2. Geological setting and stratigraphy

In Scania (Fig. 1), the lower–middle Cambrian boundary interval displays a broad transition from regressive to transgressive conditions. These are represented by a stratigraphic evolution from proximal coarse-grained siliciclastic strata to distal kerogen-rich shales, which are punctuated by discrete carbonate beds. In ascending order, the transitional interval is represented by the Rispebjerg Member (upper Læså Formation), the Gislöv Formation and the lowermost part of the Alum Shale Formation, including the Forsemölla and Exsulans Limestone beds (for a recent lithostratigraphic revision, see Nielsen & Schovsbo, Reference Nielsen and Schovsbo2007; Fig. 2). The Rispebjerg Member is a thin, 1–3 m, homogeneous succession of calcareous medium- to coarse-grained quartz sandstone, with phosphorite clasts mainly restricted to the basal conglomerates, and impregnation of authigenic phosphorite in its uppermost part (de Marino, Reference Marino1980a; Bergström & Gee, Reference Bergström, Gee, Gee and Sturt1985; Pedersen, Reference Pedersen1989). Except for some fragmentary shelly fossils, the only fossils observed in Scania are ichnofossils near the top of the formation (Lindström & Staude, Reference Lindström and Staude1971; Bergström & Ahlberg, Reference Bergström and Ahlberg1981). The Rispebjerg Member is unconformably overlain by the Gislöv Formation, a thin (0.7–5.7 m) heterolithic unit that consists of fossiliferous silty limestone, siltstone and claystone. It is rich in glauconite and phosphatic clasts (de Marino, Reference Marino1980b), and has yielded trilobites indicative of the Holmia kjerulfi and Ornamentaspis? linnarssoni assemblage zones (late early Cambrian: Bergström & Ahlberg, Reference Bergström and Ahlberg1981). The contact between the Gislöv and the Alum Shale formations is laterally variable: at Forsemölla (Figs 1, 2), the contact is situated at the base of a medium grey to black shale (the so-called ritskiffer or ‘drawing shale’; middle Cambrian according to Nielsen & Schovsbo, Reference Nielsen and Schovsbo2007), which is about 50 cm thick and has yielded a few linguliformean brachiopods and hyoliths (Bergström & Ahlberg, Reference Bergström and Ahlberg1981). In other outcrops, the contact is a simple discontinuity surface underlying the Forsemölla Limestone Bed (Nielsen & Schovsbo, Reference Nielsen and Schovsbo2007). This limestone unit has yielded paradoxidid remains belonging to Hydrocephalus sp. (Forsemölla section) and Paradoxides paradoxissimus (Brantevik section: Bergström & Ahlberg, Reference Bergström and Ahlberg1981), suggesting the presence of a stratigraphic gap that is broadly equivalent to the Acadoparadoxides oelandicus Superzone (Fig. 2).

Figure 1. Map of Scania, southern Sweden, showing major Cambrian outcrop areas (black) and the location of Brantevik, Forsemölla, Hardeberga, and the Almbacken borehole; modified from Martinsson (Reference Martinsson and Holland1974, fig. 7).

Figure 2. Stratigraphic units of the lower–middle Cambrian transition in Scania; modified from Nielsen & Schovsbo (Reference Nielsen and Schovsbo2007).

The middle Cambrian part of the Alum Shale Formation has a thickness of 20–40 m and consists of dark grey to black, kerogen- and sulphide-rich shale with lenses or beds of dark grey limestone. They accumulated over long periods of time, changing from middle Cambrian oxygenated to Furongian dominantly dysoxic, anoxic or euxinic conditions (A. T. Nielsen, Copenhagen, pers. comm. 2009). Four distinctive and correlatable, middle Cambrian limestone beds, all less than 1 m thick, are interbedded within this monotonous clayey succession: the Forsemölla (formerly referred to as ‘fragment’), Exsulans, Hyolithes and Andrarum Limestone beds (Ahlberg, Reference Ahlberg and Ahlberg1998; Ahlberg & Bergström, Reference Ahlberg, Bergström and Ahlberg1998; Axheimer & Ahlberg, Reference Axheimer and Ahlberg2003; Nielsen & Schovsbo, Reference Nielsen and Schovsbo2007).

From a biostratigraphic point of view, sediments attributed to the Holmia kjerulfi Assemblage Zone and correlative strata in Poland (Holmia Zone) display an association of olenellid and ellipsocephalid trilobites, the former dominated by Holmia and Kjerulfia, and the latter by Ellipsocephalus, Proampyx?, Ornamentaspis? and Strenuaeva. The presence of eodiscoids (Calodiscus lobatus and Runcinodiscus cf. index) in the Holmia kjerulfi Assemblage Zone suggests a correlation with strata containing representatives of the Serrodiscus bellimarginatus–Triangulaspis annio–Hebediscus attleborensis assemblage (Robison et al. Reference Robison, Rosova, Rowell and Fletcher1977; Ahlberg, Reference Ahlberg1983; Geyer, Reference Geyer2005) in the North Atlantic region and Siberia, for example, with the ‘Callavia’ Zone in Avalonia and the middle Botoman substage in Siberia.

The latest early Cambrian zone in Scandinavia, the Ornamentaspis? linnarssoni Assemblage Zone (previously referred to as the Strenuella linnarssoni Zone), was established by Kiær (Reference Kiær1917) for a succession of shales and a few interbedded limestones (the Evjevik Limestone) that overlie the Holmia Shale (Tømten Member of Nielsen & Schovsbo, Reference Nielsen and Schovsbo2007) in the Mjøsa area, southern Norway. This zone has been recorded with certainty only in southern Norway and Scania. In Scania, the base of the zone is placed at the first appearance of O.? cf. linnarssoni (Kiær, Reference Kiær1917) within the middle part of the Gislöv Formation (Bergström & Ahlberg, Reference Bergström and Ahlberg1981). The trilobites of the zone are dominated by ellipsocephalids, but also include rare specimens of olenellids (Holmia kjerulfi and Kjerulfia lata: Ahlberg & Bergström, Reference Ahlberg and Bergström1978; Nikolaisen, Reference Nikolaisen1986). The zone may be tentatively correlated with the lower part of the Protolenus Zone in Poland (Moczydłowska, Reference Moczydłowska1991).

