Introduction

Ammonites, that is, representatives of the order Ammonitida (subclass Ammonoidea, class Cephalopoda), are iconic marine invertebrates of the Mesozoic Era. They provide a text-book example of victims of the end-Cretaceous mass extinction. However, data amassed lately suggest that some ammonite populations from the Netherlands, Denmark and USA (New Jersey) briefly survived Cretaceous–Paleogene (K–Pg; Maastrichtian–Danian) boundary perturbations1,2,3,4,5,6,7,8,9,10,11. Yet, the survival hypothesis is still regarded by some as controversial12 or questionable13.

The present study deals with ammonite specimens from the lower Danian Cerithium Limestone Member (Rødvig Formation) as exposed at Stevns Klint, Denmark (for lithostratigraphy, see Surlyk et al.14). This regional K–Pg boundary succession is famous for its excellent exposure, rich fossil content and presence of geochemical signatures of K–Pg boundary events14,15,16,17,18,19,20,21. For these reasons, and also in view of its natural beauty, the Stevns Klint area has been declared a UNESCO World Heritage Site22.

The Cerithium Limestone Member is the first carbonate unit of the Paleogene portion of the Stevns Klint succession. It has long been known to record the earliest phases of the biotic recovery in shallow-marine environments after the end-Cretaceous mass extinction4,23,24,25,26,27,28,29,30,31.

The hypothesis that ammonite fossils from the Cerithium Limestone Member are in fact Danian survivors was suggested by Surlyk and Nielsen2, first criticised by Machalski32 and subsequently supported by Machalski and Heinberg4 and Machalski et al.5,6. Earlier, these fossils had routinely been regarded as redeposited (remanié) Maastrichtian fossils33,34,35,36. Indeed, the occurrence of the Cerithium Limestone in a series of disjunct basins between eroded crests of the underlying Maastrichtian bryozoan mounds, made the redeposition hypothesis worthy of careful consideration as an alternative for the survival hypothesis32. Strong ichnological and early diagenetic overprinting of the K–Pg interval at Stevns Klint adds to difficulties in its interpretation23,24,34,37,38.

Machalski and Heinberg4 proposed that ammonites did survive K–Pg boundary perturbations on the basis of specimens recovered mainly as a byproduct of processing of Cerithium Limestone bulk samples in search of minute bivalves23,24. Their study material also included some incidental outcrop finds, housed in museum collections, with limited documentation. Here, we reassess the survival hypothesis based on a much wider range of new observations, chiefly of a taphonomic and sedimentological nature, gained by us during fieldwork at the localities of Sigerslev and Rødvig along Stevns Klint. That project was initiated in co-operation with eminent Danish palaeontologist Claus Heinberg (1945–2021), who worked on bivalve turnover across the K–Pg boundary39; the present paper is dedicated to his memory.

Regional background

The study area is located south of Copenhagen (Sjælland, north-eastern Denmark; Fig. 1a, b). A belt of outcrops is located along the coastal cliff called Stevns Klint. The coastal profile is 14.5 km long and up to 41 m in height14. During the Cretaceous to Paleogene transition, the Stevns Klint area was located in a cool-water carbonate ramp setting with facies shifts controlled by sea level oscillations18,40.

The upper Maastrichtian to lower/middle Danian strata at Stevns Klint were described in detail by Surlyk et al.14 and Rosenkrantz et al.41 (their Fig. 4). In brief, the succession starts with white chalk of the upper Maastrichtian Sigerslev Member, which is truncated by a hardground with Thalassinoides burrows. This is sequence boundary 1 in Surlyk18, marked as SB 1 in Fig. 1c (based on Surlyk18 and Machalski et al.42). Higher up, there is a bryozoan-rich uppermost Maastrichtian chalk, up to 4–5 m in thickness, referred to as the Højerup Member or Grey Chalk (Fig. 1c, d). This is a highly fossiliferous unit, with chiefly benthic fauna dominated by calcite-shelled biota43. The Højerup Member is composed of low-amplitude bryozoan mounds (bioherms) (Fig. 1d). The shallow basins (in the Danish literature on Stevns Klint, understood to be interbiohermal depressions) between crests of these mounds are filled in with Danian deposits (Fig. 1c, d), starting with the lowermost Danian marly clay, named Fiskeler Member or Fish Clay14,37,44. This contains sparse fish debris and yields geochemical signatures of the Chicxulub impact at the base15. The Fiskeler is typically 5–10 cm thick and wedges out completely towards the basin margins14,37.

The overlying Cerithium Limestone Member owes its name to the presence of distinctive moulds of cerithiid gastropods. This is a 30–120-cm-thick unit, infilling major parts of the intra-mound basins (Fig. 1c, d). Although earlier workers regarded the Cerithium Limestone as a homogeneous micritic limestone unit, recent investigations have demonstrated its clearly bipartite nature, at least in some places30,31. At Rødvig, for instance, the lowermost part of the Cerithium Limestone is developed as a thin layer of bryozoan-dominated calcitic skeletal debris, thought to represent reworked Maastrichtian material, while the main body of the unit is composed of micritic limestone, representing autochthonous Danian sediment31.

