Introduction

Representatives of the suborder Involutinina are a significant constituent of early-middle Mesozoic communities of carbonate platforms and have a strong potential as paleontological tools. Globally distributed, they have shown a rapid diversification and dispersion during the Triassic. Their distribution throughout carbonate platforms is wide, from restricted, lagoonal, shallow water deposits up to deeper, “basinal-like” environments (Piller 1978). The Involutinina are, however, usually strongly affected by diagenetic alterations and are thus only occasionally preserved in the fossil record. Their poor preservation introduces significant difficulties for their description, classification and use in paleoecologic, stratigraphic, and phylogenetic studies.

The genus Involutina Terquem, 1862, which typifies the suborder, is a major constituent of latest Triassic and Early Jurassic carbonate rocks. It is particularly abundant in “basinal-like” environments and rapidly recovered from the Triassic—Jurassic extinction event. Although a potential biostratigraphic and paleoecologic marker, it remains an unsatisfactory tool. Its specific to generic recognition, strongly dependent on the wall preservation, is intricate and details of its innermost structure are imperfectly described. Based on the thorough examination of remarkably well preserved specimens, this study aims to improve our knowledge of the morphology, phylogeny, and paleoecology of Involutina.

Institutional abbreviations.—MHNG, Museum d'Histoire Naturelle de la ville de Genève, Genève, Suisse.

Geological setting

The studied material comes from the Adnet area in the Northern Calcareous Alps (Austria). Located approximately 12 kilometers southeast 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 Schnöll Quarry (Quarry XXXI according to Kieslinger 1964). There, samples have been collected in the “marmorea-crust”, a multiphased ferromanganese layer (hardground) that forms the boundary bed between the Early Jurassic Schnöll Formation (= “Enzesfelder Kalk” in Blau and Grün 1996, 1997) and the Adnet Formation. According to Böhm et al. (1999) and Böhm (2003), the “marmorea- crust” is a guide horizon, latest Hettangian to earliest Sinemurian in age.

Fig. 1.

A. Map of Europe locating Austria (dark gray). B. Map showing the location of the Northern Calcareous Alps in the territories of Austria and Germany (grey). C. Enlargement of the area of Salzburg showing the Adnet locality. D. Location of the sampled quarry (Quarry XXXI). The quarry numbering follows that of Kieslinger (1964). E. Lithologic section of the quarry XXXI and samples location within the “marmorea-crust” (arrow).

The red limestone of the “marmorea-crust” is particularly rich in echinoderms, gastropods, ammonites, and foraminifers. In our thin sections, foraminiferal assemblages are dominated by aragonitic foraminifers, and Involutina liassica is the most common form. Like other aragonitic fossils, specimens of Involutina are completely recrystallized but locally, an early, pervasive ferromanganese impregnation of their wall-microporosity has permitted the fine preservation of their original wall architecture. This particularity allowed us to observe features generally obliterated by diagenesis.

The Involutinina classification: its origin and limitations

The suborder Involutinina Hohenegger and Piller, 1977 unites tubular foraminifers with an aragonitic wall structure. Piller (1978) and di Bari and Laghi (1994) have defined two major models for the Involutinina mode of test construction. In the “Triadodiscus model”, the laminar deposits or first order lamellae (L1 lamellae sensu Piller 1978) are discontinuous but would form by stacking one continuous laminar extension or second order lamella (L2 lamella sensu Piller 1978) per whorl. In the “Aulotortus model”, the L1 lamellae are continuous but are laterally tapered such that they form two distinct L2 lamellae per whorl (or one L2 lamella per half whorl), which are successively interfingered in the umbilical region. A third model was proposed by Piller (1978, 1983) for Involutina and Trocholina. This model is close to that proposed by di Bari and Laghi (1994) for Triadodiscus and can be considered as a variant of the “Triadodiscus model”, like the “Lamelliconus” and “Prorakusia” submodels of di Bari and Laghi (1994).

