The NAFB is a classical foreland basin situated at the northern margin of the Alps and belongs to the biogeographic entity of the Paratethys (Fig. 1A). It extends from western France (Haute-Savoie) through Switzerland, South Germany to Austria. The thickness of the eroded sediments deriving from the uplifting Alps can reach up to more than 4 km near the mountains (e.g., Lemcke, 1988). In Germany, these deposits are divided into five main units based on depositional environments: the Lower Marine Molasse (UMM; Note that here we use the German abbreviations: Untere Meeresmolasse; Early and Middle Oligocene); the Lower Freshwater Molasse (USM; Untere Süßwassermolasse; Late Oligocene-Early Miocene) rich in terrestrial fossils (e.g., Werner, 1994), the Upper Marine Molasse (OMM; Obere Meeresmolasse; Early Miocene); the Upper Brackish-water Molasse (OBM: Obere Brackwassermolasse) and the Upper Freshwater Molasse (OSM; Obere Süßwassermolasse; Fig. 1C). The latter two units are the main subject of this work. During the Alpine nappe tectonism, the molasse sediments were thrusted, and the deposits piled up in a stack of southward dipping thrust sheets at the northern margin of the mountains, the Folded Molasse (Fig. 1B).

During the OMM (Ottnangian), the Paratethys was connected with the western Mediterranean Tethys by the “Burdigalian seaway”, and a basin-wide transgression flooded the NAFB (e.g., Rögl, 1998). While the marine conditions continued in the Late Ottnangian of Switzerland, the southwestern German Molasse basin was affected by a regression during the Middle Ottnangian (e.g., Lemcke, 1988). The regression resulted in the development of the Graupensand River, which eroded the underlying OMM and USM (Graupensandrinne; e.g., Doppler et al., 2005), and the channel was infilled by the Grimmelfingen Formation, which contains land mammals in its basal part (e.g., Sach and Heizmann, 2001). The uppermost part of the overlying Kirchberg Formation, probably a short transgression of the Swiss Molasse Sea (e.g., Doppler, 2011, Lemcke, 1988 and Reichenbacher, 1989), yielded the oldest Megacricetodon remains of the area (Reichenbacher et al., 2004).

Following the OBM, the OSM is characterized by strong temporal and lateral variability of continental facies; while radial alluvial fan sedimentation was prevalent in the southern/southwestern part of the basin, east/west fluvial deposits accumulated in the northern part leading to strong differences in the division of lithostratigraphic units from east to west (Abdul Aziz et al., 2008, Abdul Aziz et al., 2010, Doppler, 1989 and Doppler et al., 2005 and references therein). In the northern realm, some specific deposits, such as freshwater calcareous muds, with clear influence from Swabian Alb are termed as “sub-Jurassic”, and contain remarkable fossil mammals like hominoids (Böhme et al., 2011, Heizmann and Begun, 2001 and Ziegler, 1995). Finally, important marker beds, derive from the consequences of a meteor impact (see below) and volcanic activities that produced several layers of bentonite, clays formed by the weathering of volcanic ash in a water-bearing environment. At least four of these volcaniclastic horizons are reported from Switzerland, and three of them have been radiometrically dated at 14.20 ± 0.08, 14.91 ± 0.09 and 15.27 ± 0.12 Ma (Ub-isotope techniques). In South Germany, at least four horizons have also been dated (⁴⁰Ar/³⁹Ar) at 16.10 ± 0.20, 15.6 ± 0.40, 14.55 ± 0.19 and 14.50 ± 0.20 Ma (Abdul Aziz et al., 2008, Abdul Aziz et al., 2010 and Rocholl et al., 2008).

