Danian dinoflagellate zonation, the C-T boundary and the stratigraphical position of the fish clay in southern Scandinavia

Two Swedish borehole cores covering the Danian, and three Danish outcrop sections covering the Lower Danian, were analysed for dinoflagellates. This study suggests that the Swedish sections, in contrast to previous investigations, are the most complete sections across the Cretaceous-Tertiary boundary in the Danish Embayment. As a consequence, the previously established biozonation can be modified to include one more basal Danian zonule. A detailed study of the dinoflagellates of the Fish Clay in Denmark suggests that this layer was formed in a low salinity environment in eastern Denmark and in a stable marine environment in western Denmark. In addition, the Fish Clay is shown to be biostratigraphically older in eastern Denmark than in western Denmark.


INTRODUCTION
The Danish Embayment is a sedimentary basin containing Mesozoic and Cenozoic deposits. It is a part of the Danish-Polish Trough, which extends as a belt in a north-westerly to south-easterly direction from Denmark across Poland. The Danish Embayment was connected to the North Atlantic and Tethys through narrow sea passages (Posaryska, 1965). As a consequence of the regression that took place in the Late Maastrichtian (Vail et al., 1977;Hancock & Kauffman, 1979: Morner, 1980, the environment in the Lower Danian of Scandinavia was a shallow shelf environment (Hultberg & Malmgren, In press (a)).
Danian deposits are characterised by strongly undulating structures which have been interpreted as bioherms (Cheetham, 1971) or megaripples (N.A. Morner, University of Stockholm, personal communication, 1984), formed under the influence of strong, unidirectional currents. These structures contrast strongly with the laminated structure of the Late Maastrichtian deposits. Danian deposits in Scandinavia consist of biogenic limestones, containing an average of about 75% CaCO, (Jargensen, 1979). The most common constituent of the Danian limestones is bryozoan skeletons, but skeletal fragments of echinoderms, molluscs, nannoplankton, and foraminifera are also abundant.
The Cretaceous-Tertiary boundary is very distinct in Denmark, where it is marked by a change from Maastrichtian chalk facies to a thin clay layer. This clay, which can be seen at all known Danish C-T boundary sections, has been called "Fish Clay", because of its content of fish remains. This clay has been considered to be of lower-most Danian age (Rosenkrantz, 1966). It is a formation of four thin clay layers, immediately overlying the Maastrichtian chalk. The lowermost layer is a 2-4cm thick marl bed of a whitish grey colour. Overlying it is a 2-5cm thick dark brown to black bed with abundant pyrite concretions of varying size, which is in turn overlain by a 3-5cm thick dark grey marl bed, which gradually changes into a light grey marl horizon, about 7 cm thick.
Many hypotheses have been proposed about the biotic changes across the C-T transition. The rapid extinction of many plant and animal groups (for example the dinosaurs) have been explained by an instantaneous, global catastrophe. The most widespread theory is that of an extraterrestrial body impact, which drastically affected the earth's biosphere (Alvarez et al., 1980;Alvarezet al., 1983). The argument for this theory is the conspicuous peak of iridium found at various C-T boundary sections around the world (Alvarez, 1983), one of which is the Stevns Klint section in eastern Denmark.
The purpose of this study was to expand the existing Danian dinoflagellate stratigraphy of southern Scandinavia to include new zones at the bottom and the top of the Danian, to study the completeness of the lowermost Danian at various localities, and to investigate the timerelationships and depositional environment of the Fish Clay at various localities.

MATERIAL AND METHODS
Two borehole cores, D105 and D106, from the Limhamn area, southern Sweden, and three Danish outcrop sections, Stevns Klint, Dania, and Kjalby Gaard Relative abundances of dinoflagellates were generated by counts of at least 300 specimens per sample. The census counts included all species found in the slide.  (Christensen et al., 1973). Each of these was sampled separately for dinoflagellates; three different sections, about 150m apart, were sampled from the cliff at Stevns Klint. Dinoflagellate preparations were initiated with the removal of carbonates by treating the samples with HC1 (Wilson, 1974b). The residues were centrifuged and submitted to a heavy-liquid separation with ZnC1,. The float fraction was decanted, washed through a sieve with a mesh size of 10,um, and mounted on slides, using glycerol jelly as a mounting medium. The bottom fraction in the heavy-liquid separation was routinely scanned for dinoflagellates. Two slides were made from each sample, one of which was stained with fuchsin.

