Introduction

The Family Strombidae arose during the Eocene and diversified during the Late Oligocene and Early Miocene (Williams and Duda 2008 and references therein) and is currently represented by about 100 species, restricted to the tropics (Abbott 1960). Most strombids live in shallow water environments. They are among the most eye-catching gastropods, and attract a broad community of scientists and collectors. From the Eocene about five species have been described within the genus Strombus sensuAbbott (1960), and during LateOligocene and Miocene times strombids experienced their first main radiation with about 40 species described from the Miocene (Wieneke et al. 2011). The phylogeny and the generic affiliations of the various fossil and extant taxa, however, have only been slowly resolved. After early attempts of Abbott (1960), no major breakthrough was achieved before the 21st century when a set of papers tried to clarify the validity of genus-rank taxa and to allocate species groups to these genera (Kronenberg and Vermeij 2002; Kronenberg and Lee 2005, 2007; Bandel 2007). Molecular data generally support their conclusions (Latiolais et al. 2006) and suggest very complex phylogeographic patterns. This newly established system has also been gradually applied to fossil taxa (Harzhauser and Kronenberg 2008; Wieneke et al. 2011). The correct allocation of all these taxa is crucial for supportingmolecular datawith the fossil record.

The deposits of the various European basins, which were covered by the Miocene Paratethys Sea, present an outstanding archive of fossil strombids. Since the 19th century, the Middle Miocene strombids of the Paratethys and the protoMediterranean Sea have been usually treated as Strombus coronatus Defrance, 1827—a largely Pliocene species—or Strombus bonelli Brongniart, 1823, which is an Early Miocene species. Generally, robust shells with strongly developed spines were treated as S. coronatus whilst more slender specimens with reduced sculpture were tentatively treated as S. bonelli. Many authors, such as Hörnes (1853), Hoernes and Auinger (1884), Strausz (1966), and Bałuk (1995), emphasized the problems in assigning the specimens to one of the taxa due to the presence of intermediate morphologies. This vague taxonomic concept resulted in an apparent stratigraphic range of c. 20 Ma., from the Late Oligocene to the Pliocene, for Strombus coronatus. Especially during the early 19th century, during what might be called the “pioneer phase of conchology”, authors were unaware of the exact stratigraphic age of the various localities and this led to unrealistically long stratigraphic ranges for species. Similarly, biogeographic patterns have been obscured, as even eastern African fossils were treated as Strombus bonelli (Collignon and Cottreau 1927). Herein, we try to evaluate the taxonomic concepts applied to the various Paratethyan and some protoMediterranean representatives of Persististrombus to clarify their biostratigraphic and biogeographic value.

Fig. 1.

Distribution of the Persististrombus species plotted on a map of the Middle Miocene Paratethys Sea (modified from Harzhauser et al. 2007). Stippled area represents the distribution of Persististrombus inflexus (Eichwald, 1830). 1, Persististrombus exbonellii (Sacco, 1893); 2, Persististrombus pannonicus Harzhauser and Kronenberg nov. sp.; 3, Persististrombus lapugyensis (Sacco, 1893); NAFB, North Alpine Foreland Basin. Note that this map shows an Early Langhian situation, whereas the distribution data unite Langhian and Serravallian occurrences.

Institutional abbreviations.—GBA, Geological Survey Vienna; NHMW, Natural History Museum Vienna. All measured shells are stored in the collections of the Natural History Museum Vienna (Austria), the Krahuletz Museum, Eggenburg (Austria), and the Naturalis Biodiversity Center, Leiden (The Netherlands) as indicated in the supplementary data table (SOM, Supplementary Online Material at  http://app.pan.pl/SOM/app58-Harzhauser_Kronenberg_SOM.pdf).

