Typology and fossil assemblage of Sandbian ( Ordovician ) ‘ baksteenkalk ’ : an erratic silicified limestone of Baltic origin from the northeastern Netherlands and adjacent areas of Germany

This study seeks to describe ‘baksteenkalk’, an erratic silicified bioclastic carbonate of the Upper Sandbian from the eastern part of the Netherlands. To date, baksteenkalk has received little attention among palaeontologists. This is to be regretted on two grounds. First, baksteenkalk contains a varied fossil flora and fauna comprising many species, several of which are not or only rarely found in coeval rocks. Second, owing to a complicated silicification process, fossils, in particular algae, have preserved exceptional anatomical details. The primary aim of this study is to arouse the interest of professional, in particular, Estonian, palaeontologists in baksteenkalk. Based on lithology and fossil assemblage (most conspicuously, with regard to the algal flora), two basic types of baksteenkalk are distinguished. A list of species, differentiated for both types, is provided. It is argued that baksteenkalk reflects the ecology of a shallow, subtropical, epicontinental sea. The distribution of erratics, facies and fossil content point to an origin within the North Estonian Confacies Belt, probably west of Estonia. Baksteenkalk survived as an erratic because it was already silicified at its place of origin. A potential source of silica may have been Upper Ordovician bentonite layers. The causes and mechanisms of the silicification process which gave shape to baksteenkalk are not yet understood, however. The palaeontology of baksteenkalk is compared with that of two other erratic Sandbian silicified carbonates of Baltic origin: German Backsteinkalk and ‘Lavender-blue Hornstein’.


INTRODUCTION
'Baksteenkalk', literally 'brick-limestone', is the Dutch name for an erratic, entirely or partially leached, silicified carbonate, containing a Late Ordovician fossil assemblage, predominantly reflecting the Upper Sandbian Stage (Haljala Regional Stage C III -D I ).Baksteenkalk occurs in fluvio-glacial deposits of the Appelscha Formation (early Pleistocene) in the northeast of the Netherlands.It is in particular common in gravel pits of the Twente district (Overijssel province) and adjacent German territory, i.e., the Niedergrafschaft Bentheim.For convenience, this area is referred to as the 'WWW-area', after the villages Wilsum, Wielen and Westerhaar (Fig. 1).
Baksteenkalk is one of various types of erratic Ordovician silificied carbonates occurring in the gravel assemblages of the Appelscha Formation.The fossil assemblages in these carbonates represent mainly two periods of the Ordovician: the Haljala/Keila stages (C III -D II ) and the Pirgu/Porkuni stages (F I c-F II ).On the basis of lithological characteristics and fossil content, at least six different groups of silicified carbonates can be distinguished.
1. Baksteenkalk, the subject of the present study, will be treated extensively below.2. Lavender-blue Haljala-and Keila-hornstein (D I -D II ).
In unweathered condition bluish-black or brownishyellow, hard cherts; leached, porous parts with lighter colours, ranging from lavender-blue to bluish-white.Two types can be distinguished among these hornsteins: boulders with a knotty, irregular surface, containing crinoid segments and brachiopods, and massive blocks and slabs, often showing infiltration bands, in which sponges and cyclocrinitid algae are the dominant fossils (Van Keulen et al. 2012).3. Lavender-blue Pirgu-hornstein (F I c-F II ).Cherts consisting of irregular, undulating lumps or layers marked by distinct colours.Holes and burrows are often filled with chalcedony (achates) and quartz crystals.Several types can be distinguished among these cherts.One type contains a fossil fauna of stromatopores, tabulates and rugosa; another type is characterized by the presence of Plectatrypa; a third type with slag-like crusts is dominated by fragments of Incuhinzia syltensis (Schallreuter 1990;Van Keulen et al. 2012).4. Brown Pirgu-hornstein (F I c-F II ).In unweathered condition massive, hard, brownish-grey chert; boulders either with a light grey leached crust or entirely leached.Leached parts coarsely grained, crumbly, with uneven surface.Fossils of algae (Vermiporella fragilis, Palaeoporella variabilis) and bryozoans are common.Other fossils comprise sponges, gastropods, brachiopods, tabulates and crinoids.Trilobites are few.Algae and brachiopods are often preserved as massive, grey or white china-like chalcedony. 5. Öjlemyrflint of the 'Wielener type' (F I c-F II ).In unweathered condition splintery, brown-greyish chert; many blocks with one curved, whitish, leached side, possibly fragments of concretions.The fine-grained silicified limestone in the leached crusts of the Öjlemyrflints also occurs as homogeneous slabs and oblong blocks.The putative concretions may originally have been embedded in these blocks of silicified limestone (which is often referred to by the unfelicitous name 'Öjlemyrkalk' [literally 'Öjlemyr limestone']).
The fossil fauna is rich in ostracods, bryozoans, brachiopods and trilobites (Erratencrinurus kiaeri, Harpidella sp., Ascetopeltis bockeli, Pharostoma sp., Trochurus sp.) but poor in gastropods and sponges.Fossils tend to be concentrated in layers and are more or less deformed by compression.6.Small pebbles containing (casts of) goethitic ooids.
These probably originate from the Estonian 'Linsenschicht', dating either from the Lower Kunda Stage (B III ) or the Lower Aseri Stage (C I a).They are rare in the gravel of the Appelscha Formation, but less so on Sylt (Hinz- Schallreuter & Schallreuter 2005;Van Keulen et al. 2012: 171).
Efforts to differentiate the silificied carbonates were only made in 1993 (Rhebergen 1993).Previously, Dutch palaeontologists distinguished between 'baksteenkalk' or 'rolsteenkalk' on the one side and lavender-blue silicifications on the other (e.g.Krul 1963: 148).The scarcity of erratic silicified limestones which they had at their disposal may have prevented them from making a more refined division.Today, silicified limestones are extant in substantial numbers due to the extraction of sands from various sandpits, but before the second half of the 20th century sandpits were not exploited on an industrial scale.If collectors were aware of the occurrence of various groups among the erratic silicified limestones, they were not able to distinguish between them because of lack of appropriate names, so they came to use the term 'baksteenkalk' as a generic term for all (greyish) Ordovician silicified limestones.In historical perspective this indiscriminate use may be pardonable.However, after Rhebergen's attempt at classification, the use of the term 'baksteenkalk' in an unspecific way should be avoided.
As unfelicitous as the way in which the term 'baksteenkalk' has been used, is the name as such.It is derived from German 'Backsteinkalk', which, like baksteenkalk, denotes a type of erratic silicified carbonate rock of predominantly Upper Sandbian age.The name 'baksteenkalk' implies its being identical to Backsteinkalk, whereas in fact it differs from the latter in several respects.1. Boulders of Backsteinkalk are Saalian and Weichselian glacial erratics occurring over wide areas in the northern and in particular the northeastern parts of Germany.Boulders of baksteenkalk, on the other hand, are dropstones which fell from melting ice floes associated with the river system known as the 'Baltischer Urstrom' or the Eridanos (Rhebergen et al. 2001;Van Keulen et al. 2012).Their occurrence is more or less restricted to fluvio-glacial quartz sands of the Appelscha Formation which in the eastern parts of the Netherlands are exposed in ice-pushed ridges of the Saale glacial.
2. The rocks are petrologically different.Most boulders of Backsteinkalk are hard and solid and fossils cannot easily be extracted from them.Baksteenkalk, on the other hand, is largely, though to various extents, porous.Rocks tend to break along the surface of fossils, since these are often separated from the matrix by a thin layer of empty space caused by the dissolution of shells.3. Backsteinkalk and baksteenkalk show marked differences in the composition of their fossil assemblages.In part this is due to the fact that their stratigraphical ranges as well as their geographical provenances are different.Krueger (1995: 670) specifies the difference between baksteenkalk and Backsteinkalk by referring to the former as 'backsteinkalkartige Gerölle'.