The Swabian Jura and its eastern continuation (Franconian Jura) are part of the Jura Mountains and are separated by the Nördlinger Ries meteor crater (see below). The Jura plateau delimits the northern part of the Molasse basin in our study area (Fig. 1A). Formed during the Jurassic period, three layers of limestone are defined as the Black, Brown and White Jura. Because of its extremely high concentration of calcium carbonate, the White Jura is karstic. Fissures and cave infillings are sporadically rich in fossils so that they attracted the early mammal palaeontologists and provided spectacular specimens (e.g., Dehm, 1935 and Schlosser, 1902) that range from the Oligocene to the Quaternary (e.g., Rummel, 1993, and references therein). Some quarries provided an incredible amount on fossil-enriched fissures, such as the White Jura near Petersbuch (Fig. 1D) were more than 120 fossil-bearing infillings are documented. These often contain many small mammals as the result of the accumulation of raptor pellets (e.g., Bolliger and Rummel, 1994, Klietmann, 2013, Klietmann et al., 2015, Prieto and Rummel, 2009a, Prieto and Rummel, 2009b, Prieto and Rummel, 2009c and Rosina and Rummel, 2012, and references therein).

The Nördlinger Ries is 24 km in diameter, whereas the Steinheim basin, formed by the impact of the smaller asteroid, is less than 4 km in diameter. This is one of the best-preserved impact structures of a binary asteroid. The age of this event is a matter of debate, but the impacts range between 14.6 and 15 Ma (e.g., Buchner et al., 2013 and references therein). Herein we follow Rocholl et al. (2011) who propose 14.94 ± 0.07 Ma for the Ries event.

Among the Ries impact ejecta, the distal ones are excellent marker beds. Isolated boulders mainly consisting of Upper Jurassic limestones are dispersed along the OSM in the form of large-sized “Reuter‘scher Block” (e.g., Sach, 2014a and references therein), while smaller pieces are found in the so-called Boulder Horizon (Brockhorizont; e.g., Sach, 1997, Sach, 2004 and Sach, 2014b) up to 180 km away from the crater (Fig. 2).

Both impact structures were filled by meteoric water forming an oligotrophic lake for the Steinheim basin and a shallow evaporitic saline lake for the Ries crater (see Tütken et al., 2006 and references therein). The deposits from Steinheim have provided an important and large collection of fossil animals and plants (Heizmann and Reiff, 2002 and Tütken et al., 2006). In particular, the historical study based on the planorbid snails (Hilgendorf, 1867) is seen today as one of the first fossil evidences of Darwin's evolutionary theory (Rasser, 2013, and references therein). The Ries lake deposits are much poorer, with the notable exception of fossil birds. Locally, important fossil-enriched layers occur, such as in the clays from Wemding, which contain a rich fauna including insects (e.g., Bolten et al., 1976). Similarly, the fissure and pocket fillings at the top of the spring mounds (travertines) may contain rich and extremely well-preserved small mammals, such as at Goldberg and Steinberg (e.g., Heizmann and Fahlbusch, 1983, Rachl, 1983 and Ziegler, 1983).

Maar lakes are characterized by anoxia that is a prerequisite for an exceptional preservation potential for fossils, as shown, for instance, by the Eocene Eckfeld and Messel localities. Within the scope of this paper is the Randecker Maar, which has been intensively studied by Rasser et al. (2013 and references therein; Fig. 1 and Fig. 2). It belongs to the Urach-Kircheim volcanic field. It has a diameter of ca. 1200 m and a maximum depth of 130 m. Of the 363 taxa identified, plants and insects are the most represented groups. The mammal fauna correlates to the OSM C + D (see below).