A few pygidia of paradoxidid trilobites have been recovered from the Forsemölla Limestone Bed, which seems to represent the lowermost part of the middle Cambrian Paradoxides paradoxissimus Superzone (Bergström & Ahlberg, Reference Bergström and Ahlberg1981). This limestone unit has also yielded fairly rich faunal assemblages, including trilobites, linguliformean brachiopods, echinoderm ossicles, hyolithids, helcionellids, lapworthellids, hyolithelminthids, bradoriids, palaeoscolecids, phosphatocopids and conodont-like elements (Bengtson, Reference Bengtson1976; de Marino, Reference Marino1980b; Streng, Geyer & Budd, Reference Streng, Geyer and Budd2006; Streng et al. Reference Streng, Holmer, Popov and Budd2007). The middle Cambrian Acadoparadoxides oelandicus Superzone is generally absent in Scania and corresponds to a hiatus between the Gislöv Formation and the Forsemölla Limestone Bed. Locally, however, a succession of dark grey to black shales is present between the top of the Gislöv Formation and the base of the Forsemölla Bed, for instance, in the Almbacken drill core (Axheimer & Ahlberg, Reference Axheimer and Ahlberg2003). This shale unit was assigned to the A. oelandicus Superzone by Nielsen & Schovsbo (Reference Nielsen and Schovsbo2007).

3. Facies associations

Three sections and one drill core have been studied in order to characterize the facies associations of the Gislöv Formation and the Forsemölla and Exsulans Limestone beds: (1) the stratotype of the Gislöv Formation, about 1 km SW of Brantevik in eastern Scania, (2) the Forsemölla section in eastern Scania, (3) a section through the upper lower Cambrian at the Hardeberga quarry in south-central Scania (Bergström & Ahlberg, Reference Bergström and Ahlberg1981) and (4) a drill core from Almbacken, east of Lund in south-central Scania (Axheimer & Ahlberg, Reference Axheimer and Ahlberg2003) (Fig. 1). The lithological succession and the facies associations in the outcrops were briefly described by de Marino (Reference Marino1980a, Reference Marinob) and Bergström & Ahlberg (Reference Bergström and Ahlberg1981), who also provided references to earlier studies dealing with the lower–middle Cambrian transition in Scania. Graphic logs and correlation lines are shown in Figure 3.

Figure 3. Stratigraphic logs of the study areas, with reference to Bergström & Ahlberg's (1981) nomenclature, and reconstruction of the sedimentary architecture of the Gislöv and Forsemölla units.

3.a. Bioclastic limestone bearing millimetre-scale skeletons (Gislöv Formation)

The limestone consists of well-bedded to partly nodular wackestone to packstone, usually thin-bedded and slightly to moderately burrowed (Fig. 4a, b). Centimetre-thick fining-upward units, resting on scoured surfaces, are common (Fig. 4c). Intra-bed irregular surfaces are commonly lined with anastomosing swarms of clay seams, up to 1 cm thick. Trilobites and calcitic brachiopods are abundant, whereas echinoderm ossicles, linguliformean brachiopods, chancelloriid sclerites, sponge spicules, bradoriids (Dies Álvarez et al. Reference Dies Álvarez, Gozalo, Cederström and Ahlberg2008) and other skeletonized microfossils of uncertain affinity are subsidiary. The fossils are predominantly fragmented and abraded, and commonly display a chaotic (random orientation) fabric, although they are locally imbricated or arranged in crude low-angle laminae. The content of silt- to fine-grained sand-sized siliciclastic material is highly variable, and rare glauconitic and phosphatic clasts are encountered in thin-section.

Figure 4. Thin-section photomicrographs of the Gislöv Formation, and the Forsemölla and Exsulans Limestone beds. (a) Millimetre-scale burrow (arrowed) partly infilled with silt-grained quartz and a darker mixture of calcite and iron oxides; Bergström & Ahlberg's (Reference Bergström and Ahlberg1981, fig. 6) level A at Brantevik; scale bar = 1 mm. (b) Portion of the Almbacken drill core (depth: 31.80 m) showing the mottled aspect of the limestones in the Gislöv Formation due to burrowing; scale bar = 2 cm. (c) Millimetre-thick, packstone-to-wackestone, grading units, underlain by erosive surfaces (arrowed), and embedded in a micritic matrix; lower part of Bergström & Ahlberg's (Reference Bergström and Ahlberg1981, fig. 9) level C at Forsemölla; scale bar = 1 mm. (d) Cross-cutting feature of a solution vug, occluded by a mosaic of calcite crystals, and subsequently microstylolitized; Bergström & Ahlberg's (Reference Bergström and Ahlberg1981, fig. 6) level C at Brantevik; scale bar = 1 mm. (e) Ravinement deposit of Bergström & Ahlberg's (Reference Bergström and Ahlberg1981, fig. 6) level E at Brantevik, showing a trilobite-dominated packstone and dispersed apatite (ap), glauconite (gl), pyrite (py), and quartz (Q) clasts; scale bar = 1 mm. (f) Same level exhibiting two phosphate-walled linguliformean fragments; scale bar = 0.5 mm. (g) Centimetre-long sclerites of trilobites underlain by shelter porosities occluded with mosaics of calcite crystals; base of Exsulans Limestone Bed at Brantevik; scale bar = 2 mm. (h) Packstone rich in trilobites (t), brachiopods (b), echinoderm ossicles (e) and sponge spicules (s) embedded in a pyritic cement; bioclasts covered by fibrous calcite cements; top of Exsulans Limestone from core drilling at Almbacken; scale bar = 2 mm. (i) Detail of conglomeratic lag lining a channel, containing quartz arenitic (qa), litharenitic and pyritic (py) granules embedded in a litharenite cemented by calcite; base of Bergström & Ahlberg's (Reference Bergström and Ahlberg1981, fig. 10) level B at Hardeberga quarry; scale bar = 2 mm.

