Abstract

A new quantitative biostratigraphical method, based on an ‘index of stratigraphical abundance’ (ISA), has been used to refine the stratigraphical range of 50 dinoflagellate cyst taxa. Most of these correspond to Quaternary representatives of Achomosphaera or Spiniferites. Some extinct taxa are also discussed when their morphology is close to that of recent species (e.g. Spiniferites lenzii, S. twistringiensis and S. pseudofurcatus), as well as some species possibly related to the motile Gonyaulax spinifera complex. The ISAs of S. mirabilis and S. hyperacanthus are similar, as well as for S. elongatus and S. sphaericum. Although initially presumed as indicating morphotypes of a same taxon, it seems rather that the ISAs similarites must be interpreted as an indication of close ecological preferences. The emergence of forms of Spiniferites with low to absent septa appears relatively late (compared to the range of the genus), but the first occurrence of intergonal processes is relatively early, indicating that it is probably a character deeply rooted into the generic genome. The previously calculated phylogeny of some Gonyaulax motile cells shows a different tree pattern than the stratigraphical first appearance succession of their presumed cyst equivalents.

Introduction

Spiniferites Mantell 1850 is one of the most geographically widespread and stratigraphically long-ranging dinoflagellate cyst genus (Hultberg and Malmgren 1995). This finding holds true both in the ancient (post Jurassic) and recent (Pliocene-Quaternary) series as shown by the numerous records of this genus, especially for the type species, Spiniferites ramosus (Ehrenberg 1837) Mantell 1854. Indeed, this taxon presents an exceptional case of stratigraphic longevity, as it is reported in the fossil record from the Upper Jurassic to Present, with a thriving distribution. With a stratigraphical longevity exceeding 140 Ma, Spiniferites ramosus is probably one of the longest living among the Quaternary dinoflagellate species, and perhaps among all species in the living realm.

The genus Spiniferites encompasses more than 100 species (Fensome and Williams 2004) among which 20 have at least been recorded in Quaternary strata. In connection with the “Quaternary Spiniferites Workshops” held in Montreal (2014) and Ostend (2015) (see the general Introduction of this volume), we herein focus primarily on Quaternary and Neogene species. Stratigraphical range charts were published for some Spiniferites species (e.g. Harland 1978; Sarjeant 1979), and the one of 55 Spiniferites species by Hultberg and Malmgren (1995) remains a reference work, but over the last twenty years, new data were published and some new species have been erected. In addition, representations of species occurrences by a simple line between inception and extinction provide limited information.

To go beyond this method based on the only alternative presence/absence (i.e. occurrence), a new approach is here proposed based on an ‘index of stratigraphical abundance’ (ISA) calculated for various Spiniferites species. It is an occurrence probability of a taxon for each geological stage, which is the geochronological resolution selected here. On one hand, this allows to display the time interval in which a taxon is the most abundant relatively to other dinoflagellate cyst species, and secondly to compare the ISA’s variations between different taxa.

The motile equivalent of Spiniferites ramosus is not clear and might be attributed to Gonyaulax spinifera-type (Claparède and Lachmann 1857) Diesing (1866) motile cells (e.g. Wall and Dale 1970; Lewis et al. 1999; Ellegaard et al. 2003; Rochon et al. 2009). Other types of cysts have been attributed to that motile complex, sometimes not belonging to the genus Spiniferites (e.g. Bitectatodinium tepikiense, Nematosphaeropsis labyrinthus, Tectatodinium pellitum; cf. ‘Heterospory’ in Head 1996; Lewis et al. 2001).

The present work is an opportunity to test whether dinoflagellate cyst taxa, which have been attributed to the motile Gonyaulax spinifera complex (Dodge 1989), present similarities in their 'index of stratigraphical abundance'. The study should also clarify the stratigraphical origin of the Spiniferites species still present in the modern seas.

Material and methods

The databases

In order to collect the occurrences of the selected taxa in the most exhaustive way possible, a very large number of records is needed. The PALYNODATA online database (Palynodata Inc. and White 2008) offers such an opportunity by providing a large number of species and references that list the geographical and stratigraphical local occurrences of each taxon. For the present work, a total of 2,279 publications from the organic-walled dinoflagellate cyst literature depicting 2,333 geological sequences (field sections or boreholes) was extracted from the PALYNODATA database. As the database has not been updated since 2006, it was completed with a personal database that contains data before and after 2006 for well-dated sequences, preferably with a dating not based on the dinoflagellate cysts themselves. It indexes occurrences of 1,585 organic-walled dinoflagellate cyst species from 214 additional publications (396 geological sequences) of global dinoflagellate cyst literature.