This group is used in the sense of Oliver Pérez and Peláez-Campomanes (2013) and Oliver Pérez (2015). It thus includes Megacricetodon aff. collongensis, M. bavaricus, M. aff. bavaricus, M. bezianensis, M. lappi, M. aunayi and M. vandermeuleni, of which the latter two have not been yet recognized from the NAFB. Berger (2010) assigns some teeth from Georgensgmünd to M. bourgeoisi, a species originally only known from France. This is a feasible proposal and it deserves attention, but the taxonomic assignment in our opinion lacks justification, and so does the proposed synonymy of M. bourgeoisi with M. bezianensis. Pending a detailed comparison between these taxa, we prefer to retain these species in order to simplify the biostratigraphy. In Germany, the lineage M. aff. collongensis–M. lappi is mainly characterized by a size increase through time (Fig. 3; e.g., Abdul Aziz et al., 2010), and the limit between the closely related species is hard to define. The forms closer to M. lappi show a stronger lingual anterolophid than in the older species (Prieto et al., 2009). In the present work we use the size of the m1 in order to characterize the species occurring in different biostratigraphic units. We do not consider valid the proposal of Aguilar (1995) of a M. bavaricus–M. germanicus lineage. Mein and Freudenthal (1971a) defined the subgenus Collongomys for the species M. lappi. This proposal was mainly based on differences in mandibular characters, but has not been followed afterwards. Indeed, a comparison with mandibles of M. bavaricus is needed (M. Freudenthal, pers. com. 2012), and it could have consequences for the traditional arrangement of the lineage.

Prieto and Rummel (2009a) describe Megacricetodon aff. germanicus from Petersbuch 68. They recognize a strong resemblance with the species from Zeglingen as stated by Kälin, 1993 and Kälin, 1997b. The same year, Kälin and Kempf (2009) emended their original assignment into M. gersii without detailing their reasons. Afterwards, Maridet and Sen (2012) provided new data on the material of the type locality for this species (Sansan, France) and they further showed the similarities in size between the two species. Their morphology is also similar. As a result, the German specimens are assigned to M. gersii. In many aspects, M. gersii resembles M. germanicus from Anwil, and it may be justified to include them in the same lineage. On the other hand, M. gersii and M. germanicus co-occur in a single Swiss locality, namely Grat 930 m (Kälin and Kempf, 2009), together with two other small-sized Megacricetodon species. Unfortunately, a detailed description is not provided, and these data likely derive from unpublished new material because Bolliger (1992a) does not report such a diversity of Megacricetodon in the locality. Thus we cannot judge whether M. gersii and M. germanicus are sympatric species (and define a lineage) or not in the NAFB. In addition, Prieto (2007) described the Megacricetodon material from different fissure fillings from Petersbuch. Especially in Petersbuch 6 and 18, he recognized m1 morphotypes that better allocated the specimens to the “M. germanicus group”, while other teeth correspond to a M. gregarius-like species. These differences are not seen in the other molars. Finally, a new Megacricetodon was recently described by Oliver Pérez in his thesis (2015) based on material from the late Middle Miocene of the Calatayud Basin (Spain; Las Umbrias 22; see García-Paredes et al., 2016). The species differs from M. gersii mainly by its larger size, while it is smaller than M. germanicus. The inclusion of this form in the “M. germanicus group” should be studied in greater detail.

Megacricetodon m1s with a clearly divided anteroconid are easily recognizable in the NAFB, yet rarely recorded, most probably because of their short stratigraphic range (Fig. 4E). M. gregarius-like species are only reported in Germany from Steinheim and Petersbuch 31 (e.g., Prieto, 2007; Fig. 4F). As noted above, some m1s from Petersbuch 6 and 18 show also a clearly divided anteroconid. Prieto (2007) concluded that the Megacricetodon material from these localities belonged to one highly variable species (M. aff. germanicus). Indeed, a revision of these forms is needed, as the presence of two species could not be discarded. First of all, the assignment of the remains from Steinheim and Petersbuch 31 has not been, in our opinion, sufficiently discussed, and the same applies to the Swiss material (Bolliger, 1994a, Bolliger, 1994b and Kälin et al., 2001). The “M. gregarius” morphotypes from Petersbuch 6 and 18 differ slightly from the species from Steinheim, and are assigned as M. cf. gregarius.

The low-crowned species C. transversus was defined by Heissig (1995) on the basis of a few molars from Steinheim. Prieto and Rummel (2009c) added C. gudrunae from Petersbuch 31 to this genus. Collimys gudrunae, though also low-crowned, is clearly larger than C. transversus. An intermediate form is recognized in Petersbuch 26 and currently assigned as C. aff. transversus. These three forms represent a single lineage.