The local abundance of fenestral-like cavities may impart a mottled aspect to the Gislöv Formation. Primary pore types include intra- and interparticulate shelter pores, up to 3 mm in size. The latter are lined with early isopachous fibrous calcite crystals, up to 30 μm in length, and later by mosaics of inclusion-rich calcite cements. Equant sparry calcite cements (anhedral to subhedral crystals, 50–100 μm in size) occur geopetally in mouldic pores, and syntaxial cements commonly surround the echinoderm ossicles. Spar-filled fenestral-like vugs, cutting both matrix and skeletons, are common in some strata (Fig. 4d).

The presence of low-angle lamination and millimetre-scale fragmented bioclasts, arranged in graded units and underlain by erosive contacts, suggests that the substrate was winnowed, probably by the action of waves and storm episodes. The patchy abundance of bioturbation associated with infaunal deposit feeders indicates that benthic colonization of the sea-floor was not inhibited by synsedimentary or early diagenetic cementation. The diversity of the shelly fauna suggests that deposition took place in an open-sea environment of normal salinity. Partial dissolution is suggested by the presence of a secondary vuggy porosity, recognizable by the embayed morphology of the pores and their cross-cutting of pre-existing bioclastic fabrics. This reflects circulation of undersaturated waters. As the vuggy cavities were apparently not connected with the surface, precluding a karstic origin, they must be associated with other solution processes. Finally, anastomosing swarms of clay seams represent late-diagenetic dissolution of carbonate and relative enrichment of clay displaying solution-seam fabrics.

4. Palaeoenvironments and stratigraphic discontinuities

According to a proximal–distal facies distribution, proximal facies are considered to be at the Hardeberga quarry (WNW), whereas distal sediments should be situated at Brantevik (ESE). The pattern of deposition across the lower–middle Cambrian transition can be subdivided into four asymmetric deepening-upward trends (or parasequences), 10–100 cm thick, marked by regionally developed flooding events associated with erosive discontinuities (here named D1 to D5; Figs 2, 3).

Discontinuity D1 is located at the Rispebjerg/Gislöv contact, and marks a prominent lithological change. It is recognized as an erosive surface that marks a sharp change from a siliciclastic platform, represented by the amalgamated channels and shoals of the Rispebjerg Member (de Marino, Reference Marino1980a; Bergström & Gee, Reference Bergström, Gee, Gee and Sturt1985; Pedersen, Reference Pedersen1989), to a mixed (carbonate–siliciclastic) platform covered by shell pavements that gradually deepened (TST₁), grading into distal shales.

Discontinuity D2 subdivides the Gislöv Formation in two parts. It appears in outcrop as a rectilinear erosive surface that includes local undulating relief of up to 15 cm and occasional steep, centimetre-scale downward-cutting steps. Both basal lags and erosive tool marks are observed. D2 covers limestones in proximal outcrops and shales in distal ones. The succeeding deepening-upward parasequence (TST₂) consists of different facies in a proximal–distal transect; proximal outcrops (Hardeberga and Almbacken) show a siliciclastic-dominated unit, which grades from a sandstone, bearing calcite cement in patches and layers, that wedges out eastward disappearing in less than 50 km, to a silty limestone that locally passes into deeper shales.

Discontinuity D3 is a conspicuously smooth erosive unconformity with an undulating topography. It separates the Gislöv and Forsemölla units and marks the appearance of conspicuous linguliformean brachiopods (see Streng et al. Reference Streng, Holmer, Popov and Budd2007), and glauconite and phosphate peloids and clasts. The surface is associated with a substantial hiatus where the entire Acadoparadoxides oelandicus Superzone is generally missing. Based on these features, the formation of D3 is attributed to a sharp relative sea-level fall (although subaerial exposure features are absent) and erosion, followed by transgressive abrasion of an early-cemented shell substrate. The overlying ‘E level’ (sensu Bergström & Ahlberg, Reference Bergström and Ahlberg1981, fig. 6; Fig. 3 herein), up to 2 cm thick at Brantevik and underlain by an erosive ravinement surface, is composed of polyphase intraclastic elements, abundant glauconite and phosphate clasts. The onlapping level is associated with a distinct drowning surface and was overlain by TST₃ (Eriksson, Reference Eriksson2004; Catuneau et al. Reference Catuneau, Abreu, Bhattacharya, Blum, Dalrympe, Eriksson, Fielding, Fisher, Galloway, Gibling, Giles, Holbrook, Jordan, Kendall, Macurda, Martinsen, Miall, Neal, Nummedal, Pomar, Posamentier, Pratt, Sarg, Shanley, Steel, Strasser, Tucker and Winker2009). The transgressive unit consists of another condensed shelly limestone composed of densely packed phosphatized clasts and crudely laminated skeletal packstone/wackestone embedded in a glauconitic-rich matrix.

Discontinuities D4 and D5 are major hardgrounds that display irregular surfaces with reliefs up to 10 cm, and mark the top of the Forsemölla and Exsulans Limestone beds, respectively. They are characterized by partial dolomitization, which decreases downward, accumulation of phosphatic clasts and ubiquitous pyrite–phosphate nucleation. Dolomite and pyrite clasts are incorporated into the lowermost centimetres of the overlying black shales, suggesting early diagenetic dolomitization and pyritization processes. Both hardgrounds formed during stepwise drowning, leading to the demise of carbonate production and deposition of kerogenous shales (Alum Shale), representing deeper (darker, quieter and cooler) waters.