Both the basins infilled by the Cerithium Limestone Member and crests of the Maastrichtian mounds are truncated by an erosional hardground with large systems of Thalassinoides burrows and glauconite mineralisation14,34,38. The hardground originated through sediment omission, cementation and erosion during an early Danian sea level drop (sequence boundary 2 in Surlyk18, SB2 in Fig. 1c). It is overlain by another set of bioherms45, composed of Danian bryozoan limestones of the Korsnæb Member (Fig. 1d). At their base, there is a transgressive lag composed of glauconitised and phosphatised pebbles and fossils (including ammonites) derived from underlying strata4,23,37,45.

The typical Cerithium Limestone fauna, that is the one from the main part of this unit as defined by Störling et al.31, is dominated by moulds of minute, originally aragonitic bivalves and gastropods23,24,29. The most distinctive feature of this fauna (and its equivalent at Nye Kløv, Jylland, northwestern Denmark) is that it is almost totally deprived of low magnesium calcite-shelled biota, such as brachiopods, pectinid and ostreid bivalves, serpulids and cyclostome bryozoans, while the high magnesium calcite-shelled biota are dominated by minute, non-echinoid echinoderms23,24,46.

In terms of planktic foraminifer stratigraphy25, the Fiskeler Member belongs to the P0 Zone, the Cerithium Limestone Member to zones Pα to P1a, and the overlying Korsnæb Member to zones P1b to P1c (Fig. 1c). With regard to ammonite zonation47,48,49, the Sigerslev Member belongs to the Hoploscaphites constrictus crassus Zone and the interval from the Højerup Member to the Cerithium Limestone Member to the H. c. johnjagti Zone (Fig. 1c).

Materials and methods

The ammonite specimens and samples studied stem from two localities along Stevns Klint (Fig. 1b). These are part of a large, disused chalk quarry at Sigerslev (Stevns Kridtbrud), and a coastal section between Rødvig and Korsnæb, referred to here as Rødvig. The most important specimens, and all geological samples analysed for microfacies, originate from the Cerithium Limestone Member exposed in a single basin at Sigerslev, referred to as the “studied basin” in Fig. 2a and in Table 1, but mentioned simply as the Sigerslev basin further on in the text, and from a basin at Rødvig located 200 m north of access to the beach.

Table 1 Summary of data on the Danian ammonite specimens studied.
Full size table

This collection was amassed during 2009 and 2013 fieldtrips and during subsequent fieldwork by staff members from the Østsjællands Museum at Faxe. All specimens studied have been deposited at this museum (collection acronym OESM). The ammonites were collected through careful hammering of the Cerithium Limestone so as to ensure that the specimens were derived from this unit proper and not from burrow fills or adjacent lithified Maastrichtian chalk. Precise provenance levels in relation to the reference levels (base of the Fiskeler, top of the Cerithium Limestone) were noted, whenever possible. Photographic documentation was done in the Photo Laboratory at the Insitute of Paleobiology of Polish Academy of Sciences (further on abbreviated IPAL) using reflex camera NIKON D-5 with NIKKOR 55/2.8 micro manual.

Seventeen samples were selected for preparation of thin sections for microfacies analysis, including 16 from Sigerslev and one from Rødvig. From the Sigerslev basin, we analysed seven thin sections from the Højerup Member, seven from the Cerithium Limestone Member, and two from low in the Korsnæb Member. A single thin section was analysed from the lowermost part of the Cerithium Limestone Member exposed at Rødvig. Thin sections were prepared in the Paleontology Laboratory at IPAL. The rock samples were sliced with a Buehler Isomet Low Speed saw with a diamond-embedded blade. In view of their soft and brittle nature, the samples were embedded in Araldite 2020 epoxy in a vacuum chamber to prevent damage. The slices were then glued to glass slides with Araldite 2020 epoxy. Subsequently, the samples were manually ground down and polished using Buehler Metaserv 2000 grinder.

The investigations of thin sections and photographic work were carried out at the Microscope Laboratory and Microanalysis Laboratory, University of Warsaw, using a Nikon ECLIPSE E600WPOL optical microscope equipment in camera Nikon Digital Sight DS-5Mc and NIS-Elements AR Software. The scan of an ammonite thin section was carried out at the high-resolution digital microscope KEYENCE VHX-7000. Microfacies were classified according to the Dunham classification50 and the modified scheme of Embry and Klovan51. Thin sections are housed at IPAL under collection number ZPAL Am. 27/1–17).