A questionable record of C. gudrunae corresponds to a single m1 from Petersbuch 18. Kälin and Kempf (2009) reported Collimys in Anwil. Prieto (2007) and Prieto and Rummel, 2009c and Prieto and Rummel, 2009d did not identify the genus in neither Kleineisenbach nor in Giggenhausen, perhaps due to its strong resemblance with Democricetodon freisingensis. Indeed, one single tooth from Kleineisenbach might be attributed to Collimys (Fig. 4A), but is not included in the “C. transversus group”.

In Germany, and possibly in the NAFB as well, the lineage Collimys hiri–C. longidens–C. dobosi is mainly marked by size increase of the molars (Prieto and Rummel, 2009d) (Fig. 4B–D). In addition, Prieto et al. (2011) included a single m1 from Aumeister in this group. The anagenetic trend in size increase of the molars led Prieto and Rummel (2009d) to propose that the lineage is of great use for biostratigraphic purposes at the transition between the Middle and the Late Miocene. In contrast, this has been nuanced by Prieto et al. (2014) who suggest that given the present state of knowledge these taxa should not be used for long-distance correlations. Finally, the Collimys molars from Mörgen are relatively small and show a trilobed anteroconid (Fig. 4B). This character is observed in some specimens of C. dobosi (Hír, 2005).

The study of the Lartetomys material is currently ongoing by the authors and colleagues. Initial results indicate that two species are present. Lartetomys (= Mixocricetodon) dehmi is present in the NAFB and fissure fillings (e.g., Ziemetshausen 1b, Petersbuch 39; Rummel, 1997). A second smaller species is scarcely recorded in Petersbuch 41 and Burg-Balzhausen (Rummel, 1997 and Seehuber, 2009). The reader is referred to Prieto (2012a) and Prieto and Scholz (2013) for useful references regarding the species assigned to Lartetomys and Karidomys, and a discussion of previous synonymies.

Mirrabella is one of the rarest genera in the Miocene of Germany. A single M1 is reported from Rembach (Ziegler and Fahlbusch, 1986; as Cricetidae n. g. n. sp., Plate 10, fig. 2; see also De Bruijn and Saraç, 1992: Plate 13, fig. 3). De Bruijn et al. (2013) ascribe this specimen to Mirrabella cf. tuberosa, a species only recorded in the Early Miocene from Greece and Turkey.

There is no fauna equivalent to the Swiss Democricetodon franconicus–Megacricetodon collongensis interval zone (Glovelier) in the German molasse because the deposits of the OMM have not provided any small mammals. In contrast, the German fissure fillings from Petersbuch and Erkertshofen have provided a nice picture of the small mammal faunas of this period (Fahlbusch, 1966, Fahlbusch, 1970, Heissig, 1978, Rummel, 1993, Wu, 1993, Ziegler, 1989 and Ziegler and Fahlbusch, 1986).

In Germany, the differences between OSM A and OSM B are indeed characterized by the presence/absence of some eomyid rodent taxa. Kälin and Kempf (2009) remark that the first occurrence (FO) of Keramidomys in Switzerland is not contemporaneous with the last occurrence (LO) of Ligerimys, because both genera are never found in association, and because some localities from the same period do not record any Eomyidae at all. However, both genera occur in Germany. First, Sach and Heizmann (2001) listed them in Langenau 1. The size of the Megacricetodon, very small in this locality, is not in agreement with the expected size at the initial occurrence of Keramidomys. The fossil material has never been described, and our first observations (JP) show that the presence of Keramidomys is indeed doubtful. Secondly, Prieto (2015) reviewed the material from Eitensheim, originally assigned to Keramidomys n. sp. by Fahlbusch (1975). Indeed two genera are present in the collection, namely Keramidomys and a small Ligerimys species. However, we cannot exclude that the material of this site mixes fossils of slightly different ages, and question the biostratigraphic value of such purported co-occurrence. In Germany, Ligerimys, as well as Melissiodon, become very rare near the end of the OSM A (herein as OSM A-REI). No Eomyidae are recorded in Reisensburg (Seehuber, 2015), but the locality contains Melissiodon. In contrast, a single tooth of Ligerimys is present in Günzburg 2. The first occurrence of Keramidomys is in Pöttmes (e.g., Prieto, 2010). As a consequence, some localities lacking eomyids are difficult to assign to a given OSM zone.