As a result, the onset of the four episodes of carbonate productivity: (1) took place in the central part of a mixed platform, where the nucleation of carbonate factories was bracketed between proximal, shallow-marine areas dominated by terrigenous deposition and distal, deeper-water areas where the sedimentation of clays dominated; (2) represents retrogradational (transgressive) carbonate sands, which were successively transported landward, covering increasingly larger sectors of the platform; and (3) ended by the onset of either erosive (D2 and D3) or hardground (D4 and D5) discontinuities. Although the limited lateral extension of exposures precludes determination of large-scale sedimentary geometries, carbonate belts seemingly migrated as extensive, low-relief sheets or shoals, consisting of sediments dominated by skeletal remains that became finer with increasing distance from shoals and increasing water depth. Wave and current activity varied from just sufficient to disarticulate and transport skeletal material locally (e.g. Exsulans Limestone Bed) to sufficiently intense to accumulate beds of broken, millimetre-sized skeletal material (Forsemölla Limestone Bed). Evidence of physical abrasion in shells is mostly confined to the high-energy nearshore sediments of TST₁, TST₂ and TST₃. Relatively slow sedimentation rates for the limestones are indicated by microboring, and phosphate, pyrite and glauconite precipitation.

At the top of the Forsemölla (TST₃) and Exsulans (TST₄) units, a shift to a long-lived dysoxic–anoxic event is represented by a middle Cambrian–Furongian interval enriched in organic matter (kerogenous Alum shales), episodically punctuated by high-energy events and nucleation of carbonate productivity. The record of D4 and D5 suggests sudden and substantial deepening of the platform. As this deepening approximately coincided with a gradual sea-level rise in West Gondwana (Álvaro & Clausen, Reference Álvaro and Clausen2005), the stepwise character of the platform drowning in Scania, and the presence of complex internal and laterally rapidly changing, centimetre-thick sedimentary packages, multiple generations of intraclasts, and intra-bedded diastems of the Gislöv and Forsemölla units suggest a different subsidence regime. As noted above, these limestone beds can be traced more than 400 km north of Scania, where they maintain similar thicknesses, attesting to the flatness of the sea-floor in Baltica during Cambrian times. The kerogenous Alum shales represent the onset of a generalized drowning throughout NW Baltica. This phase of relative sea-level rise may have been important for the establishment of new biogeographic connections leading to the onset of the middle Cambrian Acado-Baltic Province (sensu Sdzuy, Reference Sdzuy1972) and also for the exchange of nutrient-enriched waters.

5. Microboring of shells

In thin-section, most skeletal carbonate grains of facies associations 3.1 and 3.2 show a halo of micrite affecting the outer margins as a result of complete microboring. In some cases, severe alteration has even produced entirely micritized grains. Some haloes show distinct microboring holes, occasionally in dense concentrations. These holes are either filled with carbonate micrite or microsparite, or a combination of both. Three ichnomorphotypes can be identified:

1. (1) Simple, cylindrical, straight to slightly curved microborings with a single, circular aperture, up to 50 μm in diameter and 300 μm long, and transverse to the skeletal wall. These are commonly unbranched microborings, although some simple branching is also observed (Fig. 5a, b).

2. (2) Straight to sinuous, cylindrical microborings subparallel to the skeletal wall, displaying similar size ranges (Fig. 5c). Their ichnotaxonomic assignment is uncertain, as no complete sections connected with the surface have been observed.

3. (3) Scattered globular microborings, up to 200 μm in diameter, disconnected in thin-section (Fig. 5d), although their clusters are probably interconnected and represent incompletely preserved filamentous microborings of larger diameter than those described above.

Figure 5. Thin-section photomicrographs of microboring textures from Bergström & Ahlberg's (Reference Bergström and Ahlberg1981, fig. 6) level E at Brantevik; all scale bars = 1 mm. (a) Mottled aspect of trilobite sclerites. (b) Detail of boxed area showing straight and Y-shaped (arrowed) microborings transverse to the skeletal wall. (c) Straight microboring parallel to skeletal wall. (d) Globular microboring, probably representing transverse sections of straight microborings. Trilobite sclerites are subdivided into a prismatic outer layer (pl) and the principal homogeneous layer (hl), the latter preferentially microbored.

The microboring organisms were host-selective. Microboring is not observed in bioclasts preserved as sparry mosaics with blocky or isopachous bladed textures, formed as a result of dissolution of unstable minerals like aragonite. Trilobite sclerites are predominantly microbored (up to 60% of observed skeletons in some thin-sections), whereas calcitic brachiopods are rarely microbored, and echinoderm ossicles and linguliformean brachiopods show no microboring features. In addition, trilobite sclerites are not uniformly microbored. The cuticular microstructure of some trilobite sclerites consists of two layers: (1) a ‘prismatic’ outer layer, up to 30 μm thick with ornamentation features such as ridges and crests and (2) a principal homogeneous layer, 200 to 300 μm thick. Microboring developed predominantly in the inner homogeneous layer, reflecting its post-mortem character. Variation in microboring size appears to reflect differences in the maturity and composition of the microboring community, presumably composed of endolithic algal, fungal and/or cyanobacterial populations that rapidly infested the shell substrates. Chazotte, Le Champion-Alsumard & Peyrot-Clausade (Reference Chazotte, Champion-Alsumard and Peyrot-Clausade1995) suggested that endolithic microbial infestation is very rapid in shallow water, occurring within weeks to two months.