Geological setting

The discussed strombids derive from three palaeogeographic areas: the Paratethys Sea, the proto-Mediterranean Sea, and the Eastern Atlantic. In terms of biogeography, all three belonged to the Proto-Mediterranean-Atlantic Region during the Early to Middle Miocene (Harzhauser et al. 2002). Its geographic extent roughly corresponds to the limits of themodern Mediterranean-Atlantic Region (sensu Briggs 1995). The mentioned strombids from Turkey, Greece, Northern Italy, and France belong to this biogeographic unit, which was part of the ProtoMediterranean-Atlantic Region during the Burdigalian (Harzhauser et al. 2002). Most of the discussed taxa are found in deposits of the Paratethys Sea, which appeared as a distinct palaeogeographic and paleobiogeographic unit around the Eocene-Oligocene boundary and lasted until the Pliocene (Rögl 1998). During its maximum extent, the Paratethys Sea spread from the Rhône Basin in France towards Inner Asia.

This sea underwent an extraordinary history of total or partial isolation, reflected in a phase of high endemism, alternating with various connections to the adjacent seas (Harzhauser and Piller 2007). Strombids from Bosnia, Austria, Romania, Hungary, Poland, and NW Bulgaria, mentioned in this paper, belong to the fauna of the Paratethys Sea (Fig. 1). The peculiar development that was forced mainly by geodynamic processes renders a system of regional stages necessary, as the regional stage boundaries do not always correlate with those of the Mediterranean standard scale (Fig. 2). A synopsis of the paleogeography of the area in several time slices was presented by Rögl (1998) and Popov et al. (2004). Harzhauser et al. (2002) and Harzhauser and Piller (2007) present a detailed introduction into the Oligocene to Miocene paleobiogeography in the circum-Mediterranean area with emphasis on the Paratethys Sea. An update of the chronostratigraphy and the use of the regional stages are presented in Piller et al. (2007). For the ages and biostratigraphic correlations of the most important localities mentioned in text see Studencka et al. (1998), Harzhauser et al. (2003), Rögl et al. (2008), Harzhauser and Piller (2007), and Zuschin et al. (2007, 2011); geographic maps showing all the localities in great detail are presented by Kroh (2005).

Fig. 2.

Oligocene-Miocene chronostratigraphy, magnetostratigraphy, and biostratigraphy after Hilgen et al. (2009) with regional Paratethys stages after Piller et al. (2007). The figure is modified from a chart produced with the Time Scale Creator program, provided by the International Commission on Stratigraphy (available at  https://engineering.purdue.edu/Stratigraphy/tscreator/index/index.php). The ranges of the treated strombid taxa are indicated as grey bars. Abbreviations: CN, calcareous nannoplankton; PF, planktonic Foraminifera.

Material and methods

We compiled morphometric data on a total of 219 Miocene to Pliocene European strombid shells. TheMiocene shells derive from the Egerian (= Chattian-lower Aquitanian), Eggenburgian (= lower Burdigalian) and Badenian (= Langhian-lower Serravallian) stages. The material comprises juvenile to adult specimens of Persististrombus nodosus (n = 8), P. praecedens (n = 4), P. inflexus (n = 36), P. exbonellii (n = 70), P. lapugyensis (n = 48), P. pannonicus (n = 6), and P. coronatus (n = 47). The measurements focused on the total height, the maximum width, the maximum width without spines, the width in dorsal-ventral direction with and without spines, of the last whorl and of the last spire whorl, and the angles of the apex and of the last whorl (Fig. 3, SOM). To evaluate the robustness of our taxonomic concept we performed principal component analysis with the software package PAST (Hammer et al. 2001).

Each taxon is described and we try to present a chresonymy (sensu Smith and Smith 1972) for each species, focusing only on relevant references with illustrations of Paratethyan occurrences (aside from P. nodosus). This strict use of “synonymy lists” avoids biasing the stratigraphic and geographic pattern by including poorly documented literature data. Moreover, we avoid a subjective interpretation of the original reference. In several 19^th century papers, new taxa are introduced as variation names within chapters on other species. These names are often not accompanied by a clear assignment to a certain genus (e.g., Sacco 1893).

Fig. 3.