Heissig (1997) proposed that the German OSM B is defined by M. bavaricus and the first occurrence of Democricetodon gracilis, D. mutilus and Anomalomys minor. However, D. gracilis is present at the end of the OSM A (Forsthart, Reisensburg; see above). Regarding D. mutilus, Seehuber (2015) could not discard the presence of the species in Reisensburg. Therefore, we do not consider it a good index for OSM B. The stratigraphic range of A. minor has been emended by Böhme et al. (2003) and does not allow discrimination of OSM A from OSM B. For this reason, it is considered herein that the older documented occurrence of Keramidomys (Pöttmes, Megacricetodon m1 mean length 1.65 mm) should be used at present as reference for the base of the OSM B. Similarly the discrimination between OSM B and C + D deserves some comments. Heissig (1997) recognized an increase in Megacricetodon size in OSM C, as well as the first occurrence (FO) of M. minor-like species and the FO of Neocometes similis. The oldest locality containing M. minor is Roßhaupten, but the species is only documented by a single m2 (Fahlbusch, 1964). The localities Engelswies/Schellenberg and Untereichen 540 m lack the small Megacricetodon, while Neocometes is present in Engelswies/Schellenberg (Ziegler, 1995). The latter localities are thus considered the oldest belonging to OSM C + D in southern Germany.

Originally, Heissig (1997) discriminated the OSM C and the OSM D based on the presence of two species of Anomalomys. This argumentation did not convince Böhme et al. (2003), who merged the two units. In contrast, Kälin and Kempf (2009) still keep them separated, considering as an important criterion the presence/absence of a (new) small-sized Megacricetodon species. This leads to some problems. Indeed, no sized-based criterion allows for clear distinctions between M. bavaricus and M. aff. bavaricus. Thus, the Swiss Megacricetodon aff. bavaricus–Megacricetodon cf. minor interval zone of Kälin and Kempf (2009) can also be referred to the OSM B in Germany. Indeed, prior to Oggenhof, the (very) small-sized Megacricetodon species (Fig. 6) is extremely rare although it occurs in some Bavarian localities. As a result the Swiss M. aff. bavaricus–M. cf. minor interval zone is not convincingly recognized in Germany. In addition, correlating this interval zone to the OSM C as defined by Heissig (1997), also implies a correlation to Sandelzhausen (reference fauna), which is not likely if one takes into account the faunal content (high number of small-sized Megacricetodon minor; Wessels and Reumer, 2009). Molar size comparable to Megacricetodon minor is not achieved in localities preceding Cricetodon occurrences (OSM E, see below; Fig. 6). Prior to that, only a few Bavarian localities provided a representative molar sample of the smaller form (Sandelzhausen, UA 565 m), thus limiting this analysis. The Swiss Megacricetodon cf. minor–Megacricetodon aff. bavaricus overlap zone has as reference locality Tobel Hombrechtikon. Members of the “M. bavaricus group” are missing. However, this could be the result of sampling bias. This assumption is supported by the abundance of M. minor in Sandelzhausen. Indeed, the larger-sized Megacricetodon species were not recognized by Fahlbusch (1964), but after intensive sampling, the species was reported by Wessels and Reumer (2009). As a result, the German OSM C + D-SCH correlates most probably to the Megacricetodon aff. bavaricus–Megacricetodon cf. minor interval zone from Switzerland, but this assumption is only based on the size of representatives of the “M. bavaricus group”. In southern Germany, the OSM C + D-SAN correlates to the Swiss Megacricetodon cf. minor–Megacricetodon aff. bavaricus overlap zone. Anomalomys minutus from Tobel Hombrechtikon (Switzerland) is only a bit smaller than A. minor and no clear morphological differences are observed (Bolliger, 1992a). We thus agree with Böhme et al. (2003) in not considering this taxon as relevant for biostratigraphic purposes. The absence of Megacricetodon in the Randecker Maar precludes any precise correlations (Rasser et al., 2013), but the presence of Keramidomys and Neocometes supports a correlation of the locality to the OSM C + D.