The merging of the two databases lead to 2,729 worldwide geological sequences from the Carnian (Upper Triassic) to today. Among them, 1,671 geological sequences provided species of Spiniferites or of the Gonyaulax spinifera complex. No stratigraphical synthesis (e.g. Sarjeant 1979; Powell 1992; Williams et al. 1993; Hultberg and Malmgren 1995) was integrated in the merged database, in order to not duplicate the data already included.

The final database therefore has two dimensions, with: the 46 geological stages, from the Carnian to the Holocene, in columns, and the taxa in rows. When a paper deals with several sites, each site was considered independently.

The occurrence of the taxon is 0 if it was not observed in the sequence and 1 if it has been recorded. The value of 0.5 is sometimes used when the dating is uncertain or imprecise (e.g. Senonian or Lower Miocene). The species referred as 'cf.' were not taken into account.

The index of stratigraphical abundance (ISA)

An ‘index of stratigraphical abundance’ (ISA) has been developed to get more information from the biostratigraphical records, allowing to go beyond the establishment of the stratigraphical range of a taxon by determining its oldest and youngest occurrences.

The ISA allows to not only to determine the stratigraphical range of a taxon as indicated by its earliest and latest occurrences but also to obtain a full occurrence probability distribution.

The ISA was calculated on the basis of the number of geological sequences in which a taxon was recorded during a time interval (= ‘nOcc’ for number of occurrences) versus the total number of sequences documented in the database (= ‘nSeq’ for total number of geological sequences) for the same time interval:

ISA = nOcc/nSeq

For each taxon, an ISA was calculated for each geological stage considering its worldwide occurrences. If necessary, the geographical area of the record can be restricted to an oceanic basin, to a (paleo)hemisphere, etc. Overall, the ISA can be considered as an estimate for the occurrence probability of a taxon during each geological stage. For example, 157 sequences of the database span the Holocene. Among these sequences, 66 record Spiniferites ramosus. The ISA of Spiniferites ramosus for the Holocene is 66/157 = 0.42. That means that in a global stratigraphical point of view, the probability to encounter a specimen of this species in a Holocene sequence is about 42%, implying that an ISA value cannot exceed 1. The database used here includes 2,729 different geological sequences and 1,585 organic-walled dinoflagellate cyst species. To calculate ISA, it is first needed to determine the number of sequences documented in the database for each geological stage (the time interval chosen here). Since the database from which it is calculated is substantial, featuring on average over a hundred publications for each geological stage (from 11 sequences for the Carnian to 308 for the Albian, cf. Table 1), the ISA constitutes a significant estimator.

Results

The stratigraphical occurrences of 50 taxa have been investigated. Most of them correspond to Spiniferites or Achomosphaera species recorded in Quaternary strata. Some extinct taxa are also depicted when their morphology is close to that of recent species (e.g. Spiniferites lenzii, Spiniferites twistringiensis and Spiniferites pseudofurcatus), as well as some species possibly related to the motile Gonyaulax spinifera complex.

Spiniferites ramosus, oldest occurrence of the genus?

Duxbury (1977), Davey (1979, 1982), Williams and Bujak (1985), Costa and Davey (1992), Stover et al. (1996) consider that, at least in NW Europe, the first appearance of Spiniferites ramosus marks the lowermost Valanginian. This is also what Hardenbol et al. (1998) followed by Ogg and Ogg (2008c) report for the Boreal/North Atlantic domain but, according to the lowermost Berriasian occurrences recorded by Monteil (1992, 1993), they consider an earlier occurrence in the Tethyan domain. That point of view is still confirmed by many experienced palynostratigraphers (James Riding, personnal communication).

Such a statement deserves a closer scrutiny since several Spiniferites ramosus occurrences were recorded in Jurassic strata. Deflandre (1938) mentioned a single specimen attributed to Hystrichosphaera furcata (now Spiniferites ramosus) from the Lower Oxfordian cliffs of Villers-sur-Mer (France). The description given by the author was very succinct, with no mention of trifid processes. Due to its poor preservation, the specimen was not illustrated. Such a record seems not suitable for biostratigraphy. Sarjeant (1960) reported one specimen of Hystrichosphaera furcata that was illustrated (pl. 14, fig. 17) from the ‘Upper Calcareous Grit’ Formation from Yorkshire (England). The age of this formation is considered as late Oxfordian (op. cit.). Still from Yorkshire in the 'Ampthill Clay' Formation, Sarjeant (1962) reported two specimens (not depicted) of Hystrichosphaera furcata from an Oxfordian pit exposure. It is of note that the collective efforts of lots of reliable palynologists over many years have never replicated these occurrences in NW Europa. Gitmez and Ertug (1999) studied five Upper Jurassic sections from Turkey and mentioned few sparce specimens of Spiniferites ramosus in only one of them. The illustrated specimen (pl. 9, fig. 4-5) does not look like Spiniferites ramosus sensu holotype. In addition, this specimen appears very fresh and could be a contamination from the Neogene upper part of the outcrop. In their synthesis, Pestchevitskaya et al. (2011) recorded Spiniferites ramosus from a little below the Jurassic/Cretaceous (J/K) boundary (late Upper Volgian) in outcrops from Sub-Polar Ural, and from the J/K boundary near Laptev Sea. Unfortunately no specimen was depicted. Other specimens have been reported from Jurassic strata, but in boreholes for which the possibility of caving cannot be excluded (e.g. Brideaux 1977, Upper Jurassic of NW Canada; Chen 1978 early-middle Kimmeridgian and Portlandian from Madagascar; Thusu and Vigran 1985 unspecified Jurassic age from Libya, considered by the authors as caved; Guy-Ohlson 1986, Toarcien-Aalenian of Sweden). These borehole occurrences have not been considered in this work.