MATERIAL AND METHODS
The material discussed here has been collected by the authors in the WWW-area and derives from estimatedly 10 000 blocks.The collections comprise thousands of fossils set in fragments of boulders.Most boulders have been broken to expose fossils, but several boulders have been left intact.All fossils from the same boulder can be tracked down by means of electronic databases.
The specimens of baksteenkalk depicted in this paper are provided either with an RGM-or a PMU-number.The former will be stored in the collection of Naturalis Biodiversity Center, Leiden (Netherlands), the latter in the palaeontological collection of the Museum of Evolution, Uppsala University (Sweden).

Typology of baksteenkalk
Entirely or partially leached, silicified carbonate containing marine fossils of Upper Sandbian age.Boulders with sub-perpendicular, straight sides and/or knobby surface.Diameter varies but never exceeds 300 mm.Colours vary from grey in barely leached parts to white in strongly leached parts.

Type 2
More or less rectangular slabs or blocks with abundant fossils appearing concentrated in levels.In comparison with type 1, the fossil assemblage is characterized by different species.
Type 2A (Fig. 3).Rectangular slabs or blocks with a flat bottom and a top often marked by depressions and cavities.Blocks may comprise coarse-grained portions alongside fine-grained ones.Micro-and macrofossils are abundant and often appear concentrated in one or two level(s).As in type 1, fossils occur as moulds, which are sometimes filled in with chalcedony, predominantly in brachiopods, rugosa and trepostomate bryozoans.Fossils are not, or only slightly, compressed (e.g.Mastopora concava).Individual blocks always contain a variety of species (polyspecific).Occasionally, representatives of one taxon dominate, e.g.gastropods (Brachytomaria baltica), cyclocrinitids (Coelosphaeridium sphaericum, Mastopora concava), trepostomate and cryptostomate bryozoans, brachiopods (Sowerbyella plana).Blocks frequently contain tubes, pipes or pockets filled with debris of crinoid cirrae, bryozoans, small brachiopods and monaxons.Also frequent are filled-in burrows, which can be recognized as shades of grey different from the surrounding rock.The algal flora has a composition different from that of type 1.The fauna is more varied than that of type 1. Type 2A makes up about 15% of all baksteenkalk.