For Scanning Electron Microscope (SEM) observations, several rock samples from the Cerithium Limestone and adjacent units at or near the Sigerslev basin were placed on stubs with double-sided adhesive tape and sputter-coated with a conductive carbon film. Analyses were conducted in the Electron Microscopy and Electron Microprobe Laboratory at IPAL, using a Philips XL20 scanning electron microscope. The instrument was operated at an acceleration voltage of 25 kV, a beam current of 98–103 nA and a spot diameter of 3.5 μm.

Results

New ammonite specimens from Cerithium Limestone

We studied ten ammonite specimens from the Cerithium Limestone Member exposures at Sigerslev and Rødvig. We identified a single scaphitid, Hoploscaphites cf. constrictus johnjagti Machalski, 200548, and four baculitid taxa: Baculites vertebralis Lamarck, 180152, Baculites cf. vertebralis Lamarck, 180152, Baculites sp. and Fresvillia sp. (see Table 1 for concise summary of the material). Specimens of H. constrictus and B. vertebralis are well known from the uppermost Maastrichtian and Danian strata of Stevns Klint4,36,47,48. In contrast, our record of the genus Fresvillia Kennedy, 198653 is new for Denmark, and for the Danian worldwide. Reference is made to Ammonite data in Supplementary Information for more details on the specimens collected.

All specimens studied are internal or external moulds left after dissolution of the originally aragonitic ammonite conchs (Figs. 3 and 4a–c). This preservation is typical of all aragonite-shelled biota in the Danish succession23,36. One of the baculitid specimens displays a micritised replica of a septum (Fig. 3b).

The fragmentary preservation of the specimens studied may reflect physical damage at the outcrop (Fig. 3h), or mechanical or biogenic damage prior to the final burial of shells in sediment (Fig. 3g). Ammonites are often preserved as fragments at Stevns Klint, while all baculitid specimens from this area available in museum collections are fragments36. Amongst the studied lot from the Cerithium Limestone, the scaphitids are preserved as pieces of body chamber walls, while the baculitids are phragmocone fragments (Fig. 3).

Fig. 1
Fig. 1
Full size image

Geological setting of studied ammonite-bearing Danian deposits in Denmark. (a) Location of study area within Denmark (rectangled). (b) Southern portion of Stevns Klint cliff section with main sites having yielded Danian ammonite fossils4 (also the present paper). (c) Stratigraphical column of the Stevns Klint succession (based on Surlyk18 and Machalski et al.42). (d) Idealised section to show stratigraphy of the Cretaceous–Paleogene boundary interval at Stevns Klint (based on Machalski et al.6). K – Cretaceous, Pg – Paleogene, hg. – hardground, Mb – Member.

Fig. 2
Fig. 2
Full size image

Geographical and geological setting of the Cerithium Limestone Member basin studied at Sigerslev. (a) Aerial view of Sigerslev quarry (https://marinas.com/) with position of Cerithium Limestone basin studied. (b) Overall view of basin studied; rectangled shows approximate provenance sector of baculite specimen OESM 13290 (Fig. 3a, b, h, i). (c, d) Close-up view of section. (e) Post-omission burrow in Cerithium Limestone infilled with bryozoan-rich sediment from overlying Korsnæb Member. (f) Close-up view to show external mould of Metacerithium in the Cerithium Limestone.

Fig. 3
Fig. 3
Full size image

Ammonite specimens from the main part of the Cerithium Limestone Member (as defined by Störling et al.31) at Stevns Klint. (a, b, h, i) Baculites vertebralis (OESM 13290), top of unit; views h and i are field photographs. (c, d) Fresvillia sp., part and counterpart (OESM 13291). (e) Baculites sp. (OESM 13293). (f) Baculites vertebralis (OESM 11541). (g) Hoploscaphites cf. constrictus johnjagti (OESM 13287). Specimens in a–e and g–i are from Sigerslev quarry (a–d, g–i are from sampled basin, e from another basin); specimen in f is from Rødvig (from yet another basin than specimen illustrated in Fig. 4).

None of the ammonite moulds from Sigerslev show taphonomic signatures which would indicate their reworking and/or redeposition. Such features are present, for example, on a specimen of Hoploscaphites constrictus johnjagti Machalski, 200548, preserved in a worn, glauconite-coated clast of Maastrichtian chalk, from the base of the Korsnæb Member (Machalski and Heinberg4, their Fig. 4). There are also no traces of epizoans on the specimens studied, which would have indicated prolonged exposure periods of empty conchs, either on the sea-floor or when floating in the water column.

The presence of a void after dissolution of the aragonite shell in one of the baculitid moulds (Figs. 3e and 6a, S3c, d) indicates that the shell was still present at the final burial stage of the fossil, having dissolved only after lithification of the carbonate mud during hardground formation4,34,36.

Ammonite-bearing settings

The most meaningful data on local depositional settings of Cerithium Limestone Member ammonites have been gathered through field exploration and sampling of two basins, one at Sigerslev (“studied basin” in Fig. 2a) and the other near Rødvig. Below, we present these observations, following the bipartite subdivision of the Cerithium Limestone into the main part, encompassing the bulk of the unit, and the lowermost part just above the Fiskeler Member30,31.