These records show that there are no unequivocal reports of Spiniferites ramosus from the Jurassic supported by photographic evidence. Cretaceous occurrences (particularly Neocomian one) are numerous enough to be considered as significant.

Berriasian (lowermost Cretaceous) records of Spiniferites ramosus show a scaterred distribution since it was reported in various outcrops from Iran (Ashraf 1979), India (Kumar 1980), Gulf of Mexico (Riley and Fenton 1984), Libya (Thusu et al. 1989), SE France (Monteil 1992, 1993), NW Siberia (Lebedeva and Nikitenko 1999).

If we consider the bibliographical data as reliable, particularly those from outcrops, it appears that the geographical extension of Spiniferites ramosus appears restricted during the Berriasian, and almost worldwide during the Valangian.

Thus, we can consider that the first occurrence datum (FOD) of Spiniferites ramosus can be placed in the lowermost Berriasian and the Valanginian worldwide extension could corresponds to its first common occurrence (FCO). After that, the species presents consistent ISA values above ca 0.2 from the Valanginian until present days (Figure 1). Two ISA optima are clearly distinct around Maastrichtian and Quaternary times with an interposed minimum centered on the Middle Miocene (Figure 1).

Davey (1979) defined Spiniferites ramosus as a biostratigraphical marker of the ‘Spiniferites ramosus’ dinozone (Davey 1979) covering the mid-Early Valanginian to the top Valanginian in the Boreal realm. That marker could correspond to the FCO of the species.

Several other species of Spiniferites first appear during the Berriasian (e.g. Spiniferites alatus Duxbury 1977, Spiniferitesmultibrevis” now Spiniferites twistringiensis (Maier 1959) Fensome et al. 1990) which makes uncertain whether Spiniferites ramosus would be the oldest representative of the genus Spiniferites. Achomosphaera who is a genus morphologically very close to Spiniferites might have been the first of the two to appear since Achomosphaera neptunii (Eisenack 1958) Davey and Williams (1966) was recorded as early as the Tithonian (uppermost Jurassic; Harding et al. 2011).

Neogene and Quaternary Spiniferites plexus

The first and last occurrences of the selected species are presented in Table 2. Some species deserve further attention and are discussed below. The ISAs of the Quaternary and Pliocene species that first appear before the Cretaceous/Paleogene boundary are shown in Figure 1. Those which inception is post-Cretaceous are shown in Figure 2.

Both Achomosphaera andalousiensis Jan du Chêne (1977) and Spiniferites septentrionalis Harland (1977) possess processes with fenestrate/trabeculate distal tips and the distinction is not always easy (see Londeix et al. 2018). Some authors consider these species as synonym, and it is difficult, in the absence of clear illustrations, to know which of these two species has been recorded. It is probable that the compiled record of each of these species does not refer to a unique taxon. Nevertheless, the FAD of Achomosphaera andalousiensis seems to be upper Langhian (Powell 1986b; McCarthy and Mudie 1996, see also Dybkjær and Piasecki 2010) while the first appearance of Spiniferites septentrionalis is upper Tortonian-Messinian (Harland 1979). Spiniferate cysts with trabeculate/fenestrate distal tips are still present today. However, it is difficult to assign them to one or the other of these taxa since they present intermediate features (e.g. Morzadec-Kerfourn 1979; Mangin 2002). Nevertheless, Spiniferites septentrionalis is present in Upper Quaternary (Harland 1977) sediments and typical Achomosphaera andalousiensis have been recorded and illustrated by Head (1997) from Piacenzian strata. Achomosphaera andalousiensis gave its name to a dinozone established by Piasecki (1980) that covers almost the entire upper Miocene (e.g. Piasecki 1980; Manum et al. 1989; McCarthy et al. 2013).