Distinctive features of the main types
The distinction between the two main types of baksteenkalk is based on the following aspects.

Matrix
Both types consist of silicified bioclasts varying in grain size.In type 1, fine-grained bioclasts prevail.In type 2, bioclasts are more varied in grain size; coarse-grained portions are frequent, but occur alongside fine-grained portions.

Shape of boulders
Type 1 occurs as blocks with rounded knobs; type 2 is commonly found as slabs with right angles.

Size and distribution of fossils
In type 1 large fossils, exceeding 20 mm in size, are uncommon.Fossils larger than 2 mm in size tend to be concentrated in nests.In type 2 large fossils of gastropods, brachiopods, trilobites, cephalopods and bryozoans are more numerous.They occur either densely packed in one or more layers or hitched together in clusters.Tubes filled with debris of crinoids, monaxons and bryozoans are frequent in type 2 but lacking in type 1.

Epifaunal overgrowth and fecal pellets
Unlike type 1, type 2 is characterized by fossil remains showing a variety of traces and marks left by the activity of benthic organisms.Brachiopod shells, calcareous algae and hard parts of trilobites are overgrown by encrusting bryozoans or damaged by etching bryozoans, such as Corynotrypa.Thread-like structures in the interior of the cyclocrinitid alga Coelosphaeridium sphaericum were interpreted by N. Spjeldnaes as fungi (pers. comm. 1986).Shells of gastropods, cephalopods and brachiopods are sometimes filled with Tomaculum problematicum, little pellets which are currently understood as faeces deposited by an unknown organism (Eiserhardt et al. 2001;Bruthansová & Kraft 2003).Groups of polygonal dots on the inside of gastropod shells are possibly connected with Tomaculum.Rather common is Arachnostega gastrochaenae, burrows in shells of gastropods and brachiopods made by an unknown organism (Vinn et al. 2014).

Trace fossils
Trace fossils are abundant in baksteenkalk but each type has its own set.Long tubes with a core of chalcedony  surrounded by porous debris are frequent in type 1 but less so in type 2 (Fig. 7).On the other hand, the tubes filled with coarse fossil debris of crinoids and bryozoans which were mentioned above, are restricted to type 2 (Fig. 8).Concentrically structured spheres of porous   debris, possibly transverse sections of burrows, are more frequent in type 2 than in type 1 (Fig. 9).Chondrites is common to both types, but differs in appearance.

Composition of the fossil assemblages
Central to the distinction of the two main types is the composition of their fossil assemblages.In particular algae, which are ubiquitous in baksteenkalk, provide a sound basis for the distinction.Species of Apidium (A. pygmaeum, A. sp.['claviformis']) and Cyclocrinites (C.porosus, C. cf.schmidti) are almost never found together in one block.Hoeegonites, being a frequent companion of Apidium, has never been observed together with Cyclocrinites.Other species of algae known from baksteenkalk, i.e., Coelosphaeridium sphaericum, Mastopora concava, Vermiporella fragilis and Solenopora spongioides (the algal nature of which is contested, see Riding 2004) occur together with Apidium as well as with Cyclocrinites and as a consequence are not exclusive to a particular assemblage.Thus, the algal flora allows us to distinguish two distinct assemblages, corresponding to the two types of baksteenkalk (Table 1).
This basic distinction can be extended to elements of the fauna.The brachiopod Bilobia aff.musca frequently occurs together with algae of assemblage 1, but is not found together with Cyclocrinites.The same applies to the trilobite Nieszkowskia inermis.Conversely, the trilobites Hemisphaerocoryphe pseudohemicranium and Paraceraurus elatifrons occur exclusively with algae of assemblage 2. Other species which are probably restricted to assemblage 2 are Astamena cf.inaequalis, Kurnamena taxilla, Sowerbyella plana and Cymbularia roemeri.
The two assemblages share a high number of species.Some of the species which are common to both assemblages hold a prominent place as companions of species exclusive to assemblage 1.These are Otarozoum peri, Harpidella planifrons, Atractocybeloides berneri, Chasmops marginatus, Illaenus jewensis, Solenopora spongioides and Plumulites sp.Species which are predominantly but not exclusively found in assemblage 2 are Deaecheospira inflata, Brachytomaria baltica, Worthenia sp., Turbo balticus, Megalomphala contorta and Orthotheca sp. .
A few aspects of the flora and fauna of baksteenkalk have been described, but hardly any work has been done on large groups, such as brachiopods, bryozoans, bivalves and echinoderms.To date, species determined in baksteenkalk have never been listed in a systematic way.As a first attempt to fill this gap, a list of taxa which we were able to determine (or discern) in our collections is presented here (Table 2).
The species listed in Table 2 are from the collections of the authors, but for the sake of completeness (which is of course a relative notion) a few rare species known from other collections have also been included.The actual number of species which have left fossil remains in baksteenkalk is certainly much higher than the number of species listed here.This is due to several factors.First, part of the fossils could not be recognized for lack of the necessary expertise on our part.Second, for important groups, like crinoids and bryozoans, determination up to species level is hampered by their state of  preservation: crinoids have disintegrated and bryozoans are preserved as casts which do not reveal distinctive details.Third, part of the fossils probably represent new, hitherto undescribed taxa.
Many species have a very low abundance in baksteenkalk.In the list these too have been assigned to a particular type but given the low number of available specimens, it is impossible to be definite about their distribution over the two types.