SYSTEMATIC PA LA EONTOL OGY
This section includes all the species that have been used in the establishment of a modified dinoflagellate biostratigraphy for the Danian of Scandinavia and the species restricted to the Fish Clay in this area. In total, approximately 50 species were recognised.
(Plate 2, figs. 3, 4) Holotype. Slide Sk 16-1, from Stevns Klint, Denmark. Coordinates x = 103.6, y = 23.8. Diagnosis. Proximate cyst composed of one wall layer or two wall layers, closely appressed. Polyhedral in shape with prominent apical horn. Cyst wall is marked by vacuoles of different size, giving wall a reticulate appearance. These bacuoles are evenly distributed on cyst wall. No trace of paratabulation is present except the archeopyle. Paracingulum is marked by a slight depression in cyst wall. Parasulcus is marked by a conspicuous area of vacuoles, much smaller than those at other locations. Archeopyle is precingular, type P, formed by the detachment of paraplate 3". Operculum is free. Archeopyle is very large and may be enlarged. Occurrence. S. reticulaturn is found at Stevns Klint in two samples, 10 and l l m above the Fish Clay. This species has not been encountered in any of the other sections.
Remarks. This species can easily be distinguished from all other species of Spongodinium by its extremely vacuolate periphragm. The only other species with similar vacuoles is S . delitiense. However, S. delitiense has unevenly distributed vocuoles on a smooth periphragm.
Diagnosis. Shape of endophragm is subspherical to slightly lenticular. A slight antapical indentation is present in most specimens. Endophragm surface finely granulate on dorsal surface with coarser granulation at paraplate boundaries, thus faintly indicating para-H ultbe rg cingulum and paraplates 3", 4", and 5". Ventral surface smooth. Margin of endophragm bears a few solid, unevenly distributed processes of varying thickness, bases of which are isolated, but the processes gradually become confluent with the periphragm. Proximally, the processes fuse with a perforate periphragm, which distally changes into a homogeneous, almost nonperforate periphragm with no indication of the processes. Periphragm terminates with an irregular margin. Proximal part of processes pointed out laterally from endophragm, but are together with the periphragm turned in a ventral direction, terminating in front of ventral part of endophragm, and pointing towards center of endophragm. In apical view, processes and periphragm form a "wing-shaped'' structure. Antapically, periphragm (in very well preserved specimens), shows a slight indentation. Archeopyle is apical, type tA, with a free operculum. Primary archeopyle suture zig-zag, indicating six precingular paraplates.

DINOFLAGELLATE ZONATION
The Danian dinoflagellate flora in southern Scandinavia is characterised by well preserved dinoflagellate assemblages, marked by high abundances of Spiniferites ramosus, Senoniasphaera inornata , Fibrocysta axialis and Areoligera coronata . The species composition of the Danian assemblages is similar to that of the Maastrichtian floras. The C-T transition is not marked by pronounced extinctions and appearances of dinoflagellate species, but mostly by changes in the relative abundance of dinoflagellate taxa. Hansen (1977) studied the dinoflagellate biostratigraphy of the uppermost Maastrichtian and the Lower Danian localities. He divided the DanianDanea mutabilis Zone (Hansen, 1977) into the Senoniasphaera inornata and the Hafniasphaera cryptovesiculata Subaones. The S. inornata Subzone was further subdivided into the Carpatella cornuta , Xenicodinium rugulatum , and X . lubricum Zonules. Kjellstrom & Hansen (1981) applied thiszonation to the borehole cores D104, D105, and D106, from Limhamn, southern Sweden. They suggested a stratigraphical range of the Limhamn cores from slightly above the lower boundary of the X . rugulatum Zonule to slightly below the upper boundary ot the X . lubricum Zonule.
