Ceratobuliminoidea, and Conorboi doi dea). As for calcitic foraminifers, distinct robertinids may have independently developed a planktonic lifestyle as both low and high conical Favusellidae (probably originated from two Triassic genera: Oberhauserella and Praegubkinella, respectively) are abundantly found in slope and basinal deposits of Middle-Upper Jurassic environments (BouDagher-Fadel et al. 1997; Wernli and Görög 2007).

In addition to the similarities observed in their evolution mode, the first known calcitic multichambered foraminifers are morphologically close to various Robertinida representatives. According to Fuchs (1973, 1975) and Tappan and Loeblich (1988), aragonitic Robertinida would be the direct ancestors of all calcitic foraminifers previously classified in the orders “Buliminida”, “Globigerinida”, and Rotaliida by Loeblich and Tappan (1987). This monophyletic hypothesis, contested by molecular data (see molecular trees in Bowser et al. 2006 and Pawlowski et al. 2013, in which Robertina branches apart from all “Rotaliida” sensu lato), also lacks strong support in fossil data. The advent and rise of the order Rotaliida sensu lato (i.e., sensu Pawlowski et al. 2013, including the former orders Buliminida, Globigerinida, and Rotaliida of Loeblich and Tappan 1992) is one of the most obscure topics. Prior to the Early Cretaceous radiation of the group, no unequivocal data exists to prove or refute the prior existence of Rotaliida sensu lato, entailing a 100 Myrs uncertainty in their time of origination (Fig. 6). Actually, all Jurassic rotaliid-alike foraminifers may in fact possess monocrystalline or aragonitic walls (see Blackmon and Todd 1959; Wernli 1971; Piuz 2004). Due to common diagenetic alterations of foraminiferal walls, phylogenetic relationships have generally been presumed, at best from rough morphological similarities or stratigraphic compatibilities. Assumed direct phyletic links between the aragonitic family Oberhauserellidae and the calcitic families Turrilinidae (as proposed by Tappan and Loeblich 1988) or Guembelitriidae (as suggested by Fuchs 1975) are not even morphologically supported. With regard to their morphologic characteristics and stratigraphic distribution, the new genera Velleditsiella and Rossanella would offer more appropriate phylogenetic links. The “toothplate” extending between the chambers of calcitic Middle Jurassic “Turrilinidae” might indeed represent the result of the evolution of the Trochosiphoniinae siphon or the Conorboididae scroll-shaped structure, which are both related to the apertural system. It is, however, unknown whether Jurassic “Turrilinidae” are true turrilinids (i.e., with hyaline-radial calcitic walls). Some Early Jurassic biserial foraminifers originally described as calcitic bolivinids (i.e., “Bolivina” liassica and “B.” rhumbleri) are for example proved (by mineralogical determination methods) to be aragonitic in composition (Blackmon and Todd 1959).

Phylogenetic relationships between aragonitic and calcitic multichambered foraminifers lack constraints, but according to our data, a paraphyly and/or polyphyly of the Mesozoic-Recent robertinids and “rotaliids” sensu lato can be envisaged. Potential direct phylogenetic links between agglutinated and calcitic forms have been disregarded. Some Cretaceous rotaliids (e.g., gavelinellids, Rotorbinella) display an inner microgranular layer, which could be reminiscent of an unknown (but direct) microgranular/agglutinated ancestor. Similarly, the monophyly of Robertinida is not fully supported by stratigraphic and morphologic evidence (dashed lines in Figs. 5, 6). A polyphyletism cannot be excluded. The acquisition of an aragonitic wall from an agglutinated or microgranular ancestor is not unique to the order Robertinida. It occurred prior to the Triassic in the Fusulinida lineage, with the origination of the aragonitic superfamily Staffelloidea (Vachard 1990; Vachard et al. 2003, 2010) and may have re-happened later through independent lineages.

Conclusions

The discovery of missing links in the Robertinida lineage has improved our understanding of the early evolution and diversity of the order Robertinida. The family Oberhauserellidae should not be considered as the only Robertinida that spanned the T/J boundary, which impact on the Robertinida diversity has been overestimated. Together with oberhauserellids, the subfamilies Trochosiphoniinae and Praereinholdellinae survived the end-Triassic major biotic crisis, and the latter group notably gave rise to the superfamilies Ceratobuliminoidea and Conorboidoidea. In these superfamilies, the development of various inner-chamber structures may have favored survival during the hostile, post-crisis Lower Jurassic period.

The reconstruction of the Robertinida evolution is of primary importance to define potential long term phylogenetic links between aragonitic, microgranular/agglutinated, and calcitic multichambered foraminifers. Up to now, only few Robertinida have been sequenced in molecular studies. These robertinids are found on a branch of multichambered Textulariida (e.g., see Bowser et al. 2006: fig. 5.3), separated from calcitic multichambered foraminifers (= Rotaliida sensu lato, see Pawlowski et al. 2013). Although molecular results on few species must be considered with great caution, this data supports a polyphyletic Mesozoic origination of calcitic and aragonitic Globothalamea. The first calcitic “rotaliid” families are morphologically close to some Jurassic Robertinida but a transition between these forms cannot be reliably supported. The original nature and structure of the wall of most Jurassic foraminifers that have been classified in the calcitic orders Rotaliida and Spirillinida remain undefined. Such forms may, as the Early Jurassic species “Bolivina” liassica and “B.” rhumbleri, have possessed aragonitic walls. A re-evaluation of the Jurassic diversity of the aragonitic order Robertinida and its potential relationships with other foraminiferal groups is a requisite.