The aim of this paper is not to discuss these two major models. However, it has to be noted that the reliability of each model has not been irrefutably proved. The distinction between the two modes of test construction results from the observation of a small number of very well-preserved specimens. As inadequately oriented sections can be misleading for the observer (see Piller 1983: fig. 4), a model established only on a few specimens is unreliable. In addition, the correspondence between the “Triadodiscus model” (di Bari and Laghi 1994: pl. 4: 2) and the associated high-quality illustrations of Triadodiscus (di Bari and Laghi 1994: pl. 3: 4; pl. 4:. 1) is questionable. Finally, in specimens known to represent the “Aulotortus model” (e.g., Aulotortus, Coronipora, Frentzenella), the diagenesis often obliterates the lamellae, giving the impression that only one L2 lamella is formed per whorl. This diagenetic alteration particularly affects the L1 lamellae, which are especially minute and discontinuous in Triadodiscus.

The present high rank Involutinina systematic subdivision is partly based on these two major models (Zaninetti 1984; Zaninetti et al. 1987; Loeblich and Tappan 1987). Nevertheless, because of preservation problems, architectural models have been identified in only few forms. The suborder is typified by the genus Involutina in which the lamellae arrangement remains uncertain, entailing confusion in the lineage classification. Actually, the systematic position of the involutinins is for the most part hypothetic and phylogenetic links between different lineages are speculative. In spite of that, in the latest proposed phylogenetic tree (di Bari and Laghi 1994: fig. 7), the majority of Involutinina taxa, Involutina included, have been postulated to have originated from Triadodiscus (“Triadodiscus model”).

Structure and morphology of Involutina

In Involutina, the test architecture is unvarying. All involutins are non-septate, perforate, and possess papillose lamellae in the umbilical region. As explained by Piller (1978, 1983), papillae are originated by local elongations of the aragonite needles forming the laminae (L1 lamellae). The resulting laminar thickenings render difficult the examination of the L2 lamellae that may, according to the section orientation, appear falsely interrupted. Contrary to the previous schemes (Piller 1978, 1983), our examination of numerous centered, axial sections clearly shows that Involutina possesses at least two L2 lamellae per whorl that are, as in the “Aulotortus model”, successively interfingered in the median part of the umbilical region (Fig. 2A, B, D, E, H).

In contrast to the test architecture, the morphology of elements constituting the test of involutins varies considerably. In Involutina liassica (Jones in Brodie, 1853), the type-species of Involutina, a large range of variability in the test size and shape, the tubular chamber morphology and its position related to the previous whorl, the lamellae thickness, the papillae size, number, and repartition, and the perforations size and their connection exists (Fig. 2A–G). Dimorphism is well-pronounced (see Schweighauser 1951: figs. 5–8). Megalospheric forms display a larger proloculus and a lower number of coils than microspheric forms but, as for other involutinids (Koehn-Zaninetti 1969), always show smaller tests. Within the same morphotype, the test may be more than twice as big or as thick, depending on the tubular chamber height and the laminar deposits thickness. The mode of coiling, generally planispiral, may also show some irregularities or oscillations (e.g., Schlagintweit and Piller 1990: pl. 1: 1, 14). The papillose lamellae may be more or less pronounced, forming more or less prominent and numerous papillae on the test surface. The perforations are highly variable in diameter, randomly distributed, and possibly form large canals that may merge (Fig. 2A, C, E, F). Lastly, along the ontogeny of the same specimen, the tubular chamber morphology may change significantly due to the optional development of a tube floor (Fig. 2A–F). The latter observation contests the model proposed by Piller (1978, 1983) for the mode of construction of the tubular chamber.

Involutina model

The current models of test construction of Involutina (Koehn-Zaninetti 1969; Piller 1978; Blau 1987b) do not integrate the whole complexity of the form. They are based on specimens in which the aragonite needles, laminar deposits (L1 lamellae), and lateral laminar extensions of the tube wall (L2 lamellae) are only partially preserved. In our material, only relics of the aragonite needles and the laminar deposits are preserved but in some specimens, the outline of the laminar extensions is entirely emphasized by the ferromanganese impregnation. Based on a detailed examination of Involutina liassica, we herein propose a new model for the Involutina structure (Fig. 3). This structural model derives from the study of several randomly oriented sections and does not correspond to a single specimen. Its difference from previous models is largely founded on the arrangement of the laminar extensions that are interfingered in the umbilical region (as in the “Aulotortus model”). Our model clearly contrasts with Piller's model (1978) in which only one lamella is formed per whorl and contests the latest Involutinina phylogenetic tree proposed by di Bari and Laghi (1994).