The occurrence of Cricetodon in OSM C + D is reported by Heissig (1997) but without details, so this occurrence is not considered herein (see comments in Prieto and Scholz, 2013: 306). Seehuber (2009) correlates Hohenraunau (Germany) to the OSM F, but reports a single M1 of Megacricetodon lappi. This specimen was found 5 m under the lowest reworked Boulder Horizon (Seehuber, 2009: 22). In contrast, the Cricetodon remains derive from reworked sediments and are thus younger. (Seehuber, 2009: 220) could not exclude that the fauna is mixed. For this reason, the locality is not considered to be relevant here. A representative sample of Cricetodon meini has been recovered in Burg-Balzhausen (Germany). That these are almost the smallest representatives of the group in the NAFB lead Seehuber (2009) to correlate the locality to the OSM E. Surprisingly, the locality lacks M. lappi, although numerous fossils have been recovered. Here we preliminarily consider the absence of this species as a sampling effect (see above: comments on Megacricetodon aff. bavaricus from Sandelzhausen). Similarly, the absence of Cricetodon in Mohrenhausen (Germany) is hard to explain. Kälin and Jost (2014) argue that OSM E-EBE equivalents have not been yet discovered in Switzerland. Indeed, M. lappi is smaller there and the localities lack Cricetodontini (equivalent to OSM E-UA565 m).

The German unit OSM E’ is briefly described in Böhme et al. (2003) who noted that it differs from OSM E only by a cricetid rodent species, called in that work Megacricetodon aff. gersii, but the authors do not provide further details on this taxon. This unit is restricted to Derching 1b, a locality that has not been well studied although it is considered to just precede the Ries impact. In the context of this work, the small mammal material from Derching has been studied, and we have not found cogent arguments to attribute this material to M. aff. gersii, but consider it a somewhat large M. minor (see Fig. 6). For this reason, the locality is included in OSM F herein. Heissig (2006) introduced OSM?F for Laimering 5 (Germany) on the basis of differences in the sedimentary cycle. From a biostratigraphic point of view, there are no differences with the OSM F species and we place the locality into this unit. Kälin and Kempf (2009) correlate the Swiss Megacricetodon lappi–Democricetodon gracilis interval zone, the Democricetodon gracilis–Megacricetodon gersii interval zone and the Megacricetodon gersii–Megacricetodon similis interval zone to the OSM F. The Megacricetodon lappi–Democricetodon gracilis interval zone records the first entry of Cricetodon in the Swiss part of the NAFB. As shown above, the genus occurs earlier in the German part (OSM E-EBE) but these faunas are not documented in Switzerland. Kälin and Kempf (2009) also recognized Anomalomys cf. gaudryi as an important newcomer into this unit. A revision of the German Anomalomys is indeed necessary to decide on the relevance of the genus for biostratigraphic purposes in Bavaria. However, Abdul Aziz et al. (2008) recognized the FO of A. gaudryi at the base of the OSM F in the same area. The characteristic occurrence of Karydomys/Lartetomys in Switzerland is also observed in Germany. The following Swiss interval zone (Democricetodon gracilis–Megacricetodon gersii interval zone) differs basically from the former by the LO of D. gracilis. Notably, eomyids and Anomalomys are completely missing. The LO of D. gracilis is at present not well defined in Germany. Finally, the Swiss Megacricetodon gersii–Megacricetodon similis interval zone is characterized by a species assemblage strongly resembling the fauna of the French site Sansan. In Germany, the distinction between C. aureus and C. sansaniensis is not always clear in the literature and the recognition of the species in the NAFB is in general a matter of debate. Rummel and Kälin (2003) assigned the collection from Zeglingen to Cricetodon sansaniensis, and preliminarily attributed some other Swiss material to this species (see also Bolliger, 2000). They also discussed that these forms differ slightly from the type material in size and morphology. Regarding the German record, they consider the fossils from Petersbuch 31 and Kirrberg Tongrube as C. cf. sansaniensis (see also Rummel, 2000 and Seehuber, 2009). Some other German occurrences are more problematic. Cricetodon sansaniensis is reported in Höll (Scholz, 1986a and Scholz, 1986b), but this sample has not been re-studied. Only the measurements of a single specimen have been given (Heissig, 1990; see also comments in Prieto and Scholz, 2013). The large collections from Steinberg and Goldberg, originally assigned to this species by Heizmann and Fahlbusch (1983), are now thought to belong to C. aff. aureus (Rummel, 2000). Cricetodon sansaniensis is described from the Bohlinger Schlucht (Giersch, 2004). The fauna contains both M. minor and M. aff. gersii and is tentatively correlated to the Swiss Megacricetodon gersii–Megacricetodon similis interval zone herein (OSM?-BOH).