Measurements used for the statistical analysis.Abbreviations: a,maximum height; b, height of last whorl; c, height of last spire whorl; d, apical angle; e, basal angle; f, dorsal width; g, dorsal width without spines; h, ventral width; i, ventral width without spines.

Systematic paleontology

This paper was triggered by studies of species previously allocated to Strombus (Lentigo) by Lozouet and Maestrati (1986). These were followed by other studies by Jung and Heitz (2001) and Kronenberg and Lee (2007). In all of these papers, the influential monograph by Abbott (1960) has been the starting point of the assignment of the lineage to Lentigo Jousseaume, 1886 (type species by monotypy: Strombus lentiginosus Linnaeus, 1758). Lozouet and Maestrati (1986) focused on the affinities of the extant Strombus granulatus Swainson, 1822 with the Paleogene European representative of the genus. Thereafter, Jung and Heitz (2001) tried to shed light on the little-known history in the Americas. Kronenberg and Lee (2007) pointed out that the allocation to Lentigo was incorrect, and described a new genus, Persististrombus, to incorporate a number of these species, based on both shell morphology and DNA research (Latiolais et al. 2006). The genus was established for the extant Strombus granulatus Swainson, 1822 in the Panamic Province. Kronenberg and Lee (2007) list several extinct American Persististrombus species, to which Strombus goeldii Ferreira and Cunha, 1957, described from the Paribas Formation (Lower Miocene) of Brazil, should be added. Later, this concept was also applied to Strombus latus Gmelin, 1791 from the AfricanEastern Atlantic Province and the Pliocene Strombus coronatus Defrance, 1827 in the Mediterranean Sea (Harzhauser and Kronenberg 2008).

Taxonomic concept.—“The species problem” has been already discussed in numerous contributions (see Hausdorf 2011 for a recent overview). Unfortunately it is impossible to extract DNA or examine soft parts of fossil strombids. Thus, we can rely only on shell morphology. As far as our species concept is concerned, we take a pragmatic approach. When morphometric data combined with a passing in geological time (eventually coinciding with a change in paleoenvironmental conditions) show that samples differ from one another, we consider them to be distinct morphospecies. Although this morphospecies concept is inadequate for modern biology, the interpretation of the paleontological record is obviously limited to it. In the case of strombids, however, the molecular analysis of numerous extant strombid taxa, performed by Latiolais (2003) and Latiolais et al. (2006), was not in conflict with the traditional morphospecies assignments. This points to a useful level of reliability of recognizing strombid species based on shell morphology.

Simpson (1961) defined a lineage as “an ancestral-descendant sequence of populations” forming an evolutionary species. Herein, we consider the long-lived and widespread Persististrombus inflexus to represent such a lineage. The geographically rather isolated and stratigraphically restricted taxa are treated herein as offshoots of thismain-lineage. Offshoot is a widely used but not strictly defined term. It may refer to species or to higher taxa (e.g., Strong and Köhler 2009). The term offshoot is used herein sensu Grant (1963), who stated that ancestral species may form offshoot populations which develop into daughter species. In this sense, the term was frequently used to describe a speciation event from the predominant lineage with phenotypic differentiation in both vertebrates and invertebrates (Eldredge and Gould 1972; Johnson 1980; see also Miller 2001). The trigger, which allowed offshoots to develop into distinct morphospecies, might largely have been geographic separation, in the case of Paratethyan strombids. A comparable but much more pronounced speciation event due to geographic isolation has been documented for the Sarmatian Paratethys, when more than 100 new mollusc species evolved endemically in that completely isolated sea (Papp 1954; Harzhauser and Kowalke 2002).

Fig. 4.

Miocene–Pliocene strombid gastropods from Europe. A–C. Persististrombus nodosus (Borson, 1820) from the Burdigalian (Early Miocene). A. NHMW1911/0006/0290, Pelona, France. B. NHMW 2013/0299/001, Saucats, France. C. NHMW1851/0017/0116; Colli Torinesi, Italy. D. Persististrombus praecedens (Schaffer, 1912) from the Burdigalian of Loibersdorf in Lower Austria, NHMW1866/0011/0004. E. Persististrombus coronatus (Defrance, 1827) from the Pliocene of Tresanti in Italy, NHMW A2588. Dorsal (A₁–E₁), ventral (A₂–E₂), lateral (A₃–E₃), and apical (A₄–E₄) views. Scale bars 10 mm.