Remarks: When Harland (1977) erected Spiniferites septentrionalis, he mentioned Philip C. Reid according who (personnal communication) the new species was synonymous with Spiniferites ramuliferus as recorded by Reid (1974). In that work Reid considered the specimen figured by Downie and Singh (1970) as Hystrichosphaera ramosa (Figure 3) as synonymous of Spiniferites ramuliferus. Neither Reid nor Downie and Singh evoked the presence of perforated or trabeculate process terminations. Synonymy of Spiniferites septentrionalis with Spiniferites ramuliferus sensu Reid (1974) (non Deflandre) is not followed here, however Spiniferites ramuliferus sensu Reid (1974) and Hystrichosphaera ramosa sensu Downie and Singh (1970) (non Ehrenberg) seem to correspond to the same taxon.

The first formal occurrence of Achomosphaera callosa Matsuoka (1983) is Londeix and Lopes (2014) from Aquitanian stratotype, but it is likely that this species appeared earlier in the Chattian (Powell 1986b as Achomosphaera cf. crassipellis) or in the upper Rupelian (Stampian, Londeix, unpub. data).

Achomosphaera ramosasimilis (Yun 1981) Londeix et al. 1999 was firstly recorded in the Santonian from Germany but the synonymy provided by the original author ranges from the Neocomian (as Hystrichosphaeridium ramuliferum, Gocht 1959) to the Upper Oligocene (as Achomosphaera ramulifera, Cookson and Eisenack 1974). Londeix (1990) recorded this species in the Upper Valanginian from SE France and it was also recorded in the Gelasian from Sicily (Londeix et al. 1999). The youngest occurrence might be in Lower Pleistocene from Labrador Sea as Achomosphaera ramulifera (de Vernal and Mudie 1989).

Remarks: In the note accompanying the original description of Achomosphaera ramosasimilis Yun (1981) states that the shape and arrangement of the processes and the structure of the central body are equivalent to Spiniferites ramosus ramosus. This taxon is here considered as a concept that coresponds to Spiniferites ramosus without septa between the processes.

The earliest suitable records of Spiniferites bentorii (Rossignol 1964) Wall and Dale (1970) seem to be from Aquitanian by Edwards (1986) and Powell (1986b) respectively from South Carolina and NW Italy. The Paleogene records of Spiniferites bentorii (e.g. Frederiksen 1969; Bujak 1973; Islam 1984; Powell et al. 1996; Gedl 2005) are herein considered as not suitable since sparse, not illustrated, or with questionable identification.

The early record of Spiniferites bentorii subsp. truncata (Rossignol 1964) Wall and Dale (1970) in the Upper Cretaceous from Western Xinjiang (China) by Yu and Zhang (1980) appears very isolated compared to the other records for this taxon. This record is not taken into account here and the stratigraphical range of Spiniferites bentorii truncata is considered to be limited to its restricted and rare occurrences int the Quaternary.

The lowest occurrence of Spiniferites bulloideus Deflandre and Cookson 1955 might be Ypresian (Auffret and Graus-Cavagnetto 1975) but this record appears discrepant since the other oldest occurrences are significantly younger: Chattian (Londeix, unpub. data) and Aquitanian (Londeix and Lopes 2014).

Spiniferites cruciformis Wall and Dale in Wall et al. 1973 is endemic to the Black Sea, Marmara Sea, Caspian Sea, Aral Sea, and eastern Mediterranean (Zonneveld et al. 2013) and occurs mainly in brackish environments. Applying biostratigraphy in a such context is generaly not easy, and in the present case the oldest occurrences of this species are poorly dated: hypothetically within the range of Pliocene to Pleistocene (Eaton 1996) or Middle Pleistocene (Koreneva and Kartashova 1978). The undisputed oldest occurrences were recorded in the upper Pleistocene (Wall et al. 1973; Mudie et al. 2002, 2007; Londeix et al. 2009). Williams et al. (1998) consider the stratigraphical range of Spiniferites cruciformis from 0.17 to 0.01 Ma. The species might be not extinct since it is present in some surface sediments of the ‘paratethysian’ area (Zonneveld et al. 2013).

The earliest clear occurrence of Spiniferites elongatus Reid 1974 seems to be from Tortonian from Northern Japan (Matsuoka 1983; Matsuoka et al. 1987). Closely related morphologies to Spiniferites elongatus were sporadically found in Rupelian strata in a section from Belgium (De Coninck 2001) and from a borehole from the Norwegian-Greenland Sea (Poulsen et al. 1996). All these records are earlier than the Pleistocene FAD stated by Williams et al. (1998) and Ogg and Ogg (2008a) and are in agreement with the Late Miocene occurrence stated by Bujak and Matsuoka 1986 and Stover et al. (1996).