Main part of the Cerithium Limestone (Sigerslev)

The Cerithium Limestone Member of the Sigerslev basin yielded five ammonite specimens. These represent Hoploscaphites cf. constrictus johnjagti (specimens OESM 13287 and OESM 13288; Fig. 3g), Baculites vertebralis (OESM 13290; Fig. 3a, b, h, i), Baculites sp. (OESM 13292) and Fresvillia sp. (OESM 13291; Fig. 3c, d). Of these, the best-preserved specimen, OESM 13290, was recovered from the very top of the Cerithium Limestone Member (Fig. 3h, i), while other specimens come from the upper dozen or so centimetres of that unit. This is definitively the “main part” of the Cerithium Limestone sensu Störling et al.31. In the Sigerslev basin, we failed to identify any deposit which would match the “lowermost part” of the Cerithium Limestone as understood by those authors.

The Sigerslev basin (Fig. 2b) measures c. 13 m in length (south to north), infilled with the lowermost Danian Fiskeler Member (5 cm in thickness), which is overlain, with sharp contact, by the lower Danian Cerithium Limestone Member, c. 26 cm thick. Both the basin and adjacent crests of the uppermost Maastrichtian bryozoan mounds (Højerup Member) are truncated by an erosional hardground surface, overlain by c. 2-m-thick bryozoan limestone of the lower Danian Korsnæb Member (Fig. 2b–d), covered in turn by Quaternary deposits. The Cerithium Limestone in the Sigerslev basin is a strongly weathered, pale-yellow limestone, apparently homogeneous on a macroscopic scale. It reveals a pseudo-nodular texture due to penetration by abundant glauconite-lined Thalassinoides burrows of the omission suite (sensu Bromley38. These burrows pipe down from the overlying Korsnæb Member, being filled with bryozoan debris (Fig. 2e, f). The limestone is intersected by stylolites and dissolution seams. Brown spherical bodies, up to 1 mm in diameter, interpreted as possible remains of prasinophyte algae17, are discernible on the rock surfaces (Fig. 3d, e).

In addition to ammonites, we recovered a few macrofossils from the Cerithium Limestone Member in the Sigerslev basin, all of them belonging to taxa with originally aragonitic shells. These are moulds of gastropods, including Metacerithium (Fig. 2f), small solitary corals and minute bivalves. Coral moulds are by far the most common macrofossils in studied deposits as we noted a total of 18 specimens. We failed to note any calcitic macrofossils in the Cerithium Limestone of the Sigerslev basin.

Underlying and adjacent Maastrichtian chalk (Sigerslev)

In contrast to the Cerithium Limestone Member, the uncemented chalk of the uppermost Maastrichtian Højerup Member exposed just below the Sigerslev basin yielded relatively common macrofossils, represented exclusively by taxa with originally calcitic shells. These are bryozoan fragments, cups (not moulds) of solitary corals, micromorphic brachiopods, shells of pectinid and ostreid bivalves, and spines (radioles) of regular echinoids, including the distinctive club-like spines of Tylocidaris54.

The cemented chalk of the Højerup Member, which constitutes the eroded crest of the bryozoan mound, flanking the Sigerslev basin to the south, yielded much more abundant macrofossils; masses of coarse bryozoan and echinoderm bioclasts were also observed on the rock surfaces. In addition to the calcitic forms, we noted well-preserved moulds after aragonitic shells, mostly fragments of baculitids (Baculites vertebralis and Baculites sp.) as well as bivalve and gastropod moulds. This is a fossil assemblage typical of the cemented parts of the Maastrichtian mounds at Stevns Klint23,43.

Lowermost Cerithium Limestone (Rødvig)

A single specimen of Baculites cf. vertebralis (OESM 13295) from the Rødvig basin belongs here (Fig. 4a–c). In comparison to Sigerslev, the infill of this basin is much thicker (c. 80 cm), with the hardground zone restricted to its upper part, which is clearly distinguishable by its hardness and yellowish colour from the soft, brittle pale grey limestone below.

Specimen OESM 13295 was recovered, in an oblique position, from the lowermost part of the Cerithium Limestone, at a level c. 10 cm above the base of the Fiskeler Member (Fig. 4a–d). The contact between the Fiskeler and Cerithium Limestone at this spot is of a transitional nature, similar to what is seen at many other sites along the cliff31,44,55. Above the black clay layer at the base of the Fiskeler, there is a series of thin clay laminae intercalated with carbonate layers (unit IV of Christensen et al.44). Clay laminae contain sparse fish skeletal detritus and become more widely spaced, and more marly upsection, passing eventually into a complex, compacted pseudo-conglomeratic microstylolitic fabric composed of carbonate burrow infills, lenses and clasts separated by horsetail clayey laminae (Fig. 4d). This is unit V of Christensen et al.44 (their Fig. 1) and the “microstylolitic net” of Ekdale and Bromley55 (their Fig. 8). Still higher, a limestone matrix prevails, crossed by stylolites and dissolution seams (unit VI of Christensen et al.44). At the level with the baculitid, we observed well-preserved macrofossils, including echinoid spines (up to 3 cm long) and bryozoan fragments (up to 2 cm long), plus abundant bioclasts of the same origin.