Taxonomy of Involutina

The genus Involutina was introduced by Terquem (1862: 450), prior to the advent of the International Rules for Zoological Nomenclature (IRZN). Terquem (1862) described two species within the genus: Involutina silicea Terquem, 1862 and Involutina jonesi Terquem and Piette in Terquem, 1862 (Terquem 1862: 450–451 and 461, respectively), but did not mention which one was the type-species for the genus. Some authors have considered I. silicea as the type-species (e.g., Loeblich and Tappan 1954). However, in an earlier work, Bornemann (1874) recognized differences in the wall composition of the two species and, by placing I. silicea into Ammodiscus Reuss, 1862, he only retained I. jonesi in Involutina. Because, as stated by Brady (1864), I. jonesi is a junior synonym of I. liassica (Jones in Brodie, 1853), Nummulites liassicus Jones in Brodie, 1853 (= now Involutina liassica) must be regarded as the type-species of Involutina (Brady 1864; Bornemann 1874; Wicher 1952; Kristan 1957; Koehn-Zaninetti 1969; Gušić 1975; Piller 1978; Loeblich and Tappan 1987).

It is generally believed that the genus Involutina is characterized by a high interspecific variability and numerous species have been introduced into the genus based on differences observed in the test and tubular chamber morphology. Our study, however, emphasizes that these traits display a large range of variability at the specific level, calling into question the validity of these species. Very few forms can be convincingly separated from I. liassica even with a thorough statistical analysis. For example, there is no reliable criterion that permits the distinction between I. farinacciae Brönnimann and Koehn-Zaninetti, 1969 and I. liassica. Likewise, in our material, along the ontogeny, the tubular chamber section of I. liassica appears oval to triangular (e.g., Fig. 2F) and is formed either by a complete or a semi tube (see Piller 1978 for definition) (Fig. 2A–F). Therefore, the validity of I. turgida Kristan, 1957 based either on the interpretation of Kristan (1957) or Piller (1978) is dubious (see also Gušić 1975).

In the literature, the diversity of Involutina had already been revised downwards. Some species (I. silicea Terquem, 1862, I. aspera Terquem, 1864, I. polymorpha Terquem, 1864, and I. limitata Terquem, 1864) have been excluded from Involutina on account of the agglutinated nature of their wall (Bornemann 1874) and the Early Cretaceous species Involutina stinemeyeri Church, 1968 as it would not possess the morphological characteristics of the genus (Brönnimann and Koehn-Zaninetti 1969). Among the remaining species, most have been placed into synonymy with the type-species Involutina liassica (Jones in Brodie, 1853) (see Bornemann 1874; Wicher 1952; Kristan 1957; Koehn-Zaninetti 1969; Gušić 1975; Piller 1978). For example, the species Involutina ticinensis (Schweighauser, 1951) is regarded as the microspheric form of I. liassica (Kristan 1957; Koehn-Zaninetti 1969).

Systematic palaeontology

The classification here used has been developed after Cavalier-Smith (2003) for the subphylum and Zaninetti et al. (1987) for the suborder, superfamily, family, and subfamily. The class and order ranks of the Involutinina are still a matter of debate and will not be discussed in this manuscript.

Fig. 2.