INTERPRETATION Palaeoecology
The fossil assemblages are dominated by calcareous algae, bryozoans, crinoids, brachiopods and gastropods.These organisms represent benthic biotic communities.The prominence of calcareous algae implies that the seafloor was in the euphotic zone.The lithological differences between types 1, 2 and 3 of baksteenkalk reflect slightly different environmental conditions.

Type 1
This type formed from a seafloor composed of finegrained bioclastic calcareous mud.The unevenly shaped, unstratified blocks, with small fossils either being concentrated in nests or disorderly dispersed over the matrix, indicate a repeated shift of sediment.Fossils consist of seafloor debris because trilobites, crinoids and machaeridia are nearly always disintegrated.Post mortem overgrowth of fossils and marks of etching and burrowing organisms, which are abundant in type 2, are less frequent in type 1.
Trace fossils (e.g., Chondrites) and other signs of bioturbation are common but do not occur as strongly concentrated as in type 2. All this suggests that highenergy conditions, with periodical minor mud flows burying the seafloor, prevailed (Botting & Rhebergen 2011).There are no signs of winnowing.Interesting are occasional finds of Lingula; these suggest proximity of the tidal or littoral zone, where this inarticulate brachiopod is at home.

Type 2
In the flat slabs of type 2A fossils tend to be concentrated in thin layers or lenses.Often large fossils are clustered.Overall, slabs consist of bioclasts which are more coarse-grained than in type 1.These phenomena imply that the seafloor was subject to winnowing.Although cryptostomate bryozoans, crinoids and large calcareous algae like Mastopora concava are never preserved in full, sizable fragments of these organisms may be found.Also shells of gastropods, brachiopods and occasionally even trilobites may occur undamaged.This renders it likely that moderate energy conditions prevailed.Post mortem phenomena in type 2 indicate that calcareous organic remains rested for a prolonged period of time on the seafloor (or inside the upper, well oxygenated part of it) before being covered by sediment, otherwise it would have been impossible for other organisms to leave their marks on them.The sediment was constantly reworked and churned up by burrowing organisms; this is indicated by vague spots, filled burrows and lack of stratification (Van Diggelen 1983).The high degree of bioturbation in conjunction with the prolonged stay of organic remains on the seafloor indicates a low sedimentation rate.The flat bottom of slabs might be indicative of an interruption in depositional patterns.Such discontinuity surfaces are especially pronounced in hardgrounds (type 2C).Hardgrounds may have been attachment surfaces for trepostomate bryozoans and algae (Vinn & Toom 2015).It is possible that the layer of fossil remains in certain hardgrounds results from winnowing.The coquinites (type 2B) are genuine storm depositions.The obvious variation in energy conditions which is manifested in type 2 and its subtypes implies that the seafloor lay under the fair-weather wave base and above the storm wave base.

Type 3
Blocks of this type are penetrated by evenly distributed, hollow burrows.Although burrows may run in all directions, a preference for one orientation can be observed, suggesting bedding.In specimen A of Fig. 6 one straight side is actually covered by a thin crust without burrows.The fact that burrows are intact and some burrows can be seen to penetrate the entire bed perpendicular to the bedding plane implies that these boulders represent some kind of 'snapshot'.Sudden deposition of sediment may have abruptly stopped the activity of burrowing organisms.