The stratigraphy of Spiniferites hyperacanthus (Deflandre and Cookson 1955) Cookson and Eisenack 1974 started in the Lower Cretaceous (Figure 1). Its lowest occurrence lies in the Hauterivian from Germany (Kirsch and Below 1995). Otherwise, the Cretaceous species Spiniferites lenzii Below 1982c (see Figure 1) is a morphologically very similar species since, although slightly smaller, it also bears numerous intergonal processes. This morphological similarity casts doubt on the identifcation of the Cretaceous cysts as Spiniferites hyperacanthus, and the absence of illustration for these records does not eliminate the doubt (e.g. Masure 1988; Below and Kirsch 1994; Fiet and Masure 2001). In the same way, the record of Spiniferites lenzii in Cenozoic strata (Matsuoka 1983) is remote from the latest (Maastrichtian) continuous occurrence. It might be the result of a reworking, and it is not considered here. All of these occurrences are discussed in Section 4 (see below).

The oldest record of Spiniferites mirabilis (Rossignol 1964) Sarjeant 1970 is from the lowermost Danian of Israel (Eshet et al. 1992), and is confirmed by the record of a Spiniferites sp. cf. S. mirabilis in Danian from Senegal (de Klasz et al. 1987; Jan du Chêne 1988).

The Dinoflagellate cyst Zone V of Costa and Downie (1979) covers the Upper Eocene and its lower boundary is marked by the first occurrence of Spiniferites mirabilis. It is now clear that the FAD of that species is older (Figure 2, Table 2).

The lowest occurrence of Spiniferites pachydermus (Rossignol 1964) Reid 1974 seems to located in the Serravallian (Matsuoka 1983) or the Upper Miocene (Mudie 1989; McMinn 1993; Suc et al. 1995) but in all the cases older than the Gelasian age stated by Ogg and Ogg (2008a).

Spiniferites pseudofurcatus subsp. obliquus (Wall 1967) Lentin and Williams 1973 has almost exclusively been recorded by Wall in Quaternary sediments from the Caribbean Sea. The sole other record is by Wrenn and Kokinos (1986) from an Upper Pleistocene section in Gulf of Mexico in which reworking is frequent. The stratigraphical range of Spiniferites pseudofurcatus (Klumpp 1953) Sarjeant, 1970 appears to end before the onset of the Quaternary (Figure 1, Table 2), so the specific allocation of the obliquus subspecies is questioned.

Foucher (1974) recorded in the upper Turonian from France the oldest occurrence of Spiniferites pseudofurcatus (Klumpp 1953) Sarjeant, 1970. The highest ISA values of that species spans from the Ypresian to the Serravallian (Figure 3). Some specimens have been recorded in Plio-Quaternary strata, but they occur rarely except in Mediterranean area where reworking is frequent (Versteegh and Zonneveld 1994; Corradini in Guerrera et al. 1985; Londeix et al. 1999).

The FAD of Spiniferites rubinus (Rossignol 1962 ex Rossignol 1964) Sarjeant 1970 appears to lie in the upper Burdigalian (offshore Portugal, McCarthy and Mudie 1996). The morphology of the specimen illustrated by El Beialy (1990, pl.4, fig.1) from the Middle/Late Oligocene from Egypt appears too different from Rossignol’s species to be taken into account.

The lowest occurrence of Spiniferites splendidus Harland 1979 seems to be Chattian (Zevenboom 1996), even if there are few records of this taxon. Williams et al. (1998) and Ogg and Ogg (2008a) consider the last occurrence of Spiniferites splendidus as an index species for the mid/upper Zanclean boundary (synchronous with the top of Martini’s (1971) NN13 calcareous nannofossils Zone). This datum should be considered obsolete since Mudie (1986) and Stoker et al. (1994) recorded Spiniferites splendidus in Gelasian strata. Furthermore, Mertens et al. (this volume) propose a synonymy with Spiniferites mirabilis.

The original description of Spiniferites multibrevis indicates the presence of intergonal processes, sometime up to three (Davey and Williams 1966). That feature being shared with Spiniferites hyperacanthus, the stratigraphy of this taxon is studied here. As Spiniferites twistringiensis (Maier 1959) Fensome et al. 1990 is considered to be a senior synonym of Spiniferites multibrevis (Davey and Williams 1966) Below 1982c (see Sarjeant 1983; Fensome and Williams 2004) the stratigraphical ranges of both species have been merged (Figure 1, Table 2). The overal range of Spiniferites twistringiensis appears to be Berriasian (as Spiniferites multibrevis: Ashraf 1979; Heilmann-Clausen 1987) to Serravallian (Figure 1) but it might be still present at Holocene (Grill and Quattroccio 1996).