Ornamented involutinins. A–G. Involutina liassica (Jones in Brodie, 1853), Early Jurassic, Adnet area, Austria, collection MHNG-75631, thin sections 402, 4a1, 4a2. Specimens with a structure partly preserved from diagenesis by a ferromanganese impregnation. Note the large intraspecific morphological variability. A. MHNG-75631-40a; A₁, axial section, specimen with merged canals and relatively well-developed tube floors. Note the lamellae interfingering; A₂, enlarged view of A₁. B. MHNG-75631-40b. Transverse, sub-axial section of a specimen without tube floor. C. MHNG-75631-40c. Axial section of a specimen with merged canals. D. MHNG-75631-42a. Axial section. Note relics of the fibrous wall structure (white arrow) and the lamellae interfingering (black arrow). E. MHNG-75631-42b. Axial section of a specimen presenting a tube floor in its juvenile part only. Arrows point interfingering lamellae. F. MHNG-75631-41a. Axial section of a specimen with thick umbilical masses and well-developed papillose lamellae. Note that papillae are the expression of local lamellae thickenings. G. MHNG-75631-41b. Transverse, sub-equatorial section showing transversely sectioned papillae. Note the folding of the tube wall, possibly delimiting rudimentary eggholders-like structures. H, I. Involutina sp. (Early Jurassic, Adnet area, collection MHNG-75631, thin sections 4a2, 402). Discoidal specimens partly preserved from diagenesis by a ferromanganese impregnation. H. MHNG- 75631-42c. Transverse section. Note the presence of interfingered papillose lamellae (arrow). I. MHNG-75631-40d. Axial section. Straight canals perforate the umbilical masses. J. Triadodiscus inceptus di Bari and Laghi 1994 (Carnian, Late Triassic, Italy, di Bari and Laghi material, SEM images), MHNG-2011-1-9; J₁, isolated specimen with distinct bumps; J₂, enlarged view of J₁; J₃, sectioned, polished, and etched specimen; J₄, enlarged view of J₃ showing significant laminae discontinuities; J₅, enlarged view of J₄ showing detail of a bump. Note the limited lateral extension of the bump and the good preservation of the aragonite needles. Abbreviations: e-h, rudimentary egg-holder; i, interfingered (lamellae); m.c., merged canals; p, perforation; pap, papilla; p.l., papillose lamella; s.c., straight canal; t.f., tube floor. Scale bars 50 μm.

Fig. 3.

Structural model of Involutina (in axial section). The presence of a semitube, as defined by Piller (1978), is only barely discernible in few of our specimens.

Phylogeny

There are conflicting theories as far as the position of Involutina in the involutinins lineage is concerned. Involutina has been either considered to directly derive from trochoslaminar piral forms (i.e., Lamelliconus) or planispiral forms (i.e., Aulotortus or Triadodiscus). The hypothesis that Lamelliconus is the closest ancestor of Involutina has been proposed by Kristan-Tollmann (1963). However, recent studies have demonstrated that trochospirally coiled involutinins (Trocholinidae) are a separate group (Rigaud et al. 2013a), refuting Kristan-Tollmann's (1963) statement. The possible origination of Involutina from Aulotortus as proposed by Koehn-Zaninetti (1969), Gušić (1975), and Salaj et al. (1983) or from Triadodiscus as postulated by Piller (1978), Gaździcki (1983), and di Bari and Laghi (1994) is trickier to prove. Aulotortus and Triadodiscus only differ in their laminar deposits architecture (more discontinuous in Triadodiscus) and possibly in their laminar extensions arrangement (interfingered in Aulotortus). These characteristics, strongly dependent on the test preservation, have been only partially documented in Involutina. Our study emphasizes that the laminar extensions of Involutina liassica are interfingered in the median part of the umbilical region, as in the “Aulotortus model”. This observation questions the potential phylogenetic link between Triadodiscus and Involutina and demonstrates that, contrary to Piller's (1983) opinion, Aulotortus and Involutina do not show a distinct difference in their mode of test formation.

The assumption that Involutina derives from Triadodiscus is mainly based on the observation of bumps (= “bosses” sensu di Bari and Laghi 1994) on the test surface of Triadodiscus inceptus di Bari and Laghi, 1994. In isolated specimens, these features are actually very close to the involutins papillae. This resemblance led di Bari and Laghi (1994) to regard the species as the missing link between the two genera. Our examination of numerous specimens of Triadodiscus inceptus from the collection of di Bari and Laghi (1994), however, allows to incontestably distinguish bumps and papillae, refuting a possible link between the two structures. While papillae are the external expression of local L2 lamellae thickenings (e.g., Fig. 2F; see also Piller 1983), bumps are the external expression of the laminae discontinuity characterizing the wall architecture of Triadodiscus (Fig. 2J).

On the other hand, the specimens illustrated by He and Yue (1987: “Aulotortus columnaris”, pl. 5: 21–23) represent a probable missing link between Aulotortus and Involutina. These forms, though devoid of papillose lamellae, display large pores or canals that are comparable in size to those of Involutina liassica. It is noteworthy that the syntypes of Aulotortus columnaris (He 1982: pl. 4: 1–4), however, do not show canals in their umbilical masses (only fine perforations). Hence, the specimens illustrated by He and Yue (1987) cannot be included in A. columnaris He, 1982.