Biostratigraphic implications
The fossil fauna and flora of baksteenkalk as a whole firmly points to the Haljala Stage (Idavere C III -Jõhvi D I ).Our collections do not contain species whose latest occurrence recorded is C I or C II .The upper boundary of baksteenkalk is more difficult to determine because the evidence for the presence of D II is debatable.
Two or three species of trilobites, only a few specimens of which have been found in type 2, seem to indicate the Keila Stage: Bolbochasmops bucculentus type 1 (D II a, Krueger 2013: 33, but D I according to Jaanusson 1945: 221, 222), Hemisphaerocoryphe granulatus (D II bp, Krueger 1994: 483, 484;cf. Dolgov & Meidla 2011: 628); Toxochasmops macrourus (D II b, Krueger 2013: 41).The possibility should be considered that these rare specimens are early (possibly, the earliest) representatives of the species involved.In that instance, their stratigraphical range would have to be brought down to (late) D I .
Type 1 is characterized by species of Apidium and Hoeegonites, but as these algae are not known from Estonia and Sweden, their occurrence cannot be correlated with regional stratigraphy.The trilobites observed together with Apidium pygmaeum provide diverging dates: Conolichas triconicus, C III a (Rõõmusoks 1970: 220); Nieszkowskia inermis, latest C III -earliest D I (Krueger 1995: 670); Nieszkowskia ahtioides, early D I ; Achatella kegelensis, D I (Rõõmusoks 1970: 247); Atractocybeloides berneri, D I (Krueger 1991: 225); Chasmops marginatus, Oculichasmops muticus and Illaenus jewensis, C III -D II (Rõõmusoks 1970: 247); Otarozoum peri, C III (Rõõmusoks 1970: 220).It is to be noted that allegedly 'late' elements, like the trilobite species assigned by Krueger to D II , are lacking in type 1.Overall, the fossil assemblage of type 1 indicates a stratigraphical range falling within the Haljala Stage, but not going beyond C III -early D I .
The differences in lithology and ecology noted above imply that the various types of baksteenkalk originated in slightly different environments on the open shelf (see above).Here the question presents itself whether the differences between types 1 and 2 are to be understood in terms of different environments only or whether they also reflect a chronostratigraphical difference.
Many species found in baksteenkalk are known from the Idavere and Jõhvi stages in Estonia and their fauna and flora supply the best reference material available for establishing the chronostratigraphical range of baksteenkalk.Jaanusson (1995) noted that in Estonia the boundary between these stages does not coincide with a distinct change in macro-and microfauna and moreover is situated within a lithologically uniform sequence.For that reason, he erected the Haljala Stage encompassing both the Idavere (C III ) and Jõhvi (D I ) stages.A substantial part of the Idavere Stage and the lower part of the Jõhvi Stage, the so-called Aluvere Zone (D I a), belongs to the same 'mittelordovizische Grundfauna' (Jaanusson 1945: 221, 222).Changes in the faunal composition between these stages are moreover continuous (Hints 1997).This does not alter the fact that many species in D I are 'new' in comparison with C III .Rõõmusoks (1970) listed 206 species for D I , of which 84 are not known from C III .This picture of gradual change within the framework of overall continuity also applies to baksteenkalk.As shown by the fauna list, the large majority of species is recorded for both type 1 and type 2. In rare instances, blocks of baksteenkalk hold a fossil assemblage combining species which are otherwise exclusive to one type.Thus, one block holds a specimen of Apidium pygmaeum alongside Hemisphaerocoryphe pseudohemicranium.Unusual combinations like these indicate that the transition in flora and fauna between types 1 and 2 occurred gradually.
Thus, against the background of the Estonian biostratigraphy in the Viruan, it may not seem inappropriate to differentiate chronologically between type 1 and type 2. As it is, type 1 appears to be slightly older than type 2. The fossil assemblage suggests a C III -D I a date for type 1 and, overall, a D I date for type 2.

Remarks on diagenesis
Various aspects of the diagenetic processes that gave shape to baksteenkalk are poorly understood.This is in particular true of the mechanisms of silicification and leaching.The complexity of these processes is aptly demonstrated by the variety of forms in which fossils, in particular the algae Apidium, Cyclocrinites and Coelosphaeridium, have been preserved (Van Keulen 2011, 2014).Specimens may show exceptional details in anatomy defying a simple explanation in terms of linear silicification and leaching processes.
All the same, the mode of preservation of most algae is such that it enables us to reconstruct the main stages of the diagenetic process.These stages are best exemplified by Coelosphaeridium sphaericum.In comparison with other species, specimens of this calcareous alga show a rather uniform manner of preservation and are common to both type 1 and type 2.
Coelosphaeridium sphaericum consists of a central axis with a bulbous end, bearing whorls of laterals which expand distally to form the surface of a sphaerical thallus (Fig. 12, specimen on the left).The central axis and, probably, the laterals were filled with soft organic matter (cytoplasm).At a late stage of growth, the void between the central axis and the laterals became calcified (Spjeldnaes & Nitecki 1990;Rhebergen 1994).All specimens preserved as fossils were originally calcified.
Specimens in baksteenkalk are mostly preserved as casts of the cavities that originally contained organic matter.Thus, an initial stage in the diagenetic process involves the decomposition of soft organic parts of dead organisms (1).The cavities that resulted were filled by fine-grained sediment (2).In our specimens the intralateral calcified parts have all been dissolved; roughly coeval specimens from impure limestones of the Oslo region in Norway (Furuberg Formation) show the same mode of preservation.As a consequence, the surrounding sediment must have solidified as limestone (3) before calcareous skeletal parts (aragonitic: brachiopods, hyolitha; calcitic: trilobites, bryozoans; both: algae, molluscs) dissolved (4).Otherwise the cavities mimicking the shape of the skeletal parts would not have survived.In a subsequent stage, the limestone became silicified (5).A mechanism of transport and accumulation of silica at low temperatures has been proposed by Landmesser (1995).The wide variety of modes of silicification of baksteenkalk fossils, in particular algae, points to a complicated process.We confine ourselves to stating that in the hard, unweathered cores of boulders and hardgrounds all specimens of Coelosphaeridium sphaericum are massively silicified, whereas in the weathered, soft and porous boulders they are invariably preserved as casts of fine-grained sediment due to the secondary dissolution of the intralateral silicifications (Figs 8,12,14).This indicates that boulders were subjected to leaching (6).
Geochemical analysis might help unravel the intricate mechanisms of the silicification and leaching processes during the formation of baksteenkalk.It is to be hoped that this analysis will be carried out in future.