Cyst species possibly related to the motile Gonyaulax spinifera complex

The FAD of Bitectatodinium tepikiense Wilson 1973 appears to be Lower Miocene within the planktonic foraminiferal Biozones N5 to N7, i.e. Burdigalian or may be upper Aquitanian (Powell 1988b, 1992; Stover et al. 1996) (Figure 3).

Moshkovitz and Habib 1993 report Impagidinium aculeatum (Wall 1967) Lentin and Williams 1981 from Danian sediments. Although the material comes from outcrops, precluding caving, that record is far older from other ones that are Middle (Firth 1996, as Impagidinium sp. cf. I. aculeatum) to Upper Eocene (Brinkhuis 1992; Brinkhuis and Biffi 1993, as Impagidinium sp. cf. I. aculeatum; Gedl and Leszczynski 2005, as Impagidinium aculeatum).

It is recognized that Impagidinium patulum (Wall 1967) Stover and Evitt 1978 FCO is recorded since Lower Miocene (e.g. Powell 1986a, 1992; Brinkhuis et al. 2003; Dybkjaer 2004). Its first appearance is difficult to pinpoint since many occurrences were recorded from Danian strata (Moshkovitz and Habib 1993) to Upper Oligocene (e.g. Strauss 1993; Eidvin et al. 1998; Dybkjaer 2004), sometimes considered as caving (e.g. Eidvin et al. 1998) or with morphologies distant to the type material. Since illustration is not always provided, it is difficult to state.

Head and Nøhr-Hansen (1999) consider Xenicodinium (as Tectatodinium) rugulatum to be a taxonomic junior synonym of Tectatodinium pellitum, therefore, the records taken into account herein are those of both names. The base of the dinoflagellate cyst ‘X. rugulatum Zonule’ proposed by Hansen (1977) and located at the base of the Danian marks the FAD of Tectatodinium pellitum which is just below the base of the calcareous nannofossil Zone NP2 (Stover et al. 1996; Ogg and Ogg, 2008b), and approximately corresponds to the base of the foraminiferal Zone Plb, according to Berggren et al. (1995). The older records of that taxon are thought doubtful (because questioned identification or not illustrated) by Head and Nøhr-Hansen (1999).

The FCO of Tuberculodinium vancampoae (Rossignol 1962) Wall (1967) is probably Priabonian as recorded by several authors (e.g. Costa et al. 1988; El-Bassiouni et al. 1988; Brinkhuis 1992; De Coninck 1999). The records by Fechner and Mohr (1988) and Kar (1985) respectively from Ypresian field section and Lutetian might indicate an earlier stratigraphically position for the appearance of this taxon. The record by Tea-yassia et al. (1999) from lower Maastrichtian is probably due to caving and is not considered here.

Discussion

About Spiniferites ramosus and some others

On the whole, the stratigraphical range of a species is only meaningful when we deal with the same taxon. Since the objective of this study is to depict the stratigraphical continuity of Spiniferites species recorded in Quaternary sediments, the reference morphologies are those of Quaternary species as described in Quaternary literature (e.g. Harland 1983; Powell 1992; Rochon et al. 1999). Many morphological, environmental and sometimes genetic data are available for Quaternary species. The genetic stability over time of the dinoflagellate species is impossible to establish, however, it has been agreed that Spiniferites species have sufficient taxonomical/morphological stability to follow them through geological times.

This is particularly the case for Spiniferites ramosus with a stratigraphical range of nearly 145 Ma, which seems (an oddity in the biosphere evolution) a very peculiar (even strange) case in the biosphere story. Without genetic fingerprinting, the concept of the morphological species is the lone guiding thread we can follow along such time intervals. In the case of Spiniferites ramosus, it corresponds to an ovoid central body with a smooth to shagreenate surface, with gonal processes only and distal trifid, then sometimes bifid, terminations (see Mertens et al. and Londeix et al. 2018). The length of the process is variable and intergonal process may occasionally be present (maximum one per suture). This morphological concept corresponds to the Cretaceous type of the species, illustrated by Davey in Davey and Williams (1966, fig.8) and is consistent with the Quaternary specimens (e.g. Rochon et al. 1999, pl.9, fig.4-6). Although stratigraphically distant, these two specimens illustrate the morphological standard of the species. The stratigraphical range depicted here concerns Spiniferites ramosus ramosus and excluded the many subspecies (> 32, Fensome and Williams 2004) assigned to Spiniferites ramosus.