Fig. 5.

Strombid gastropod Persististrombus inflexus (Eichwald, 1830) from the Middle Miocene of the Vienna Basin. A. NHMW2012/0080/0001, Vöslau. B. NHMW2012/0081/0001, Enzesfeld. C. NHMW1855/0002/0052, Enzesfeld, illustrated in Hoernes and Auinger (1884: pl. 19: 1). D. NHMW1874/ 00024/0041, Vöslau, illustrated in Hoernes and Auinger (1884: pl. 18: 4). E. NHMW 2012/0081/0002, Enzesfeld. In dorsal (A₁–E₁), ventral (A₂–E₂), lateral (A₃–E₃), and apical (A₄–E₄) views. Scale bars 10 mm.

Fig. 6.

Strombid gastropod Persististrombus exbonellii (Sacco, 1893). A–D. Middle Miocene of Gainfarn in Lower Austria. A. NHMW1855/0045/0420. B. NHMW1855/0045/0419. C. NHMW1846/0037/0185. D. NHMW1847/0037/0058. E. NHMW62275, Middle Miocene of Vöslau in Lower Austria. In dorsal (A₁–E₁), ventral (A₂–E₂), lateral (A₃–E₃), and apical (A₄–E₄) views. Scale bars 10 mm.

Fig. 7.

Strombid gastropods from Austria and Turkey. A. Juvenile specimen with protoconch of Persististrombus exbonellii (Sacco, 1893), NHMW 2013/0299/0002, from the Middle Miocene of Gainfarn in Lower Austria, in ventral (A₁, A₂), and apical (A₃) views. B. Protoconch of Persististrombus inflexus, NHMW 2013/0299/0003, from the Serravallian of Karaman in Turkey, in ventro-apical (B₁) and ventral (B₂) views. Scale bars 1 mm.

Fig. 8.

Strombid gastropods from Europe. A–C. Persististrombus lapugyensis (Sacco, 1893) from the Langhian (Middle Miocene) of Lãpugiu de Sus in Romania. A. NHMW1854/0035/0150. B. NHMW1855/0043/0017. C. NHMW1866/0040/0270, illustrated in Hoernes and Auinger (1884: pl. 18: 1). D, E. Persististrombus pannonicus sp. nov. from the Serravallian of the Oberpullendorf Basin. D. Holotype, NHMW1930/0006/0058, Ritzing, Austria. E. Paratype, NHMW1970/1396/0609, Brennberg, Austria. Note the different angles between the aperture and the axis in both species (A₃–E₃) and the different spine morphology (A₄–E₄). In dorsal (A₁–E₁), ventral (A_2–E₂), lateral (A₃–E₃), and apical (A₄–E⁴) views. Scale bars 10 mm.

Discussion

Morphometrics.—The comparable outline of the specimens and the broad range of intraspecific variability prevent a clear separation of the proposed taxa based only on height-width data (SOM). Moreover, juvenile specimens hardly differ morphometrically. Therefore, the data set was reduced to adult shells in Fig. 9. The best separation of the taxa in scatter plots is achieved when the relation of the height of the last spire whorl with the total height is compared to the apical angle. Especially the height of the last spire whorl is a significant character, which allows distinguishing P. coronatus-like species fromthe Pliocene P. coronatus. Fig. 9 illustrates a clear transition fromthe P. nodosus-field, via P. praecedens, and a cluster of P. exbonellii to P. lapugyensis. Although overlapping with P. pannonicus and P. lapugyensis, the bulk of P. coronatus forms a clearly separate cluster.