Provenance and palaeogeography
In Baltoscandinavia, Ordovician bedrock is mainly preserved in intracratonic basins: Vänern, Vättern, the Southern Bothnian basin and the Central Baltic basin (e.g.Van Balen 1996).Part of the Ordovician bedrock is well exposed on the Swedish mainland, on Öland, Estonia and in the area around St Petersburg.Another part lies buried below the seafloor of the Baltic and the Bothnian Sea, and these Ordovician beds are only known from drill cores.To date, no silicified sediments with fossil assemblages akin to those of baksteenkalk have been reported from the intracratonic basins.
It is generally assumed that Ordovician sediments originally covered a much larger area.Large stretches of Ordovician deposits have vanished as a result of prolonged fluvial erosion by the Eridanos river system and by subsequent glacial erosion caused by the Pleistocene glaciations (Zeck et al. 1988).Thus, the slightly tilting Ordovician beds along the northern and northwestern edge of the Central Baltic Basin probably extended further northwards.There is even evidence that Middle Ordovician beds extended well into Northwest Finland (Uutela 1998).It is possible that baksteenkalk originates from deposits which nowadays have completely disappeared due to erosion.
Information about the probable source area of baksteenkalk may be derived from the distribution of erratics (geschiebes) in Baltica.Erratic boulders of massively silicified limestone containing remains of dasyclad algae and bryozoans are known from Gotland.The algae comprise Coelosphaeridium sphaericum, Mastopora concava and, rarely, Apidium sp., species which are common in baksteenkalk (Fig. 16).The boulders closely resemble hardgrounds of baksteenkalk (type 2C; Fig. 17).Although the exact relationship of boulders to baksteenkalk is difficult to establish, there is an obvious affinity.As the direction of the ice flows by which they were transported was south to southwest, both the Southern Bothnian and the Central Baltic Basin come into consideration as source areas.These boulders may be identical with the numerous Backsteinkalkgeschiebe of the Baltic type which (1989) reported from Gotland.
Another clue as to the probable provenance of baksteenkalk is provided by lithology, and in particular by the fossil assemblage.Outcrops and borings led Männil (1966) and Jaanusson (1976) to distinguish various lithofacies, arranged in belts, in the sediments deposited in the epicontinental sea of Baltica.These 'confacies belts', a term coined by Jaanusson, are each 'defined by a combination of litho-and biofacial characteristics' and have 'a fairly stable relative position within the depositional area' (Jaanusson 1976: 308) (Fig. 18).
The Central Baltoscandian Confacies Belt has a varied lithology and was in general deposited in deeper water than the eastern confacies belts (Jaanusson 1976: 309).The lithology of the North Estonian Confacies Belt is more uniformly characterized by shallow-water carbonate and fine-clastic sediments (Nestor & Einasto Fig. 18.Ordovician confacies belts in Baltoscandia.After Jaanusson (1995).1997).New data from borings in eastern Sweden caused Jaanusson to reduce the area covered by this belt by moving its western boundary eastwards (Jaanusson 1995).Given the lithological and faunal characters of this confacies belt, it constitutes a probable source area of baksteenkalk.
An important clue for an East Baltic origin of baksteenkalk is provided by the ostracod fauna.Although we have not carried out a systematic inventory of the ostracods occurring in baksteenkalk, we were able to determine Tetrada memorabilis, Bolbina ornata, B. minor, Oepikium tenerum, Kiesowia frigida, Pentagona pentagona.These species are characteristic of one particular ostracod fauna among the seven faunas Schallreuter (1970) distinguished in glacial Backsteinkalk.This fauna, type 1B13, contains several species which are only known from Estonia and which are restricted to the C III and D I stages.Schallreuter also noted that the ostracod Steusloffia costata, which in Sweden belongs to the common ostracods of D I sediments, is completely lacking in 1B13.All this led him to locate the origin of the 1B13 Backsteinkalk in the Baltic in the vicinity of Estonia (Schallreuter 1970(Schallreuter , 1993)).In baksteenkalk, species characteristic of 1B13 are recorded from both type 1 and type 2, whereas Steusloffia costata has not been observed in either type.
The nature of the relationship between the East Baltic 1B13-type Backsteinkalk and baksteenkalk can only be established by a full comparison of their fossil assemblages.This lies beyond the scope of the present study.Nevertheless, the character of the ostracod fauna lends forceful support to the view that baksteenkalk originates from the Baltic region to the west of Estonia.
Conditions in this region were probably similar to those described by Nestor & Einasto (1997) for the west of Estonia.During the Haljala and Keila stages, bioclastic calcareous muds were deposited on the open shelf.The production of skeletal material, which was the main source of the carbonate, was slow.Accordingly, the rate of sediment accumulation in the Baltic epicontinental basin was extremely low.Estimates vary from 1 to 3 mm per 1000 years.In line with these figures, the beds represented by our C III -D I baksteenkalk cannot have exceeded 6 m thickness, since the Haljala Stage lasted ca 2 Ma.In Estonia, however, the thickness of the Haljala Stage varies from 10 to 20 m (Jaanusson 1945;Hints 1997), which seems to indicate a somewhat higher rate of sedimentation.