Microfacies analysis

We performed a microfacies analysis of samples from the Sigerslev basin and from the basin with the baculitid OESM 13295 at Rødvig, based on thin sections and SEM observations (see Figs. 4e, 5 and 6). Below, we focus on observations which are strictly relevant to the present topic. For a full description of microfacies and their photographic documentation, reference is made to Microfacies description in Supplementary Information (Figs. S1–S5).

Our observations in the Sigerslev basin reveal a variety of microfacies on both sides of the K–Pg boundary. These are chiefly wackestones and packstones with diverse grain components (Figs. 5 and 6; Figs. S1–S4). The most common grain types are calcareous dinoflagellate cysts (c-dinocysts, see Ciurej et al.56), tests of foraminifera (both planktic and benthic), siliceous sponge spicules (chiefly monaxons, preserved as empty or sparite-filled voids), as well as echinoderm (mostly echinoid), bryozoan and bivalve bioclasts. These components occur in variable frequencies, preservational states and size ranges, depending on stratigraphical unit (Fig. 5, Figs. S1–S4).

In the present context, the most important point is the identification of two key characters which allow us to differentiate the matrix of the main portion of the Cerithium Limestone Member (sensu Störling et al.31) from the Maastrichtian chalk of the Højerup Member. These characters concern abundance, mode of preservation and typical size of calcitic bryozoan and bivalve bioclasts and size of sponge spicules (Fig. 5, Figs. S1–S3). Bryozoan and bivalve-derived bioclasts in the main Cerithium Limestone are rare, poorly preserved (usually rounded) and typically of very small size (c. 100 μm). Such bioclasts are common, usually better preserved and larger in the Maastrichtian chalk, particularly at the top of the bryozoan mound (Fig. 5f). As far as the sponge spicules are concerned, these elements are more abundant and up to ten times larger (up to 1,000 μm in length) in the Cerithium Limestone than in the Maastrichtian chalk.

The infills of cephalopod moulds may provide clues as to their origin and the taphonomic processes involved57,58. Therefore, we performed a microfacies analysis of the infill of the best-preserved baculitid from the top of the Cerithium Limestone Member in the Sigerslev basin (Figs. 3h and i and 6, Fig. S3). The lower, presumably older infill consists of typical Cerithium Limestone matrix (microfacies 1 in Fig. 6a, c, d, Fig. S3b, f). In contrast, the higher, presumably younger portion of the infill (microfacies 2 in Fig. 6a, b, c, Fig. S3a–d, f) has no match in other microfacies identified across the K–Pg boundary interval at Sigerslev. Unlike the typical Cerithium Limestone microfacies, microfacies 2 contains only tiny sponge spicules (100 μm long and 20 μm thick), which is a typical feature of all Maastrichtian microfacies. However, these Maastrichtian microfacies contain abundant tests of foraminifera, particularly in the part below the Fiskeler (Fig. 5a, e) and large calcitic bioclasts, particularly at the bioherm top (Fig. 5a, f). These components are absent from microfacies 2. Moreover, the fine bioclast fraction immediately distinguishes microfacies 2 from the directly overlying bryozoan limestone of the Korsnæb Member (Fig. 5b).

Our microfacies analysis of the Rødvig section was restricted to the lowermost Cerithium Limestone Member, more specifically to the level which yielded the mould of Baculites cf. vertebralis, OESM 13295 (Fig. 4a–e). The microfacies identified is chiefly packstone and wackestone, dominated by well-preserved, coarse bryozoan bioclasts, abundant echinoderm bioclasts and benthic foraminifera (Fig. 4e, Fig. S6). The large size of the bryozoan and echinoderm bioclasts distinguishes it from all other Cerithium Limestone samples studied, but puts it very close to the microfacies present at the top of the Maastrichtian bioherm at Sigerslev (compare Fig. 4e with Fig. 5f). This type of microfacies, dominated by coarse bryozoan bioclasts, was identified at the base of Cerithium Limestone at Rødvig by Störling et al.31 and at Kulstirenden by Andersson30 and interpreted as reworked Maastrichtian material.

Fig. 4
Fig. 4
Full size image

Stratigraphical and sedimentological setting of Baculites cf. vertebralis from lowermost Cerithium Limestone Member (as defined by Störling et al.31) at Rødvig. (a) Field photograph within part of the section; hammer is 33 cm long. (b) Magnification of (a). (c) Specimen counterpart in slab preserved in museum collection (OESM 13295). (d) Cross section of slab containing counterpart illustrated in (c) to show soft sediment deformation. (e) Bryozoan pakstone with allochthonous bioclasts; note large bryozoan fragments (up to 2 mm in size). b – bryozoan bioclasts.