Accordingly, we postulate that Aulotortus is the direct ancestor of Involutina. Their test morphology (discoidal to almost globular), type of coiling (planispiral with possible irregularities or oscillations), laminar arrangement (interfingering), and tubular chamber morphology (oval to crescent- shaped in section and with or without a tube floor) are similar. Additionally, in Aulotortus, it is the thickness of the laminar extensions and not their length that predominantly controls the size of the test. The thinner the laminar extensions are, the more discoidal is the test (e.g., in Aulotortus tenuis). Finally, the high morphological variability observed in Involutina strongly reminds that of Aulotortus representatives.

Stratigraphic and evolutionary Implications

Late Triassic and Early Jurassic carbonate platforms lack diagnostic biostratigraphic markers, particularly in shallow-water deposits. Aulotortus is widely used in biostratigraphic studies and Involutina, rich in Tethyan “basinal-like” deposits, might be useful to calibrate platform deposits with deeper environments. However, the genera Aulotortus and Involutina are characterized by significant morphological unevenness, entailing difficulties in their specific recognition.

In Aulotortus and Involutina, species have mainly been established on the basis of the test shape, the laminar development, the mode of coiling, or the tubular chamber morphology. While consistent with studies on other foraminiferal groups, these criteria, single-handedly, have proved to be unreliable for the involutinid taxonomy. For instance, Aulotortus sinuosus, Involutina liassica (e.g., in Böhm et al. 1999: pl. 15: 17), and Involutina hungarica may show either a planispiral or oscillating coiling. In addition, misinterpretation of the wall structure in recrystallized specimens can lead to the establishment of artificial species. For example, the species “Aulotortus tumidus” has been falsely interpreted to possess a last “evolute” whorl (see Piller 1978: pl. 6). Similarly, in Involutina liassica, as the papillose lamellae usually follow the previous whorl outline, the last whorls may appear misleadingly evolute.

A whole revision of the species classified in these two genera is required. In both genera, specific determination should be based on steady, reliable criteria or on the combination of several unstable characters. According to our observations, only few forms show steady patterns but the combination of criteria such as the lamellae thickness, the shape of the test, and the perforation size has proved to be useful for specific identification. The recognition of a direct phylogenetic link between Aulotortus and Involutina may also facilitate the distinction between factual and artificial species. From Aulotortus to Involutina, only one morphological acquisition is required (the development of papillose lamellae) and it is likely to suppose that involutins kept the characteristics of their ancestors. Comparisons between the two genera at the specific level have emphasized strong morphological resemblances between the specimens illustrated by He and Yue (1987: pl 5: 21–23) as“Aulotortus columnaris” and Involutina liassica, and between Aulotortus minutus and the few discoidal specimens found in our material (Fig. 2H, I). Hence, as already stated by Koehn-Zaninetti (1969: fig. 21), it is probable that Aulotortus representatives have shown a parallel evolution leading to the appearance of several involutins. As a consequence of this parallel evolution, it is now possible to confirm that in both genera neither irregularities or little oscillations in the mode of coiling nor slight modifications of the tubular chamber morphology are consistent for specific differentiation and establishment.

Pending further studies on the taxonomy of the two groups, we would advise a cautious use of Aulotortus and Involutina in biostratigraphic studies.