Several interbeds of volcanic ash (metabentonites) have been reported from the Haljala and Keila stages throughout the Baltic.Drill cores in the west of Estonia (including Saaremaa and Hiiumaa) have revealed a thickness of 0.4-0.6 m for the Kinnekulle bentonite (K-bentonite) lying on top of the Haljala beds (Huff et al. 1992;Bergström et al. 1995).The Jõhvi and Idavere successions of these drill cores were found to contain other bentonite beds of varying thicknesses.Backsteinkalk from Sweden and the Baltic Sea is thought to have formed by the silicification of limestone under bentonite beds (Neben & Krueger 1973;Schallreuter 2005).A similar process might be assumed for baksteenkalk.
In Estonia, coeval carbonates under bentonite beds do not exhibit the same measure of silicification.Jaanusson (1945) reported silicified fossils from the Idavere and Jõhvi beds.Yet continuous beds of silicified Ordovician sediments do not seem to occur in Estonia.Towards the east of Estonia, bentonite beds gradually decrease in thickness, but this alone does not sufficiently explain the difference between the largely unsilicified limestone beds in Estonia and the completely silicified Backsteinkalk/baksteenkalk.Apparently, variations in environmental conditions played a part as well.Recently Bartholomäus et al. (2014) have pointed to the occurrence of 'cauliflower cherts', i.e., siliceous concretions, in Haljala limestone from Põõsaspea Cliff.They interpret these cherts as diagenetically silicified anhydritic nodules which originated in Sebkha-like conditions.According to Landmesser (1995), locally high Si(OH) 4 concentrations may form by evaporation.Conditions may have been favourable for the formation of an evaporitic environment during the Keila age, when the sea level dropped (Nestor & Einasto 1997).As the sea retreated to the south, areas to the north and northwest of Estonia were uncovered and possibly became evaporitic.Regional differences with respect to the duration of the regression phase were probably of influence on the silicification process.Yet, baksteenkalk itself does not yield evidence that it formed in Sebkha-like conditions.All we can say is that environmental conditions in the area where baksteenkalk originated seem to have been more favourable for the silicification of limestones than in the north of Estonia.
Besides bentonite, biogenic sources of silica have been considered.Jaanusson (1945) supposed a relationship between the silicified fossils from the Idavere and Jõhvi beds and the distribution of the so-called 'Kieselspongienfazies'.Bartholomäus et al. (2014) reckon with the possibility that the silicification process derived most SiO 2 from siliceous spicules of sponges, in particular of hexactinellids.However, it seems improbable that siliceous spiculae could have supplied the amount of SiO 2 required for the silicification of such a large volume of limestone.Genuine beds of spiculite have not been observed in the Middle and Upper Ordovician of Baltica.In baksteenkalk and other erratic Ordovician silicifications, spiculae, mostly monaxons and hexactines, are common but never occur in massive quantities.
During the Ordovician, the Baltica continent drifted from southern high latitudes to the equator.In Haljala and Keila times, the epicontinental cratonic sea was still in the temperate climatic zone (at ca 35° S).Tabulates and stromatoporoids, which are bound to high water temperatures, had not yet made their appearance.As a consequence, these groups are not represented in baksteenkalk.
Given this state of affairs, a sophisticated palaeontological comparison between the various types of baksteenkalk and Backsteinkalk is impossible.Yet, on the basis of the information that can be gathered from the above-mentioned publications and from personal communications, it is possible to make the following observations.
The majority of the species determined in Backsteinkalk have also been recognized in baksteenkalk.The species found in Backsteinkalk but not in baksteenkalk include Allolichas longispinus, Chasmops wrangeli, Sinuites sp.Conversely, the fauna and flora of baksteenkalk may comprise species which do not occur in Backsteinkalk.Although the absence of evidence is no evidence of absence, it may be telling that Hoeegonites kringla, Solenopora spongioides and Vermiporella fragilis, species which are frequent in type 1 baksteenkalk (cf.Rhebergen 1997), have never been reported from Backsteinkalk.
Several taxa are known to have a different abundance in Backsteinkalk and baksteenkalk.The trilobites Keilapyge laevigata and species of Lonchodomas and Neoasaphus are more frequent in the former than in the latter (Rhebergen 2001).Conversely, Atractopyge berneri (Rhebergen 2001) and Nieszkowskia ahtioides (Krueger 1995) are rather frequent in baksteenkalk but rare in Backsteinkalk.The same seems to apply to the brachiopod Platystrophia chama (H.-H. Krueger pers. comm. 1995).As to the algae, Apidium pygmaeum is depicted in Neben & Krueger (1979: 44), but the fact that it is mistaken for Cyclocrinus (= Cyclocrinites) strongly suggests that this alga, which is ubiquitous in type 1 baksteenkalk, is not common in Backsteinkalk.This raises the question of whether Apidium pygmaeum is tied to the Baltic types of Backsteinkalk, which according to Schallreuter (1984) are rare west of the Oder.
Questions like these can only be answered on the basis of a typology of Backsteinkalk and a systematic inventory of its fossils.It is to be hoped that these will be carried out in future.