Even more tentative is the correlation of the Kirrberg faunas (Kirrberg Tongrube and Kirrberg Sandgrube, Bavaria) to the Swiss Megacricetodon similis–Megacricetodon gregarius interval zone. It is based on the presence of Cricetodon cf. engesseri in these two sites (Seehuber, 2009), a rare species at present only reported in Switzerland from this interval zone from the site Chräzerentobel 505 m. The composition of the Kirrberg faunas allowed Seehuber (2009) to exclude a correlation to younger assemblages.

The Megacricetodon species with clearly divided anteroconid (M. gregarius-like; see taxonomical notes) is considered by Kälin and Kempf (2009) an excellent index fossil in the Swiss record. They extend their analysis to Steinheim and conclude that the presence of the species represents a very short interval, named as “gregarius event” by Heissig (1997). The species is, according to this scheme, never found in co-occurrence with Deperetomys, an index taxon for the following Swiss biostratigraphic unit (Anwil). In contrast to Steinheim, some German localities cannot be so easily correlated. In Bavaria, the sites from Derndorf, Tiefenried and Eppishausen are close in age, but not rich in small mammals. These localities postdate the FO of the pig Listriodon (Seehuber, 2009), which is seen also as a newcomer in Helsighausen (Switzerland; Kälin and Kempf, 2009). The first occurrence of Listriodon splendens in Central Europe occurs at around 14.2 Ma in Klein-Hadersdorf (Austria; Böhme et al., 2012). In Germany, Seehuber (2009) considers the material of this species from Stätzing to be probably older than 11.4 Ma. Furthermore, Deperetomys may be present in Tiefenried, indicating an earlier entry of the genus in the NAFB than generally accepted (Prieto, 2012b and Seehuber, 2009). The German fissure filling Petersbuch 31 (upper part) contains M. cf. gregarius (Rummel, 1997; see Prieto, 2007, figs. 36F and I), but the species is found in association with Cricetodon sansaniensis, which disappears earlier in Switzerland. Petersbuch 6 and 18 may contain two medium-sized Megacricetodon species that can only be distinguished on the basis of the m1. One resembles M. gersii or M. germanicus, and the second one shows a better divided anteroconid (Prieto, 2007, see taxonomical notes). These fissures contain also some rare Deperetomys teeth. And finally, Prieto and Rummel (2009a) consider Petersbuch 68 to be older than Steinheim on the basis of the size of M. gersii. Among the Cricetodontini, the presence of rare Cricetodon engesseri molars and the absence of Collimys agree with this correlation. A very variable Deperetomys is also present and abundant (MR and JP, pers. data). Finally, the fissure fillings of Petersbuch 35 and 48, rich in Deperetomys, are considered to be slightly older than Anwil (Prieto, 2007). To sum up, it can be concluded from these localities that the faunal succession around the “M. gregarius event” is characterized by intense but short dispersal waves that complicate the correlation of the faunas. Herein, we propose the succession of Table 1 and Fig. 6, but emphasize that it is not well supported for the faunas near in age to Steinheim.