A secondary effect of microboring was the weakening of shells, resulting in later mechanical breakage. Hence, the mixture of millimetre- and centimetre-sized bioclasts was not only controlled by hydraulic selection but also by post-mortem microboring and reworking. Since boring organisms require an onset of non-deposition intervals (Bromley & Asgaard, Reference Bromley and Asgaard1993), storm-dominated sedimentation may have controlled bioerosion, limiting the residence time of eligible hard substrates. As a result, microboring was controlled by changes in grain size, the nature of the skeletal mass bored, the rate of sediment supply, depth and frequency of storms, and the activity of infauna and epibenthos (Bromley, Reference Bromley and McIlroy2004; Glaub & Vogel, Reference Glaub and Vogel2004; Glaub et al. Reference Glaub, Golubic, Gektidis, Radtke, Vogel and Miller III2007). Lower boring frequency is directly associated with enhanced siliciclastic influx, which would have increased the burial rate of host substrates and reduced their exposure time in the taphonomically active zone. On the contrary, a high density of microboring was regulated by residence time of shelly hard substrates.

6. Precipitation of fluorapatite, glauconite and pyrite

As shown in Figure 3, glauconite and fluorapatite clasts are more abundant in strata overlying the ravinement surface, at the base of the Forsemölla Limestone Bed, and in scattered conglomeratic levels. Three kinds of phosphate debris can be distinguished in the studied units:

1. (1) Silt- to sand-sized homogeneous clasts that are subrounded to subangular in shape, and made up of amorphous collophane.

2. (2) Composite clasts that are subrounded to subangular in shape, up to 6 cm long, and display a chaotic (randomly oriented) texture (Fig. 6a). They contain a complex mixture of silt to medium-grained sand-sized allochems, composed of quartz, feldspar, mica flakes, pyrite and other opaques, rip-up shale clasts, glauconite peloids, other heterogeneous phosphatic intraclasts and bioclasts. The latter are largely composed of trilobite sclerites, brachiopod valves, hexactinellid sponge spicules and chancelloriid sclerites (Fig. 6b). Thin-sections reveal that the phosphatization involved occlusion of intra- and inter-particle porosity, rim cement overgrowth, and partial replacement of interparticle micrite by fluorapatite and collophane.

3. (3) Clasts bearing phosphatized microbial pseudomorphs. Two kinds of microbial populations are distinguished: (1) microbial biofilms or microstromatolites that grew on inter-particle pores produced complex phosphatic crusts and anisotropic phosphatic coatings (Fig. 6c, d) and (2) dispersed void-filling botryoidal (globular clusters) textures similar to ovoidal to rod-like microbial colonies (up to 5 μm in diameter). The excellent textural preservation shows that the concentration and precipitation of phosphate occurred prior to and independently from the decay of the microbial organic matter, during a very early stage of diagenesis.

Figure 6. Thin-section photomicrographs of phosphoritic clasts. (a) Polyphase clast bearing homogeneous clasts (ho) encrusted by apatitic lamellae, heterogeneous clasts (he), engulfed quartz (Q), four-lobed mosaics of micro- and cryptocrystalline silica (similar to volcanic glass shards; m), and glauconite; level B of the Gislöv Formation at Forsemölla; scale bar = 1 mm. (b) Detail of a heterogeneous phosphoritic clast including sections of chancelloriids (ch) and hexactinellid sponge spicules (ss), quartz and glauconite (gl) clasts; level E of the Gislöv Formation at Brantevik; scale bar = 500 μm. (c) Bioclastic–peloidal packstone partly replaced into apatite: original peloids and bioclasts are wholly replaced by apatite, whereas the remaining voids were occluded by apatite, calcite and pyrite; level B of the Gislöv Formation at Forsemölla; scale bar = 2 mm. (d) Detail of cements occluding primary pores, composed of microbial crusts of apatite (displaying both microstromatolitic and globular textures), and mosaics of calcite; level B of the Gislöv Formation at Forsemölla; scale bar = 2 mm.

Homogeneous clasts were sourced from centimetre-thick hardground crusts or nodules, whereas the polymictic and polyphasic character of the heterogeneous clasts indicates that discrete phosphogenic episodes alternated with reworking and depositional events. Coeval reworking of pyrite (formed under reducing and sulphidic conditions below the sea-floor/water interface) is shown by mechanical breakage and the polyphase nature of some composite pyrite agglomerates, and that of heterogeneous clasts by growth of gradually thickened fluorapatite coatings or rims around single or aggregated bioclasts and lithoclasts.

Two processes must be distinguished: phosphogenesis and phosphate concentration. Phosphogenesis is a biogeochemical process governed by microbially mediated Eh and pH of bottom and pore waters, dissolved chemical elements and sedimentary rates, whereas the concentration of the authigenic minerals is related to hydraulic reworking and/or winnowing processes (Krajewski et al. Reference Krajewski, Van Cappellen, Trichet, Kuhn, Lucas, Martín-Algarra, Prévôt, Tewari, Gaspar, Knight and Lamboy1994; Föllmi, Reference Föllmi1996; Sagemann et al. Reference Sagemann, Bale, Briggs and Parkes1999; Briggs, Reference Briggs2003). In fact, the succession of repeated cycles of sedimentation, phosphate concentration and reworking led to multi-event phosphate deposits rich in heterogeneous clasts. We concur with Notholt & Brasier (Reference Notholt, Brasier, Cook and Shergold1986) that the major source of phosphorus in the Cambrian of Scania is associated with upwelling waters. According to palaeomagnetic reconstructions, upwelling was probably induced by the northeast trade winds that influenced the northwest corner of Baltica, located at a palaeolatitude of around 35° S (Torsvik & Rehnström, Reference Torsvik and Rehnström2001; Cocks & Torsvik, Reference Cocks and Torsvik2005). Intense coastal upwelling might have provided a supply of nutrients to surface waters, resulting in high primary productivity followed by high organic carbon fluxes to the sea-floor, although phosphogenetic processes associated with condensation during successive drowning events cannot be ruled out (A. T. Nielsen, Copenhagen, pers. comm. 2009). The precipitation of apatite within these settings is stimulated by the production of pore-water phosphate generated through the microbial degradation of organic matter (preserved as phosphatized microbial pseudomorphs) and, in our case study, coincides with the sudden abundance of linguliformean brachiopods. The episodic character of phosphate precipitation is possible by maintaining the phosphatized particles at or near the sediment–water interface through repeated cycles of sudden burial and subsequent removal (the so-called ‘Baturin cycles’: Föllmi, Reference Föllmi1996). As a result, the Scanian strata are characterized by current-dominated sedimentary regimes with correspondingly low net sedimentary accumulation rates.