Fig. 5
Fig. 5
Full size image

Microfacies types across Cretaceous–Paleogene boundary in Sigerslev basin. (a) Idealised cross section through basin (not to scale). (b) Bryozoan packstone. (c) Wackestone with c-dinocysts, foraminifera and large sponge spicules. (d) Bindstone with presumed fenestral structures and large sponge spicule, composed of bioclastic packstone with very fine bioclasts (light-coloured layer at bottom) and wackestone rich in c-dinocysts (dark layer above). (e) Wackestone with c-dinocysts, foraminifera, fine sponge spicules and admixture of bryozoan bioclasts. (f) Bioclastic packstone (see text for further explanations). b – Bryozoan bioclasts, c – calcacerous dinocysts, e – echinoderms, f – foraminifera, fen? – presumable fenestral structures (empty), s – sponge spicules.

Fig. 6
Fig. 6
Full size image

Microfacies of infill of Baculites vertebralis (OESM 13290) from top of Cerithium Limestone Member, Sigerslev basin. (a) Scan of thin section prepared along transverse cross section of specimen; crossed nicols. (b) Microfacies 2 with well-preserved c-dinocysts. (c) Boundary between microfacies 2 and microfacies 1. (d) Microfacies 1 with large sponge spicules (monaxons). (e) SEM image of infill of specimen composed of micrite crystals with abundant calcareous dinocysts; note that relationship to microfacies distinguished above is not clear. b – Bryozoan bioclasts, c – calcareous dinocysts, Ce – calcareous dinoflagellate, Cervisiella operculata, cs – calcitic shell fragment, m – micrite crystals, s – sponge spicules.

Discussion

Machalski and Heinberg4 based their survival hypothesis on several lines of evidence, including consideration of the overall taphonomic patterns across the K–Pg boundary at Stevns Klint, taphonomic signatures of Cerithium Limestone Member ammonite specimens, as well as some biostratigraphical and stable isotopic data. In our reassessment of the survival hypothesis, new observations, chiefly of a taphonomic and sedimentological (microfacies) nature, are presented in preceding chapters. As will be seen below, our data provide arguments for both survival of some ammonites and redeposition of others, depending on the position within the Cerithium Limestone of the specimens studied.

Evidence for survival

Our study of the Sigerslev basin (where only the main portion of the Cerithium Limestone Member sensu Störling et al.31 has been identified) confirms a principal difference in composition of macrofossil assemblages between the Cerithium Limestone and the underlying/adjacent uppermost Maastrichtian chalk4,23,24,43. During our work we noted the absence of any low magnesium calcite-shelled macrofossils or coarse bioclasts in the Cerithium Limestone. In stark contrast, the top of the adjacent bryozoan mound (bioherm) is replete with calcitic remains. Calcite, particularly low magnesium calcite, is much more durable in the fossil record than aragonite59. Therefore, the difference between the basin and adjacent bioherm-crest must be regarded as a primary one, related to inhospitable conditions for low magnesium calcitic-shelled biota during deposition of the Cerithium Limestone23,24. If the aragonitic shells of ammonites preserved in the Cerithium Limestone had been eroded from the adjacent Maastrichtian bryozoan mound, they would have been accompanied by masses of more durable calcitic fossils, which is not the case.

All ammonite specimens from the main part of the Cerithium Limestone Member studied lack any signs of reworking and/or redeposition, like glauconitisation, phosphatisation or surface abrasion. Only the bipartite infill of baculitid OESM 13290 from the top of the Cerithium Limestone at the Sigerslev basin could pose some problems in this respect (Fig. 6). Potentially, such a two-generation infill might suggest a complex taphonomic history or even redeposition of this particular specimen. Even if only one of these infills would have had features of the Maastrichtian microfacies, it would be a strong argument for redeposition and thus against the survival hypothesis. However, this is not the case. The lower, presumably older microfacies 1 is typical of the Cerithium Limestone. The upper, presumably younger microfacies 2 is unique amongst all microfacies we distinguished on both sides of the boundary at Sigerslev, so is neutral in the redeposition versus survival debate. We speculate that microfacies 2 is a Danian sediment preserved in a shelter of the ammonite conch, which probably entered the conch from above, coming from sediment originally laid down above the present top of the Cerithium Limestone, but subsequently removed from the record by erosion. Based on an analysis of Maastrichtian bioherm geometry and topography, Heinberg23 (p. 89) estimated that “no less than 1 m of bioherm summits, and thus Cerithium Limestone, has been removed” by erosion which led to the truncation of the basins and intervening mounds. The Danian age of baculitid OESM 13290 is confirmed by SEM observations which reveal a typically Danian calcareous dinocyst Cervisiella operculata (Bramlette and Martini, 196460) in its infill (Fig. 6e). For discussions on stratigraphical range of this species the reader is referred to Machalski and Heinberg4 and Störling et al.31.