A principal component analysis on the full data set including all specimens reveals a similar pattern (Fig. 10). A separation of a cluster with subadult specimens is mainly caused by size effects. This indicates that all juvenile shells of the Persististrombus inflexus-lineage are very similar and that species-specific features arise only later during ontogeny. Adult shells cluster in separate fields: P. nodosus and P. praecedens show no overlap. Similarly, adult shells of P. coronatus, P. lapugyensis, and P. exbonellii settle in separate areas in the plot, being separated mainly by the characters “spire angle”, “height of the last spire whorl” and the “number of spines on the last whorl”. As the characters “height of the last spire whorl” and “apical angle” turned out to be significant, a further PCA was performed on a reduced data set including the relation “height/width (max)” (Fig. 11). Again, P. coronatus separates from the P. inflexus and its offshoots. The latter being split into a P. lapugyensis and a P. exbonellii cluster with specimens of P. inflexus as intermediates. The analyses document the difficulties to separate juvenile shells of Persististrombus based on the implemented measurements but perform successful for adult shells. In all analyses, a separation of the Late Miocene to Pliocene Persististrombus coronatus from Early to Middle Miocene members of the Persististrombus inflexus-lineage and its offshoots is obvious.

Fig. 9.

Scatter plot of adult specimens of various Persististrombus species based on the apical angle and the relation between total height and the height of the last spire whorl. Abbreviations refer to the specimens indicated in SOM: c, P. coronatus; e, P. exbonellii; i, P. inflexus; la, P. lapugyensis; n, P. nodosus; p, P. praecedens; pa, P. pannonicus.

Fig. 10.

Principal component analysis based on all specimens (see text for details) performed with the software package PAST (Hammer et al. 2001). Abbreviations refer to the specimens indicated in SOM: c, P. coronatus; e, P. exbonellii; i, P. inflexus; la, P. lapugyensis; n, P. nodosus; p, P. praecedens; pa, P. pannonicus.

Extrinsic factors triggering speciation.—As a whole, all specimens of the P. inflexus-lineage display a degree of variability similar to that of Recent species of Persististrombus, viz. P. granulatus (Kronenberg and Lee 2005) and, to a slightly lesser extent, P. latus (DeTurck et al. 1999). Our data show that there is a tendency for an iterative but independent development of spiny morphs. This phenomenon is not restricted to the P. inflexus-lineage and its offshoots but was also documented for other strombids (Landau et al. 2011). Absence or presence of predators has bearing on shell morphology. In laboratory conditions, the presence of the lobster Panulirus argus induced changes in both behavior and shell growth, i.e., shells exposed to the lobsters grew slower, yet the shell weight remained the same (Delgado et al. 2002). Also Herbert et al. (2004) noted the development of antipredatory traits in members of the Strombus alatus complex in presence of a predator. This pressure was concluded based on an increase in repair-marks by over 90%. Anti-predatory trait evolution in these Strombus was inferred from increases in adult mean size and lip thickness, the percentage of individuals with knobs on the last whorl, the maximum number of knobs on the last whorl and the growth rates, deduced from oxygen stable isotope sclerochronology. A similar scenario might have played a role in the development of the strongly spined species in the Paratethys. A statistical analysis of repair-marks is beyond the scope of this study. Nevertheless, our unpublished observations indicate that the frequency of repair-marks on shells of the spiny P. lapugyensis does not differ very much from that of the rather smooth P. exbonellii and is low in both cases.

Fig. 11.

Principal component analysis based on adult specimens (see text for details) performed with the software package PAST (Hammer et al. 2001). Abbreviations refer to the specimens indicated in SOM: c, P. coronatus; e, P. exbonellii; i, P. inflexus; la, P. lapugyensis; n, P. nodosus; p, P. praecedens; pa, P. pannonicus.