Comparison with Lavender-blue Hornstein
Another group of silicified erratics comprises boulders of Lavender-blue Hornstein (LBH), as mentioned in the Introduction.They fall into two different components, one of Sandbian age, the other of Katian (Pirgu to Porkuni) age.The Sandbian component ranges from the Jõhvi Stage to the Keila Stage.Van Keulen et al. (2012) described several aspects of LBH, such as typology, fossil assemblage, distribution in Europe and provenance.This study also includes a brief comparison between Sandbian LBH and baksteenkalk erratics.The results are not repeated here, but two aspects may be worthwhile mentioning.
The first aspect concerns a marked difference in the composition of the algal flora and the fauna of sponges.In coeval erratics of LBH, Apidium pygmaeum is absent, whereas Cyclocrinites is represented by more species than in baksteenkalk.Conversely, several genera of astylospongiid sponges which are frequent in LBH of Sandbian age, such as Astylospongia, Carpospongia and Caryospongia, are not found in baksteenkalk.Astylospongiid species characteristic of the 'blue' Lausitz-Sylt sponge association sensu Rhebergen & von Hacht (2000) occur in situ in the Haljala beds of the St Petersburg region and Keila beds of the Ristna Klint in Northwest Estonia.The former location has yielded Carpospongia pogrebowi, C. conwentzi and C. roemeri, the latter Syltrochos syltensis (Rhebergen 2009) and Carpospongia castanea (Krueger 2003).Also in other respects, the fossil assemblage of the D II a strata of the Ristna Klint bears resemblance to that of erratic LBH of Sandbian age (Krueger 2003).
The second aspect concerns the geographical distribution of the Sandbian LBH erratics.The majority derive from Miocene and Pliocene fluvial deposits of the Eridanos.Lavender-blue Hornstein erratics are not known from Gotland and the mainland of Sweden.As a consequence, their provenance is likely to differ from that of baksteenkalk and Backsteinkalk.Considering the similarities between the 'blue' erratic sponges and the coeval sponge assemblages in the St. Petersburg area and Ristna Klint, the Haljala/Keila LBH erratics may originate from strata which during the Neogene have been subjected to massive erosion by a tributary of the Eridanos, the Pra-Neva (Suuroja 2007).

Fig. 1 .
Fig. 1.Geographical map of the area from which most specimens of baksteenkalk have been collected.The highest concentrations occur in Early Pleistocene fluvio-glacial deposits which are exposed in Late Pleistocene (Saalian) ice-pushed ridges, the so-called WWW-area (grey).

Fig. 7 .
Fig. 7. Tube of porous debris with a core of chalcedony, a common trace fossil in type 1. RGM 1317644, Wilsum.

Fig. 12 .
Fig. 12. Natural surface of a fragment of a type 2A boulder showing specimens of Coelosphaeridium sphaericum and a transverse section of a gastropod.RGM 1317649, Itterbeck.

Fig. 13 .
Fig. 13.Natural surface of a fragment of a type 2A boulder.Bottom: a specimen of Cyclocrinites porosus showing the membrane under the vanished globella; right: orthid brachiopod, interior mould of ventral valve.RGM 1317650, Wilsum.