In summary, our observations from the main part of the Cerithium Limestone contradict redeposition of ammonites preserved in this unit, therefore pointing to the autochthonous (indigenous) nature of these fossils, and – by implication – confirming the survival hypothesis as presented by Machalski and Heinberg4. It should be emphasized on this occasion that the aragonite-shelled non-ammonite fauna known from the main portion of the Cerithium Limestone is so different from the Maastrichtian one in terms of taxonomic composition23,24,29 that this fact alone makes the redeposition of these remains unrealistic.

An additional, yet by no means decisive, argument in support of ammonite survival into the Danian may come from the presence of the genus Fresvillia (Fig. 3c, d) in the Cerithium Limestone at Sigerslev. So far, this genus was known only from the Maastrichtian outside Denmark (for details see Ammonite data in Supplementary Information). According to MacLeod61, the occurrence of Danian populations of “Cretaceous” species in areas in which they were not observed in upper Cretaceous strata may be regarded as a supportive argument for their survival across the K–Pg boundary. This criterion was used, for example, as one of the arguments in favour of the autochthonous nature of an ammonite faunule from the lowermost Danian (P0 Zone) unit IVf-7 of the Meerssen Member (Maastricht Formation) in the Maastricht area, the Netherlands. This unit contains baculitid species which do not occur in the underlying uppermost Maastrichtian deposits10,11.

The duration of ammonite survival in Denmark after the K–Pg boundary event is difficult to assess. For example, a potentially long residence of empty ammonite shells on the Danian sea-floor could lead to an incorrect estimate of the survival period. Literature review of radiocarbon datings of marine shells from nearshore and shelf environments reveal wide range of time-averaging in Recent and subfossil assemblages62. As far as cephalopods are concerned, radiocarbon dating of Nautilus macromphalus shells found exposed on a rocky floor in a cenote in the Loyality Islands (New Caledonia) yielded ages in the range of ca. 6000 to 7000 years BP63. Radiocarbon dating is obviously inapplicable for our material in view of its age and preservation, but a prolonged residence on the sea-floor may be documented, for example, by the presence of epizoans on shells of fossil and recent cephalopods64,65. The shells of epizoic biota, typically represented by calcitic oysters, serpulids, and bryozoans, are missing in the studied specimens. However, calcite-shelled biota are extremely rare, in general, in the Cerithium Limestone23,24, so the argument loses validity in this case. On the other hand, the size of our ammonites argues against post-mortem floating and long term drift. According to experiments on ammonite conchs66, small specimens such those in the present study would not have experienced much post-mortem transport.

Another difficulty in dating the Danian survivors is imposed by the diachroneity of Cerithium Limestone deposition along Stevns Klint outcrops24,25. Based on present data and those gathered by Machalski and Heinberg4, the Danian ammonites are now known from a series of Cerithium Limestone exposures along the cliff, ranging from Rødvig to Holtug outcrops (Fig. 1b). According to Rasmussen et al.25, this unit becomes progressively younger northwards, from the Parvularugoglobigerina eugubina Zone (Pα) in the south to the Parasubbotina pseudobulloides Subzone (P1a) of the Parasubbotina pseudobulloides–Globoconusa daubjergensis Zone (P1) in the north.

The Sigerslev quarry, from which our most important data originate, is located closer to the northern end of the range of ammonite-bearing Danian sites. Based on extrapolation of data supplied by Rasmussen et al.25 (their Fig. 7), this allows us to place the Cerithium Limestone ammonites from Sigerslev in the P1a Subzone. Following Rasmussen et al.25, the base of this zone is defined by the first appearance of the planktic foraminiferal species, P. pseudobulloides (Plummer, 192767), which is dated at 68 kyr after the K–Pg boundary event68. Thus, this may be taken as the minimum duration of ammonite survival, at least in the Sigerslev area. As far as the maximum duration is concerned, Machalski and Heinberg4 tentatively estimated the hardground at the top of the Cerithium Limestone to have formed 200 kyr after the K–Pg boundary, following Smit69. In any case, available data allow us to conclude that the Danish Danian ammonites survived much longer than the ammonites known from the the post-impact deposits of the USA7 and the Netherlands11. Statistical analyses of ammonite abundance and stratigraphic distribution at the Bay of Biscay (France) and Antarctica suggest that, in theory, two additional genera survived the K-Pg boundary at the Bay of Biscay70, and one survived in Antarctica71. However, no actual records of ammonite survivors have been documented at either site10. Outside of Denmark, the ammonite survivors seem to be restricted to the lowermost Danian P0 zone in terms of planktic foraminifera biozonation. The P1a Subzone ammonites from the Cerithium Limestone are therefore the last documented representatives of this group in the entire world.