Another factor influencing the radiation of the P. inflexuslineage in the Paratethys might be related to paleogeography. Three of the described species are chronologically and geographically well separated from each other (Fig. 1). This pattern might be caused by the archipelago character of the Central Paratethys during the Middle Miocene (Rögl 1998). The Paratethys was connected to the Proto-Mediterranean Sea only through relatively narrow channels (Fig. 1), whilst the Central Paratethys itself was a dead end sea, i.e., with no connections to other ocean basins. These channels in all probability have hampered faunal exchanges, as now is the case between the Mediterranean Sea and the Black Sea, the Atlantic and the Mediterranean Sea, and the Red Sea and Persian Gulf with the Indian Ocean. The rapid changes in relative sea-level combined with the tectonic activity in an active back-arc/ fore-arc basins system may have led to a series of separations and reconnections of shallowmarine faunas. Large scale separations of the north-eastern branch of the Paratethys (e.g., Polish-Carpathian Foredeep) and the southern parts (Pannonian basins complex) also had significant impact on the stable isotope composition of the sea-water (Latal et al. 2006). Thus, the geographic isolation of populations might have had a strong impact on speciation as already documented for some nassariid gastropods (Harzhauser and Kowalke 2004).

Oligocene to Holocene biogeography of Persististrombus . —The roots of the genus reach back to the TethyanOligocene. The oldest species that can be attributed to Persististrombus is Strombus radix Brongniart, 1823, from the Piedmont Basin and the Vicentin in Italy (Rovereto 1900; Fuchs 1870), the Mesohellenic Basin in Greece (MH own observation), and Bulgaria (Karagiuleva 1964). In the west, it reached even the Adour Basin in France (Lozouet and Maestrati 1986). Its easternmost occurrences are recorded from the Rupelian of the Central Iranian Esfahan-Sirjan Basin and the Kutch Basin in India (Harzhauser 2004; Harzhauser et al. 2009). This Oligocene species is unknown so far from the Paratethys. Already during the Oligocene, an eastern lineage established in the Rupelian Acropora reefs of the Arabian Peninsula, represented by the small and spiny Persististrombus bernielandaui (Harzhauser, 2007). This lineage continued into the Early Miocene and is represented by P. gijskronenbergi (Harzhauser, 2007) in Oman and P. kronenbergi Harzhauser, 2009 in Tanzania during Aquitanian times. Burdigalian species in this lineage are P. deperditus (Sowerby, 1840) in Kutch in northern India and P. quilonensis (Dey, 1961) from Kerala in southern India. The last known species in the Indo-WestPacific Region (IWP) is P. preoccupatus (Finlay, 1927) from the Late Miocene of Borneo (Beets 1941). Thus, the genus seems to have become extinct in the IWP around the Miocene/Pliocene boundary. However, a genus still pending description that probably arose from a Persististrombus ancestor, was present in Indonesia during the Miocene, and persisted until the latest Pliocene (GCK, unpublished data).

Transatlantic migration of the genus into the Americas took place during the Early Miocene when Persististrombus appears as P. goeldii (Ferreiro and Cunha, 1957) in the Paribas Formation of Brazil, and as P. aldrichi (Dall, 1890), P. chipolanus (Dall, 1890), and P. mardieae (Petuch, 2004) in the Chipola Formation of Florida (Petuch 2004). After a major stratigraphic gap, the genus re-appears in the Americas and enters the eastern Pacific. At present it is not clear whether Persististrombus got extinct and re-invaded the Americas or that specimens are not preserved or not yet discovered. Based on the morphology of P. granulatus, which is rather close to P. radix and P. nodosus we are inclined to believe the latter to be themost likely scenario.Afterwards, P. toroensis (Jung and Heitz, 2001) and P. insulanus (Jung and Heitz, 2001), possibly synonyms of P. granulatus (Swainson, 1822), are reported from the Pliocene of Panama. P. barrigonensis (Jung and Heitz, 2001), subsequently synonymized with P. granulatus by Landau and Silva (2010), occurs in the Late Miocene or Pliocene ofVenezuela, and P. obliteratus (Hanna, 1926) in the Pliocene of California (all data from Wieneke et al. 2010, see also Kronenberg and Lee 2007). Persististrombus granulatus has been reported fromthe Lower Pliocenemember of the Imperial Formation in California (Powell, 1988), and survives to the Recent in the Panamic province (Emerson and Old 1963).