As shown above, the Deperetomys hagni taxon range zone as defined by Kälin and Kempf (2009) for the Swiss molasse might be in need of revision in the view of the German record. In addition, they noticed that the genus is absent in the younger localities assigned to MN 8. In any case, the reference locality Anwil is the Swiss equivalent of Kleineisenbach and Giggenhausen (Prieto, 2007).

Clear similarities are observed between Nebelbergweg (Switzerland) and Hammerschmiede (Germany), although the Swiss locality lacks Microtocricetus (Kälin and Engesser, 2001). The genus is recorded in Nebelberg TGL localities (Switzerland). Recent advances show that the rich and diverse faunas from Hammerschmiede can be correlated to the base of the Tortonian, at 11.6 Ma, while Hillenloh is probably a bit younger, with an estimated age of 11.3 Ma (Fuss et al., 2015, Kirscher et al., 2016 and Prieto et al., 2011). The relative age of Nebelbergweg is indeed much more controversial, and the argumentation proposed by Prieto and Rummel (2009d) based on Collimys has to be critically evaluated (Prieto et al., 2014). In this context, the German locality Mörgen is highly relevant. Seehuber (2009) places it between Anwil and Nebelbergweg, and excludes a correlation to MN 9 based on the presence of Miodyromys aegercii, Euprox furcatus and Dorcatherium crassum. He recognizes some uncanny similarities with Nebelbergweg. Both localities share a very large Miodyromys species, as well as similar Democricetodon species and Collimys of the “C. longidens group”. In Mörgen, two Megacricetodon forms are recognized although they have not been identified to the species level. The smallest one resembles in many points M. minutus, while the larger is characterized by an m1 with a well-divided anteroconid. Seehuber (2009) refrains from ascribing it to M. robustus, a species endemic to Nebelbergweg, mainly because of the presence of a well-developed anterolophulid on the German teeth. The specimens from Mörgen belong in our opinion to the same Megacricetodon grade as M. robustus or M. crusafonti. The tribe Cricetodontini is not reported from Nebelbergweg, but the group is represented in Nebelberg by Hispanomys sp. (Kälin and Kempf, 2009). Cricetodon sp. from Mörgen also resembles this form, being at the same time similar to C. nievei and early Hispanomys species. Recently, a single tooth of Microtocricetus has been found in this German locality. It is an M2 that fits within the lower size range of Microtocricetus molassicus (1.72 × 1.33 mm; see Kretzoi and Fejfar, 2005). From a morphological point of view, this specimen differs from the type material in its reduced lingual anterocone (Fig. 4F; nomenclature follows Daxner-Höck and Höck, 2015). These data link strongly Mörgen to the Swiss localities, but their biostratigraphic arrangement is still unclear. As noticed above, the proposal of Prieto and Rummel (2009d) that Hammerschmiede is older than Nebelbergweg is open to discussion. On the other hand, the lack/rarity of Microtocricetus, might imply, among other possibilities, an older age for the Swiss site. Similarly, the German locality Marktl (Bavaria) is considered to be older than Hammerschmiede, although this conclusion is based on third molar morphology of Microtocricetus (Fahlbusch and Mayr, 1975). This conclusion is also partly followed by Kowalski (1993). In addition, the locality also contains the large-sized eomyid Eomyops sp., a species currently not recognized elsewhere in Europe (Prieto, 2012a). Finally, the presence of Collimys cf. longidens and Democricetodon n. sp. in Chräzentobel 655 m allows for a clear correlation of the Swiss fauna to the localities mentioned above. In conclusion, the relative ages of the localities proposed here is a working hypothesis, requiring of additional material for many sites and if possible of independent stratigraphical data.