Evidence for redeposition

The depositional setting of baculitid OESM 13295 from the lowermost Cerithium Limestone Member of Rødvig (Fig. 4) reveals a quite different story regarding its origin. The specimen floats in a mass of Maastrichtian bryozoan-dominated bioclasts (Fig. 4e) and some clearly Maastrichtian macrofossils have been found at the same level. Our observations therefore fully match those of Störling et al.31 and Andersson30, who drew attention to the presence of dense accumulations of calcitic macrofossils and coarse, mostly bryozoan, bioclasts, at the very base of the Cerithium Limestone at Rødvig. These authors regarded this deposit as an essentially allochthonous (extraneous) material, redeposited into the Cerithium Limestone basins from the adjacent uppermost Maastrichtian mounds, and we concur with this conclusion. In addition to calcitic fossils and bioclasts, the cemented bioherm crests at Stevns Klint contain numerous moulds of aragonite-shelled organisms, including abundant ammonites36. We therefore argue that specimen OESM 13295 was redeposited (probably with its original shell still preserved) into the Cerithium Limestone basin along with a mass of calcitic remains. Interestingly, a baculitid mould in a similar location, i.e. at the base of Cerithium Limestone, just above the Fiskeler, was recorded by Rosenkrantz33 (p. 29, footnote 1). The complex fabrics of the interval exposed at Rødvig, which yielded OESM 13295, utterly differs from that of the main part of the Cerithium Limestone, as observed at Sigerslev. Marly seams with fish debris intercalate with limestone layers just below the level of the specimen discussed (Fig. 4d). This suggests intermittent synsedimentary submarine dissolution of carbonate sediment55. This pulsatory dissolution was probably related to sea water acidification in the direct aftermath of the end-Cretaceous impact30. According to the latter author, the acid waters could have destabilised the unconsolidated sediment on top of the bryozoan mounds, which allowed its transport into inter-mound troughs during the earliest phases of deposition of the Cerithium Limestone. The soft-sediment deformation of the interval just below the baculitid specimen (Fig. 4d) suggests synsedimentary submarine slumping. Deformations of this kind have been documented at various levels along the Stevns Klint succession55.

In summary, OESM 13295 most probably is an allochthonous fossil, transported into the Cerithium Limestone Member basin at Rødvig from the crest of an adjacent Maastrichtian mound, probably through submarine slumping. Interestingly, the presence of baculitid shafts (“Stumpen”) was recorded from the upper part of the lowermost Danian Fiskeler Member by Rosenkrantz72, who interpreted these finds as reworked Maastrichtian fossils in view of their chalk infills. Heinberg23 (p. 89) reported on the presence of reworked Maastrichtian chalk pebbles and chalk-infilled fossils in the Fiskeler, which suggested, in his view, some degree of lithification of the mound crests prior to Fiskeler deposition. However, this issue is beyond the scope of the present paper.

Conclusions with implications

Our results for the main part of the Cerithium Limestone Member (sensu Störling et al.31) point to the autochthonous nature of the enclosed ammonites, which means that they are early Danian survivors, as suggested by Surlyk and Nielsen2 and firmly proposed by Machalski and Heinberg4. In contrast, a single baculitid from the lowermost part of the Cerithium Limestone is interpreted as a redeposited Maastrichtian fossil, suggesting that also redeposition of ammonite specimens from adjacent Maastrichtian mounds was taking place during the initial stages of infill of intervening basins.

Ammonites did not make it higher up into the Danian, because there are no remains known from above the lower Danian Cerithium Limestone Member, except for the clearly allochthonous specimen from the base of the Korsnæb Member mentioned above. Therefore, early Danian ammonite survivors provide a particularly short-term case of the “Survival without recovery” or “Dead Clade Walking” phenomenon73,74,75.

The Danish Danian ammonites clearly survived the end-Cretaceous crisis triggered by the Chicxulub impact10,76. The question remains as to what, if not the commonly invoked impact-related acidification of surface waters10, caused their final eradication from the early Danian shallow-marine habitats in Denmark and elsewhere. Locally, it is possible that the last ammonite populations fell victim to environmental changes related somehow to the sea level drop manifested by sequence boundary 2 at the top of the Cerithium Limestone (Fig. 1c). Does this discontinuity surface mark a complete withdrawal of the sea from the study area? If yes, this would suggest habitat loss as a driver of the extinction of the last Danish ammonites, in concert with traditional linking of mass extinctions to sea level oscillations77. Although no traces of emersion have ever been recorded from the Stevns Klint succession, palaeogeographical maps show a very limited range of marine sedimentation in the Danish basin during deposition of the Cerithium Limestone Member40. Given the current state of knowledge, we are unable to assess whether non-environmental biological factors, such as competition with other members of the nectobenthos, may have contributed to the ultimate extinction of the Cerithium Limestone Member survivors. While the global disappearance of the ammonites remains an open question, their populations were geographically restricted after the catastrophic events at the K-Pg boundary9, rendering them particularly vulnerable to extinction.