Hansen (1 977) encountered the lowermost Danian C . cornuta Zonule at Kjolby Gaard and Dania, but not at Stevns Klint and Limhamn. Likewise, the uppermost Maastrichtian dinoflagellate zone, the S. inornatal Palynodinium grallator. Concurrent-Range-Zone, established by Hansen (1 979), was encountered by him at Kjolby Gaard and Dania, but not at Stevns Klint and Limhamn. Therefore, he concluded that there had been a break in sedimentation, resulting in a hiatus, covering the uppermost Maastrichtian and lowermost Danian. This hiatus was considered to have an increasing stratigraphic extent eastwards (Kjellstrom & Hansen, 1981). Thus, among the localities studied by Kjellstrom and Hansen (198 l ) , Limhamn was considered to be the most incomplete section and Kjdby Gaard the most complete across the C -T boundary.
The S. inornataiP. grallator Concurrent-Range-Zone (Hansen, 1979) was defined on the basis of the joint occurrence of these two species. Hultberg & Malmgren (198%) noted that some of the specimens referred to S. inornatu by Hansen (1 979) were referable to Glaphyrocysta perforata. In the present study, it was found that S. inornata, in fact, first appears in the Danian.Glaphyrocysta perforata occurs throughout the Upper Maastrichtian and the lowermost Danian.
At Limhamn, the interval overlying the barren interval, contains high abundances of F . axiale (Fig. 3). This interval can be used to establish a basal Danian Acme-Zonule within the S . inornata Subzone. This zonule is defined by a series of samples exhibiting relative abundance of F . axiale greater than 10%. The average relative abundance of this species in this zone is approximately 19% and the average value in the Danian is approximately 6%).
The C. cornuta Zonule, suggested by Hansen (1977) to be basal Danian, was not found at Limhamn by Kjellstrom and Hansen (1981). In the present study, the F. axiale Acme-Zonule is followed by Limhamn by an interval containing abundant specimens of C. cornuta (40%) of the assemblage). This interval is 6 m thick in core D105 and6.Smthickincore D106.TheC.cornuta Zonule was found by Hansen (1 977) to be approximately 0.2m thick at KjQlby Gaard and 0.3 m thick at Dania. At Stevns Klint, the Fish Clay was considered to belong to this zonule. In the present study, no specimens of C. cornuta were found in any of the Fish Clay horizons.
The interval of low dinoflagellate content, the F. axiale Acme-Zonule, and the C. cornuta Zonule together comprise approximately 26m of Danian sediments at Limhamn (Fig. 4). At Kjeolby Gaard and Dania, the total thickness of these intervals is 0.1 m and 0.2m respectively. In contrast to previous suggestions, this indicates that Limhamn is the most complete sequence in the Lower Danian and that the stratigraphical extension of the hiatus increases westwards instead of eastwards (Fig. 5). The fact that no specimens of C. cornuta were found at Stevns Klint suggests that the interval of low dinoflagellate content here is correlatable with the similar interval at Limhamn, and that sediments belonging to the F. axiale Acme-Zonule and the C. cornuta Zonule have been eroded at Stevns Klint. The C. cornuta Zonule is overlain by the X . rugulatum Zonule, characterised by the presence of X . rugulatum and X . reticulatum and the absence ofX. lubricum. This zone comprises 12 m in core D105 and 11 m in core D106. The uppermost part of this zone is missing at KjQlby Gaard.
The H . cryptovesiculata Subzone is here found for the first time in Swedish deposits. It is defined by the first occurrence of H . cryptovesiculata. This zone was described by Hansen (1977) as uppermost Danian. However, the uppermost part of the Limhamn cores is marked by the first appearance of Pithonella organica, which can be used to establish a new, uppermost Danian subzone, defined by the first appearance of this species. The H . cryptovesiculata Subzone is 14m thick in both core D105 and core D106. The P. organica Subzone is 14m and 13m thick, respectively, it. cores D105 and D106. Fig. 5 shows a range chart for all stratigraphical index species. Fig. 6 shows a comparison between the dinoflagellate zonation established by Hansen (1 977, 1979), and the modified dinoflagellate zonation established in this paper.