Several key requirements are needed for phosphate precipitation: (1) an excess supply of P₂O₅, commonly related to biogenic sources, (2) a limited or depleted supply of Mg, (3) suitable high pH conditions and (4) a slow rate of deposition of clastic or ferruginous sediments (Marshall-Neill & Ruffell, Reference Marshall-Neil and Ruffell2004). Phosphate clasts and nodules observed in the Gislöv and Forsemölla units often encapsulate glauconitic remains, implying early glauconite formation. As Mg is required for glauconitization, glauconite precipitation necessarily depleted the pore water content of Mg, favouring precipitation of phosphate. The coeval occurrence of phosphate and glauconite debris coexisting in polyphase clasts suggests that they either precipitated simultaneously or during alternating phases with no compaction or diagenetic alteration in between. Glauconitization indicates enrichment of iron in sediment pore waters, subsequently precipitated as pyrite. An early-diagenetic suboxic environment provided the right conditions for phosphogenesis, glauconitization and the coeval presence of authigenic and reworked pyrite. This contrasts with the burrowing and the rich and diverse benthic fauna, reflecting syndepositional well-oxygenated conditions.

7. C- and O-isotope stratigraphy

Stable carbon and oxygen isotope analyses were performed on bulk rock samples from fresh intervals of the Gislöv Formation and the Forsemölla and Exsulans limestone beds at Brantevik, a section well constrained biostratigraphically (Bergström & Ahlberg, Reference Bergström and Ahlberg1981). Rock samples were selected in micritic-to-microsparitic matrix with sparse skeletal content, where neither intraparticular nor shelter cements are visible. Clean matrix was drilled with a dental microdrill and the powdered samples were digested in 100% phosphoric acid at 25 °C. The resulting CO₂ gas was analysed using an ISOPRIME mass-spectrometer for isotopic analysis at the University of Salamanca (for details, see Al-Assam, Taylor & South, Reference Al-Assam, Taylor and South1990). The isotopic composition of calcite is expressed by the conventional δ notation relative to the V-PDB reference (Craig, Reference Craig1957); δ = [(Rs/Rr) − 1] * 1000, where R = ¹⁸O/¹⁶O or ¹³C/¹²C, respectively, in the sample and in the reference. Analytical reproducibility (1σ) is 0.1 ‰ for δ¹⁸O and δ¹³C_carb.

We analysed 12 samples for carbonate δ¹³C and δ¹⁸O, spanning a mixed stratigraphic interval of 1.65 m, in which the hardground and phosphatized tops of each limestone bed were not sampled, in order to avoid influence of organic-rich substrates (Fig. 7). The δ¹³C_carb background values for the section gradually decrease from −1.4 to −7.9 ‰ through the Gislöv–Forsemölla units. After the generalized drowning of the Scanian platform, associated with widespread deposition of kerogenous shales, the recovery of carbonate productivity (Exsulans Limestone Bed) displays a sharp increase of δ¹³C_carb over 15 cm, reaching a positive value of +0.89 ‰.

Figure 7. (a) Stratigraphic log of the Brantevik section with chemostratigraphic data. (b) δ¹³C_carb and δ¹⁸O values. (c) δ¹³C_carb ‰ v. δ¹⁸O ‰cross-plot; r = coefficient of correlation.

The diagenetic processes capable of causing changes in the oxygen- and carbon-isotope composition of marine carbonate components include neomorphism and recrystallization, re-equilibration with fluids of different isotope composition, and precipitation of isotopically different diagenetic carbonate phases (e.g. Glumac & Walker, Reference Glumac and Walker1998). Fluctuation of δ¹⁸O values in the studied limestones ranges from −7.38 to −12.69 ‰. These values fit well with similar analyses compiled from Cambrian tropical low-Mg calcite-walled brachiopods (probably yielding the closer proxy to real sea-water chemistry), which vary from −7 to −12 ‰ (Wadleigh & Veizer, Reference Wadleigh and Veizer1992; Veizer et al. Reference Veizer, Ala, Azmy, Bruckschen, Buhl, Brhun, Carden, Diener, Ebneth, Godderis, Jasper, Korte, Pawellek, Podlaha and Strauss1999). In addition, the studied limestones were not significantly exposed either to meteoric waters or to marine-meteoric waters of the mixing zone, because of the wide variability of δ¹⁸O values, and lack of covariance between δ¹³C and δ¹⁸O values (see coefficient of correlation, r = 0.36), respectively (Brand & Veizer, Reference Brand and Veizer1980; Allen & Matthews, Reference Allen and Matthews1982). Although a relative depletion of δ¹⁸O values, reflecting a diagenetic resetting and post-depositional alteration, cannot be rejected, we consider that the δ¹³C values were negligibly altered.

As a result, the δ¹³C composite curve (formed by segments separated by erosive discontinuities) of the lower–middle Cambrian transition in Scania is strongly asymmetric, attaining minimum values (shift to −7.9 ‰) close to the top of the earliest middle Cambrian Forsemölla Limestone Bed. Higher up, after a demise in carbonate productivity and deposition of kerogenous shales related to drowning, the Exsulans Limestone Bed records a pronounced rise in δ¹³C to enriched values (+0.89 ‰), followed again by a new demise of carbonate productivity. The initial negative shift from low to moderately negative values coincides with the onset of the three deepening-upward parasequences described above, whereas the subsequent positive trend coincides with sedimentary features that suggest flooding and drowning in an open-marine setting.