If we accept that a species is ubiquitous because it is able to adapt to many environmental conditions, it is logical that its stratigraphical perenity is important. This is indeed what is observed for Spiniferites ramosus which proves to be a common species in modern sediments and whose current geographical distribution is very important, especially in environments with high seasonality (e.g. Zonneveld et al. 2013). Thus the long stratigraphical range of Spiniferites ramosus might be regarded as a taxonomical stability, at least in a paleontological point of view. The ‘index of stratigraphical abundance’ for Spiniferites ramosus shows two main optima: one during the Late Cretaceous and the second one during the Quaternary (Figure 1). In addition, Spiniferites ramosus ISA variations show that its FCO is in the base of Valanginian while its FOD is in the lowermost Berriasian.

The main features of Spiniferites hyperacanthus are the regular presence of several intergonal processes and low (or absent) septa. However, some modern specimens of this species show an unexpected morphological plasticity with the presence of well expressed septa (see Limoges et al. 2018). Such morphology is very close to that of Spiniferites lenzii and Spiniferites twistringiensis leading to wonder how to distinguish these taxa and whether they represent various morphotypes of the same species. The taxonomical identity of these taxa is not questioned here, but the discontinuous ISAs of Spiniferites hyperacanthus and Spiniferites lenzii are puzzling (Figure 1). Nevertheless, the presence of low to absent septa (i.e. Spiniferites hyperacanthus) appears to be a feature more frequent during Neogene and Quaternary times than in previous periods. This is also observed with Spiniferites mirabilis ISA (Figure 2), a species morphologically close to Spiniferites hyperacanthus. Spiniferites cf. Spiniferites lenzii illustrated by Oboh-Ikuenobe et al. (1998, pl.9, figs. 2-3) from the upper Maastrichtian offshore Ivory closely ressembles Spiniferites hyperacanthus. Caving is unlikely since many specimens were encountered (op. cit.). On the opposite, specimen illustrated by Lister and Batten (1988, pl.11, fig.6) as Spiniferites hyperacanthus appear very close to Spiniferites lenzii. Since it is difficult to to evaluate the records of non-illustrated Cretaceous specimens, no occurrence of this period has been ruled out here. On the other hand, the Plio-Pleistocene record of Spiniferites lenzii from Japan (Matsuoka 1983) is not taken into account.

The emergence of Spiniferites morphologies with intergonal processes appears relatively early (Berriasian-Valanginian), showing that it is probably a character deeply rooted into the genus genome.

ISA comparison for some possibly related taxa

During the Montréal and Ostend round tables dealing with Spiniferites and Achomosphaera (see Mertens et al. this volume) the synonymization of some species was discussed. The ISAs of some of them are here discussed.

Spiniferites ramosus, Achomosphaera ramosasimilis and Spiniferites bulloideus are morphologically close species. The first appearance of Spiniferites ramosus and Spiniferites bulloideus appear far apart, respectively Berriasian and no older than Ypresian (Figures 1 and 2). The time lapse between the first appearance of Spiniferites ramosus and Achomosphaera ramosasimilis appears closer, since the latter appears in the Valanginian. In addition, no similarity is observed between Spiniferites ramosus and Spiniferites bulloideus ISAs (Figure 4A) but variations of Achomosphaera ramosasimilis ISA are very similar to those of Spiniferites ramosus during the Neogene and the Pliocene, leading to consider a possible taxonomical relationship between these two taxa.

With the same approach, Spiniferites mirabilis and Spiniferites hyperacanthus ISAs appear to be correlated from the Rupelian to the Pleistocene (Figure 4B). The similarity of their ISAs reinforces the assumption they could be two morphotypes of a same taxon. The records of Spiniferites serratus and Spiniferites splendidus are too few to make significative the comparison of their ISA to that of Spiniferites mirabilis.

It is of note that the ISA of Nematosphaeropsis labyrinthus is very close to that of Spiniferites mirabilis (Figure 4B) This may suggest a possible link between the two taxa although the two morphologies are very different (intergonal processes and antapical veil in Spiniferites mirabilis, presence of trabeculae in Nematosphaeropsis labyrinthus).

Several types of cysts are possibly produced by Gonyaulax spinifera-type motile cells: e.g. Spiniferites elongatus, Spiniferites membranaceus, Spiniferites ramosus, Spiniferites bulloideus, Spiniferites mirabilis, Nematosphaeropsis labyrinthus, Bitectatodinium tepikiense, Tectatodinium pellitum, Ataxiodinium choane etc. (Rochon et al. 2009).