In the Western Tethys, the Oligocene P. radix was followed during the Aquitanian and Burdigalian by P. nodosus (Borson, 1820). Its geographic range was comparable to that of its Oligocene precursor spanning from the Bay of Biscay in the west via the Mediterranean area probably to the Qom Basin in the east (Harzhauser et al. 2002). Unequivocal Early Miocene Paratethyan occurrences are unknown so far. The small P. praecedens (Schaffer, 1912), however, dwelled as Paratethyan endemic from Chattian to early Burdigalian times. The development of Persististrombus in the Paratethys during the late Burdigalian (Ottnangian and Karpatian regional stages) is unclear as the few specimens are too fragmentary or poorly preserved for identification at the species level.

With the onset of the Middle Miocene Climatic Optimum, the genus become very successful in the Paratethys and as P. inflexus (Eichwald, 1830) is present in even the northernmost basins in the Polish-Carpathian Foredeep. Nevertheless, it did not enter the Eastern Paratethys where no strombids are known so far (Iljina 1993).

P. lapugyensis developed geographically discrete populations around 16–15 Ma. in the southern part of the Paratethys during the Middle Miocene Climatic Optimum. At the same time, P. inflexus is recorded in all northern basins such as the North Alpine Foreland Basin, the Vienna Basin and the Polish-Carpathian Foredeep (see systematic chapter). A second offshoot, the delicate and elongate P. exbonellii (Sacco, 1893), developed around 14.5–14.0 Ma. in the seagrass meadows along the western margin of theVienna Basin and disappeared soon thereafter. The evolution of this species coincides with the cooling duringMiocene Climate Transition when a drastic reduction of gastropod species occurred in the Paratethys Sea (Harzhauser and Piller 2007). The last, regionally defined offshoot developed during the Serravallian in an embayment of the Pannonian basins complex around 13.0 Ma ago, resulting in the P. coronatus-like P. pannonicus. Still, P. inflexus was represented at least as far north as the Vienna Basin. At that time, P. inflexus also occurred in the Mediterranean Sea, where it formed huge populations in the Turkish Karaman Basin (MHunpublished data) and appeared even in the Loire Basin (Glibert 1949). The poor documentation of the genus in the Middle Miocene of the Mediterranean area is probably linked to the low amount of shallow marine siliciclastic in Middle Miocene deposits. The wealth of Langhian and Serravallian deposits in the Paratethys Sea is contrasted by relatively few coeval outcrops in the Mediterranean area. Therefore, the impression that the Persististrombus inflexus-lineage experienced an extraordinary bloom only in the Paratethys Sea has to be considered with caution.

Due to the changing water chemistry of the Paratethys, the entire Persististrombus inflexus-lineage became extinct during at the end of the Serravallian together with most stenohalinemarine taxa (Rögl 1998; Harzhauser and Piller 2007). It seems to have vanished also in the Mediterranean area around the Middle/Late Miocene boundary and became replaced by Persististrombus coronatus. This species appears during the Tortonian in the Mediterranean Sea (Sacco 1893) and persisted to theMessinian of Libya from where Bandel (2007) reported on aberrant specimens. Later it becomes a very common species during the Zanclean and the Early Piacenzian (Sacco 1893; Stchépinsky 1939, 1946; Harzhauser and Kronenberg 2008). It might have originated along the West African Coast as suggested by a report of the species by Brébion (1983) from the Middle or LateMiocene ofAngola. An Atlantic distribution of P. coronatus is documented byMeco (1977) from the Pliocene of the Canary Islands.

Persististrombus coronatus disappears from the Mediterranean Sea completely with the onset of the Late Pliocene cooling (Landau et al. 2004) and seems to be extinct thereafter. During the Pleistocene Persististrombus latus (Gmelin, 1791) represents the European Persististrombus-lineage. This extant species is restricted to the African-Eastern Atlantic Province but invaded the Mediterranean Sea during the Pleistocene. There it appears during the warm phases of the Marine Isotope Stages 7 and 5 (De Torres et al. 2009) and probably also during MIS 3 (Zazo et al. 1984; Rögl et al. 1997).