Hypothesis on functional morphology and related discussion

The emergence of Involutina in the Late Triassic has recorded the acquisition of papillose lamellae in planispirally coiled Involutinina. At this time, papillae are also documented in both tubular (i.e., Trocholina, Semiinvoluta, Frentzenella) and multichambered (i.e., Cassianopapillaria) trochospirally coiled aragonitic forms. In Involutina, a strong correlation between the shape of the test and the papillae size is observed. As previously mentioned, the shape of the test is controlled by the laminar deposits thickness (subglobular forms show protuberant umbilical masses with thick laminar deposits whereas discoidal forms display thin laminar deposits). In the Adnet material, the papillae size of Involutina liassica varies considerably from specimen to specimen, even in the same thin section. The thicker the test is, the more prominent the papillae are (Fig. 2A–F). This observation has already been mentioned by Moullade and Peybernès (1974) and Schlagintweit and Piller (1990) for Involutina hungarica. The same correlation is observable from species to species. Discoidal forms show less prominent papillae than lenticular forms (Fig. 2A–F, H, I). It is noteworthy that the number of papillae is also reduced in discoidal forms. Moreover, in discoidal forms, the papillae repartition is not arbitrary (e.g., see Kristan 1957: pl. 22: 2). The papillae form a spiral on the test surface that follows the tubular chamber progression. This peculiar repartition (straight above the tube), the convergent evolution (appearance of papillae) in aragonitic forms at the end of the Triassic, and the correlation existing between the test shape/ thickness and the papillae size/quantity strongly suggests that papillose lamellae have a biological role.

The biological role of papillae in the Involutinina has been discussed by Piller (1978). As Involutina is usually found in deeper deposits than Aulotortus, Piller (1978) concluded that papillae are a probable reinforcement of the test structure related to higher hydrostatic pressure. Although it is widely thought that Involutina is indicative of slope to basinal depositional settings (Piller 1978), no occurrence of the genus in deep basinal settings (e.g., co-occurring with thinshelled bivalves and/or radiolarians limestones) have been mentioned. Involutina has been mostly found associated with an abundant and diversified fauna, in deposits showing a high affinity with the platform, suggesting an occurrence in the photic area. In the Late Triassic, Involutina occurs together with sponges, gastropods, ammonites, echinoderms, ostracods, crustaceans, bryozoans, brachiopods, and other foraminifers (Kristan 1957). In Early Jurassic deposits, Involutina is particularly abundant on the top of drowned Triassic carbonate platforms and is also found associated with abundant and diversified bioclasts. In the Adnet area, for instance, Involutina liassica is found together with echinoderms, ostracods, oncoids, other foraminifers, gastropods, bivalves, holothurians, crustacean debris, serpulids, ammonites, brachiopods, and globochaetes (Böhm et al. 1999 and personal observations). The oncoids and possibly the globochaetes (Skompski 1982) attest to the presence of Involutina liassica in the photic area. Furthermore, most involutinins associated with Involutina show a relatively thin wall on one or both sides of the test (e.g., Trocholina, Coronaserra, Coronipora, Kristantollmanna), attesting that their test does not necessitate any reinforcement to face the local hydrostatic pressure, as expected for foraminifers.

Conversely, the lamellae thinning may document a slowdown of the biomineralization process with depth increase (Hottinger 1997) and the presence of papillae, may help to maximize the amount of light penetrating the wall (Hottinger 2006). The assumption that papillose lamellae express a paleoecological adaptation to deeper environments for light catching is consistent with our data (for the same depth, thicker tests require a more prominent relief) and supports the hypothesis that the Involutinina may have held symbiont- bearing representatives (Rigaud et al. 2013b).

Accordingly, in Involutina, the “plaits or infoldings of the outer shell” interpreted as “septa” by Brady (1864) can be reinterpreted. These features, named “demi-cloisons” by Terquem (1862), are well-visible in equatorial section. They form small compartments on the tubular chamber periphery (Fig. 2G). We propose that they have served as rudimentary eggholders (see Hottinger 1977), adaptative features for symbiont positioning immediately beneath the test wall.

Concluding remarks

Prior to this study, the structure of Involutina was inadequately defined. The recognition of a direct phylogenetic link between Aulotortus and Involutina has highlighted inaccuracies in the classification, phylogeny, and possibly the stratigraphic range of planispirally coiled involutinids. An entire revision of the group is required. Pending this revision, a cautious use of Aulotortus and Involutina in biostratigraphic studies is recommended.

The evolutive acquisition of papillose lamellae is most likely a clue that Involutina had a symbiotic life. Such a mode of life implies that the involutinins can be used as depth bioindicators within the photic area. In papillose Involutinina, the papillae (number and size) might thus become relevant paleobathymetric tools.

The earliest Jurassic has been characterized by a global sea-level rise. In Involutina, the acquisition of evolutive features such as papillae and rudimentary egg-holders was most probably a significant advantage over Aulotortus to face the Triassic—Jurassic extinction event.