8. Chemostratigraphic comparison with other palaeocontinental margins

Similar asymmetric fluctuations of δ¹³C across the lower–middle Cambrian transition are known from other basins, although intercontinental correlation within the lower–middle Cambrian transitional interval is hampered by the strongly provincial character of the faunas. In the Siberian Platform, a negative δ¹³C shift to nearly −2 ‰ occurs close to the Botoman–Toyonian boundary, post-dating a turnover of shallow-platform trilobites, and followed by a rapid decline in archaeocyathan diversity (Debrenne, Reference Debrenne1991; Zhuravlev, Reference Zhuravlev1998). Brasier et al. (Reference Brasier, Corfield, Derry, Rozanov and Zhuravlev1994) reported that the subsequent Toyonian–early middle Cambrian interval was characterized by a gradual increase in δ¹³C values. This was also documented by Brasier & Sukhov (Reference Brasier and Sukhov1998) in the middle Botoman–late Amgan Kuonamka Formation, dominated by organic-rich argillaceous carbonates deposited under starved-basin conditions, where δ¹³C values gradually turned from mainly negative values to positive at the top of the Toyonian. This corresponds with a δ¹³C curve recorded from coeval Ordian to early Templetonian carbonates from the Georgina Basin (Donnelly, Shergold & Southgate, Reference Donnelly, Shergold and Southgate1988; Shergold, Reference Shergold1995), and a similar gradual increase in δ¹³C values punctuated by positive shifts across the lower–middle Cambrian transition in the western Mediterranean region (West Gondwana; Wotte et al. Reference Wotte, Álvaro, Shields, Brown, Brasier and Veizer2007).

Another distinct negative shift in δ¹³C has been documented from the Laurentian lower–middle Cambrian boundary interval in the Great Basin (Montañez et al. Reference Montañez, Osleger, Banner, Mack and Musgrove2000), where the succession is relatively conformable and complete. In Laurentia, the onset of a large negative excursion (≥ 4 ‰ to negative δ¹³C values) is located just beneath an extinction level at the lower–middle Cambrian boundary (Palmer, Reference Palmer1998a, Reference Palmerb). This negative carbon isotope excursion is interpreted as recording: (1) the introduction of ¹³C-depleted anoxic waters onto shallow-water carbonate platforms during the latest early Cambrian, (2) a decrease in organic carbon burial due to major biomass reduction or (3) a major decrease in primary producers and global oceanic productivity (Saltzman et al. Reference Saltzman, Ripperdan, Brasier, Lohmann, Robison, Chang, Peng, Ergaliev and Runnegar2000; Álvaro et al. Reference Álvaro, Bauluz, Subías, Pierre and Vizcaïno2008). The hypothesis that an anoxic water column developed below the mixed surface layer, at least locally, during the latest early Cambrian and before a major transgression, has recently received further support from sulphur isotope data (Hough et al. Reference Hough, Shields, Evins, Strauss and Mackenzie2006). Sea-water δ¹³C values from the upper part of the middle Cambrian in the Great Basin exhibit high-frequency fluctuations (shifts of a magnitude as high as 4 ‰) around a base-line value of −0.5 ‰, but the curve shows a broadly progressive fall in values during the early middle Cambrian.

No distinct shift in the carbon-isotopic composition was recognized by Guo et al. (Reference Guo, Strauss, Liu, Zhao, Pi, Fu, Zhu and Yang2005) at the lower–middle Cambrian transition in the Guizhou Province of South China, where there is a decreasing trend in δ¹³C values from positive values in the upper lower Cambrian to negative values in the lower part of the middle Cambrian. In contrast, the corresponding interval in the Wangcun section, Hunan, South China, is characterized by a distinct negative shift of about 4 ‰ in δ¹³C values across the Qiangdongian–Wulingian transition (Zhu et al. Reference Zhu, Zhang, Li and Yang2004). These authors have also described two similar excursions across the lower–middle Cambrian transitional interval in the Jiagou and Bagongshan sections of the Anhui Province, North China, although biostratigraphic data allowing precise correlation between these two palaeogeographically separated blocks are not yet available. Zhu et al. (Reference Zhu, Zhang, Li and Yang2004) proposed a direct chronostratigraphic correlation of the Qiangdongian–Wulingian transition of South China with the Laurentian Olenellus–Plagiura transition, based on this negative excursion and the FAD of Ptychagnostus atavus in these successions. Thus, the Laurentian δ¹³C negative shift at the Olenellus–Plagiura transition and the Qiangdongian–Wulingian negative δ¹³C peak in South China (Zhu et al. Reference Zhu, Zhang, Li and Yang2004) correspond well with the Redlichiid–Olenellid Extinction Carbon isotope Excursion (ROECE) at the boundary between Series 2 and 3 of the Global standard Cambrian (Zhu, Babcock & Peng, Reference Zhu, Babcock and Peng2006). The ROECE excursion can be tentatively identified also in Scania, albeit partly affected by the gap between the Gislöv Formation and the Forsemölla Limestone Bed.

9. Subtropical versus temperate waters

Throughout late Neoproterozoic and Cambrian times, Baltica lay at southerly latitudes (∼ 30–60° S: Cocks & Torsvik, Reference Cocks and Torsvik2005) and was geographically inverted: present-day southern Baltica faced the equator and NW Baltica the NW Gondwanan margin. A key palaeomagnetic datum from late Cambrian Alum Shales sampled in the Forsemölla–Andrarum area places Scania at around 35° S (Torsvik & Rehnström, Reference Torsvik and Rehnström2001). Are these palaeomagnetic data supported by the aforementioned facies associations?