The ISAs variations through geological time of Spiniferites elongatus and Impagidinium sphaericum show an impressive covariance (Figure 4C). It gives substance to similar ecological affinities between the two species that are generally associated with sediments in polar to sub-tropical waters (Zonneveld et al. 2013; de Vernal et al. 2018). It also questions whether a biological affinity might exist between the two taxa. As respectively depicted by Helenes (1986) and Ellegaard et al. (2003) the paratabulations of Impagidinium sphaericum and Spiniferites elongatus show differences in the apical and precingular 6’’ paraplates, indicating their ISA covariance is rather due to ecological similarities instead of taxonomical ones. According to Harland (1983), the motile cell of Impagidinium sphaericum would be a Gonyaulax sp. indet. That of Spiniferites elongatus is considered to be Gonyaulax spinifera or Gonyaulax elongata (Head 1996; Ellegaard et al. 2003). It is noteworthy that Helenes (1986) already indicated paratabulation similarities between what he called Sphaericum pattern (represented by Impagidinium sphaericum) and the Spinifera pattern related to Gonyaulax spinifera.

To a lesser extent, the ISA variations of Spiniferites elongatus and Bitectatodinium tepikiense also show a fair similarity (Figure 4C), which probably has to be interpreted as ecological similarities. The records of Spiniferites ellipsoideus, Spiniferites frigidus and R. amphicavata are too few to make the comparison of their ISAs significant to that of Spiniferites elongatus.

Cysts of the motile species Gonyaulax digitalis (Pouchet 1883) Kofoid 1911 are thought to be referable to Spiniferites bentorii (cf. Head 1996) or to Bitectatodinium tepikiense (Ellegaard et al. 2003). The comparison of ISA of Spiniferites bentorii and Bitectatodinium tepikiense does not show any relationship between these two taxa (Figure 4D).

No other covariance has been detected between taxa related to Gonyaulax spinifera-type motile cells.

As inferred from partial LSU rDNA sequence data, Ellegaard et al. (2003, fig.46) propose a phylogeny of some Gonyaulax motile species. According to this method, Gonyaulax baltica would have appeared first and Gonyaulax membranacea would have appeared last (Figure 5A). In this tree, Gonyaulax elongata, Gonyaulax digitalis (= Spiniferites bentorii or Bitectatodinium tepikiense, Head 1996; Ellegaard et al. 2003) and Gonyaulax cf. spinifera (= Spiniferites ramosus, Ellegaard et al. 2003) appear successively between the two former species. That topology differs from the appearance chronology of the cyst equivalent of these species since Spiniferites ramosus first occurs at ca. 145 Ma (lowermost Berriasian) and the cyst of Gonyaulax baltica only in the Holocene (less than 11 ka). In addition, the order of appearance of Bitectatodinium tepikiense, Spiniferites bentorii, Spiniferites elongatus, Spiniferites membranaceus is respectively around 13 Ma (Serravallian), 23 Ma or maybe 58 Ma (Aquitanian or Thanetian), 37 Ma (Priabonian) and 112 Ma (Albian) (Figure 5A, Table 2). The stratigraphical appearances seem therefore to show a different pattern from that obtained by genetic analysis of motile cells, but this difference can simply mean that the stratigraphical range takes into account type morphologies, and not biological species.

Conclusion

The ‘index of stratigraphical abundance’ of the Quaternary Spiniferites species show only few similarities between them. Some couples of taxa show ISAs’ similarities, such as Spiniferites mirabilis and Spiniferites hyperacanthus, and to a lesser degree, Spiniferites ramosus and Achomosphaera ramosasimilis. ISAs’ comparisons with species outside of the Spiniferites genus sometimes show striking similarities such as for Spiniferites mirabilis and Nematosphaeropsis labyrinthus, and for Spiniferites elongatus and Impagidinium sphaericum. In a first approach, the similarities of ISA between two taxa could have been interpreted as indicative of cyst morphotypes belonging to the same species. The example of Spiniferites elongatus and Impagidinium sphaericum refutes this hypothesis since these two cysts do not have the same tabulation, which excludes the possibility that they could belong to the same species. This leads to dismiss the idea to use ISAs as a tool for taxonomical discrimination. However, it could be a suitable way to discern taxa having similar ecological preferences.

The fact remains that the ISAs allow a better visualization of abundance variations of a taxon, or a morphology, over time than the classic linear stratigraphical range. The ISA proves to be a valuable tool for locating in time the FCO and the LCO.

Moreover, it would be interesting to test the ISAs of some species (of Spiniferites and others) on a regional scale and not on a global scale, as was carried out in this paper, to refine the stratigraphical interest of taxa.

Acknowledgements

I wish to express my gratitude to James Riding (British Geological Survey) for his review enriched by his subsequent advice and constructive exchanges. Stephen Louwye (Ghent University) is thanked for his helpful and constructive reviews of the manuscrit and for his patience to improve the text, as well as Tristan Cantat-Gaudin (University of Padova) and Fabienne Marret (University of Liverpool).

This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.