Implications for paleoclimate reconstructions.—The genus is represented by two extant species: P. granulatus (Swainson, 1822) in the Panamic Province and Persististrombus latus (Gmelin, 1791) in the African-Eastern Atlantic Province. The distribution data of both taxa are shown in Fig. 12 and compared with long-term summer and winter sea surface temperatures, based on LEVITUS (1994).

Persististrombus latus occurs in an area from the Rio D'Oro in Morocco/Mauritania in the north to Angola in the south (Rolán and Ryall [1999]), being limited in the north by the Canary Current and in the south by the Benguela Current (Meco et al. 1997). In both areas, the 20–21°C SST isotherms seem to be a limiting factor for its frequent distribution. Its northern-most limit is the Cape Verde Islands (Kreipl and Poppe in De Turck et al. 1999) where water temperatures are usually warmer than 16°C (Meco 1977), which seems to be the very limit for dispersal of this species.

Fig. 12.

Distribution of the two extant species of Persististrombus in the Panamic Province and the African-Eastern Atlantic Province (after Meco 1977 and Kreipl and Poppe 1999). The data are plotted on sea surface temperature maps generated with the World-Ocean-Atlas-Data visualization program of the NOAA based on the LEVITUS 1994 data sets ( http://www.esrl.noaa.gov/psd/ ).

Persististrombus granulatus occurs from the Baja California in the north via Panama and Costa Rica to Ecuador in the South to the Galapagos Islands in the west (Kreipl and Poppe in De Turck et al. 1999). Its northern boundary coincides with the cool California Current in the north and the Peru Current in the south. Sea surface temperature maps indicate that the 20–21°C isotherm is an important barrier for this species as well. The temperature requirements of both extant species, which are genetically separated for several million years, are thus very similar. Therefore, the limiting sea surface temperature for extinct species of Persististrombus may have ranged around the same level with 16°C as absolute minimum and ~20°C as realistic scenario.

Conclusions

Strombids are frequent fossils in shallow water deposits worldwide. Inadequate taxonomic concepts completely camouflaged their biostratigraphic significance in the European Miocene. Mediterranean and Paratethyan carbonate platforms often lack adequate microfossils for biostratigraphic datings and are far from being stratigraphically well resolved. The application of the herein presented taxonomic concept will allow at least a quick separation of Lower, Middle and Upper Miocene deposits.

The frequent confusion of Miocene Paratethyan strombids with the Pliocene P. coronatus in the literature was based on the spiny morphologies. Such spiny morphs formed independently several times also in other strombids (see e.g., Landau et al. 2011). The first developed already during the Chattian (P. praecedens), the second during the early Langhian (P. lapugyensis) and the third one during the Serrvallian (P. pannonicus). A fourth yet undescribed group arose during the Serravallian in the Turkish Karaman Basin. These taxa are stratigraphically and geographically disjunct and have a distinct set of morphological characters, documenting an iterative development. This iterative evolution of highly reminiscent morphologies is also evident for the geographically and stratigraphically strongly separated specimens of the extant P. granulatus from Panama and its look-alikes from the Burdigalian of France (cf. Lozouet and Maestrati 1986). The distribution of the taxa is a good example for the outcrop-area effect on biodiversity and biogeography estimates. An extraordinary wealth of Langhian and Serravallian fossils in the area of the former Paratethys Sea is contrasted by relatively few coeval faunas in the Mediterranean area. This imbalance accounts for the impression that the Persististrombus inflexus-lineage bloomed only in the Paratethys Sea, although rare occurrences in the Middle Miocene of Turkey and the Loire Basin point to a much wider distribution.

The updated taxonomic concept allows utilizing ecological and climatic data of extant congeners which indicate that the 20°C was the limiting temperature for the successful distribution of all Persististrombus species. Minimum sea surface temperatures of 16°C were probably the final barrier for reproduction.