The four episodes of shell-dominated carbonate productivity, represented by the onset of the Gislöv, Forsemölla and Exsulans units, developed on an unrimmed platform bordering NW Baltica. Among other primary factors, they were directly controlled by terrigenous input, carbonate productivity, accommodation space and nutrient supply. Frame-building fabrics are absent, and dominant carbonate facies are grainy (or shelly), and were hydrodynamically controlled by the action of traction currents. Biostratinomic processes were dominated by post-mortem skeletal disarticulation and fragmentation, and carbonate mud was commonly winnowed off-platform. The carbonate mud appears to be homogeneous, sometimes mottled as a consequence of burrowing, but rarely pelleted, except within phosphatized clasts, in which spherical aggregates of micrite (up to 50 μm in diameter) are common (Fig. 6c). Microbial activity was pervasive, as indicated by the widespread development of microboring and phosphatized microbes, but did not develop centimetre-scale stromatolites or thrombolites.

The carbonate sea-floor supported a prolific epibenthic biota, which produced abundant skeletal debris, and a significant endobenthic biota, as suggested by the presence of macroburrowing. The greatest production of skeletal grains was associated with the proliferation of multi-element organisms (such as chancelloriids, echinoderms, sponges, trilobites and brachiopods) that rapidly disintegrated after death into their constituent parts, showing abundant signs of mechanical breakage, abrasion and considerable rounding. The benthic communities were dominated by fixed, mainly filter-feeding nonreefal chancelloriid–echinoderm–sponge meadows (the CES community sensu Álvaro & Vennin, Reference Álvaro and Vennin1997). The coarsest fraction of the skeletons was transported short distances (e.g. Exsulans Limestone Bed), while the smallest grains were redistributed by currents across the platform resulting in episodic low-angle sheets or shoals of loose hydrodynamic skeletal sediment that covered the open sea-floor, intensely affected by current and storm washing and winnowing (e.g. Gislöv Formation and Forsemölla Limestone Bed).

The hydrodynamic regime is reflected by the episodic presence of intra-bed erosive surfaces, overlain by grading structures, and not encrusted by microbial veneers or mats. The absence of these microbial crusts is a key difference between the Gislöv and Forsemölla units and another coeval and mixed, condensed succession, namely, the Brèche à Micmacca in the Atlas Mountains, Morocco (Álvaro & Clausen, Reference Álvaro and Clausen2006, Reference Álvaro, Clausen, Pratt and Holmden2008; Álvaro et al. Reference Álvaro, Aretz, Boulvain, Munnecke, Vachard, Vennin, Álvaro, Aretz, Boulvain, Munnecke, Vachard and Vennin2007). The limestone interbeds of the latter also display the episodic development of chancelloriid–echinoderm–sponge meadows, affected by intensive bottom current abrasion and winnowing (responsible for the onset of intra-bed erosive diastems), and an abundance of volcanogenic, phosphate and glauconite clasts. However, in Morocco, the aforementioned diastems were encrusted with microbial mats. The main hydrodynamic factors that controlled biomat development were: low sediment accumulation rates, early lithification (grainstone facies is common), and substrate stability (the microtopography of the shelly/grainy sea-floor provided growth space, where the cemented shell surface was permanently suitable for microbial nucleation). Both the limestones of the Brèche à Micmacca and those of Scania are composite event-concentration, low-relief shoal complexes composed of parautochthonous bioclastic assemblages. These units can be described as ‘hiatal accumulations’ (sensu Kidwell & Bosence, Reference Kidwell, Bosence, Allison and Briggs1991) because of the subdivision of the coquinas by minor discontinuities and the composite, multiple-event nature of the shell concentrations.

Under what conditions do microbial crusts form during depositional hiatuses? Based on field and laboratory studies, under normal oxygenated conditions and in the absence of grazers, brief hiatuses can favour development of microbial laminated crusts, whereas prolonged episodes or longer hiatuses can be characterized by microboring infestation (e.g. Reid et al. Reference Reid, Dupraz, Visscher, Sumner, Krumbein, Paterson and Zavarzin2003). As a result, there is a limit to the time span related to these hiatal processes, as longer hiatal episodes can result in development of eukaryotic algal communities, which do not form laminated structures. Therefore, the absence of microbial mats in the lower–middle Cambrian boundary interval of Scania cannot be directly considered as an argument to support temperate waters.

Nutrient supply is another key factor to understand proliferation of microendoliths (Peterhänsel & Pratt, Reference Peterhänsel and Pratt2001), poorly developed in the Brèche à Micmacca. It can result from terrestrial influx of organic matter, marine transgression, or upwelling. Eutrophic conditions are characterized by high and oscillating levels of nutrient supply, and high net productivity, which contrast with the low and stable nutrient supply and lower net productivity of oligotrophic conditions. As a result, levels of bioerosion are high in eutrophic areas (characterized, for instance, by microendolithic infestation), associated with low or irregular patterns of carbonate productivity, and relatively low in oligotrophic environments (Wood, Reference Wood1993; Brasier, Reference Brasier, Bosence and Allison1995). Glauconite formation and phosphogenesis may also indicate the presence of increased levels of K and P, two nutritive elements. Therefore, all evidence indicates coastal upwelling on the NW corner of Baltica as a direct cause of enhanced nutrient mobilization, analogous to modern subtropical unrimmed platforms (Vogel et al. Reference Vogel, Gektidis, Golubic, Kienne and Radtke2000).

Another palaeogeographic area, where the lower–middle Cambrian transition is recognized as an erosive unconformity covered by limestones and limestone/black shale alternations, is the Georgina Basin, Australia. In the Hay River–Mount Whelan region, Donnelly, Shergold & Southgate (Reference Donnelly, Shergold and Southgate1988) reported the series boundary at the contact of the Red Heart Dolostone and Lower Hay River formations. The lower part of the latter formation consists of phosphatic mudstone and wackestone that alternate with centimetre-thick, phosphatic hardgrounds. The palaeoenvironment subsequently recorded the alternation of winnowed (well-oxygenated) and calm (anoxic) substrates, mimicking the stratigraphy of the lower part of the Alum Shales. However, the Australian sea-floor neither recorded permanent reworking nor pervasive episodic erosion, allowing for the preservation of intact phosphatic hardgrounds, which are absent in Scania.