PALAEOECOLOGY A N D RIOSTRATIGRAPHY OF T H E FISH CLAY
In this study, the four different beds o f the Fish Clay (Christensen et a / . , 1973) were analyscd separately lor dinoflagellates. The uppermost metre of Maastrichtian chalk, immediately below the Fish Clay, displays low abundances of dinoflagellates, which indicates that thc assemblages in the Fish Clay are not reworked, but primary. The lowermost bed contains a well-developed Maastrichtian dinoflagellate assemblage, containing abundant specimens oft'. grullator. In this assemblage, no dinoflagellates indicative of the Danian were t'ound.
The second bed is marked by a monospecific occurrence o f M . druggii.
The third bed is marked by an assemblage consisting ot 70'%) M . druggii.
The uppermost bed of the F i s h Clay does not contain a n y d i no fl age1 I ate 3 .
The gonyaulacacean ratio was established by tiarland ( 1 973) as a measure of relative salinity; this is the ratio between gonyaulacacean and peridiniacean dinoflagellate cysts. Harland (1973) compared this ratio between core top samples in the Caribbean (high salinity) and the nearshore waters of Woods Hole, Massachussetts (lower salinity). In the samples from Woods Hole, the abundance of peridiniacean cysts was higher. Thus, the value of the gonuaulacacean ratio was lower in the Dinof lagellate Zonation Hansen (1977Hansen ( ,1979 Kjellstrorn E  Hansen (1977Hansen ( , 1979 and the zonation established by the present study. samples with lower salinity. This ratio is usually high in the Maastrichtian and Danian (Fig. 7), which indicates open rnarinle conditions. In the second bed of the Fish Clay, the gonyaulacacean ratio has a value close t o zero (Fig. 7). It is difficult t o draw statistical conclusions from a monospecific assemblage, but this may indicate low salinity o r even brackish conditions during the deposition of this bed (average value in the Danian is 150).
T h e gonyaulacacean ratio in this bed has a value of 6.4 which. again, indicates a low salinity environment (Fig. 7).
More than 100,000 dinoflagellate specimens have been recorded in the Fish Clay from Stevns Klint. P . grallator was found in beds one and three, but no dinoflagellates indicative of the Danian occur in any of the beds. The specimens ofP. grallator in the Fish Clay are considered t o be in situ. If they were reworked, P. grallator would also be present in bed two. Thus, according t o dinoflagellate stratigraphy, the Fish Clay at Stevns Klint should be placed in the uppermost Maastrichtian, not in the lowermost Danian.
Unfortunately, the Fish Clay contains very little information on other microfossil groups, such as planktonic foraminifera and nannofossils. N o stratigraphically useful planktonic foraminifera are encountered in any of the Fish Clay samples. Nannofossils are present in low numbers in the Fish Clay, however, they are severely dissolved and appear t o represent a reworked Maastrichtian flora. All Fish Clay samples, except those from bed two, contain a moderately well preserved fauna of benthic foraminifera. The samples contain abundant Maastrichtian representatives, such as Arenobulimina obesa, Gavelinella pertusa , and Stensioina yomeruna. N o representatives of the Danian Eoglobigerinu dunica assemblage were found. This means that there is no evidence from other microfossil groups contradicting the Maastrichtian age of the Fish Clay at Stevns Klint. i ACKNOWLEDGEMENTS 1 am grateful to Dr. Bjorn A. Malmgren, Department of Paleontology, University of Uppsala, for his supervision of this project and his help during the preparation of the manuscript. He also examined the samples for planktonic foraminifera. Jan Gabrielson, Department uf Palaeontology, University of Lund provided the data on benthic foraminifera. I thank Dr. Maurits Lindstrom, University of Stockholm, for valuable comments on the manuscript. I am also grateful to Birger Schmitz, University of Stockholm for his valuable help during the field work and stimulating discussions. I thank Borje Sawensten, University of Stockholm, for technical assistance, and Inger Arnstrom and Solveig Jevall, bniversity of Stockholm for drafting assistance.