Abstract: Rögla is the northernmost locality yielding Mesozoic plant fossils in Scania, southern Sweden, and is one of the northernmost Rhaetian assemblages in Europe. The assemblage consists of over 500 specimens collected 50–60 years ago, of which 139 yielded identifiable plant remains referable to 15 plant species; another 19 specimens are tentatively assigned to four species because of their fragmentary preservation. The flora includes sphenophytes, ferns, cycads, bennettitaleans, seed ferns of uncertain alliance, conifers and some leaf remains that are tentatively assigned to ginkgophytes based on their epidermal anatomy. The species-level composition of the assemblage is consistent with a Rhaetian age and is similar to well-known floras from nearby Höganäs and Bjuv, except for the absence of cycads belonging to Nilssonia, which are very common in most other Scanian floras. The fossil assemblage is interpreted to derive from multi-storey vegetation occupying moist habitats on a coastal plain. Strong affinities are evident with the coeval floras of Jameson Land, Greenland, reinforcing the concept of a distinctive North Atlantic floristic sub-province at the close of the Triassic.
T he Rhaetian–Bajocian (Late Triassic – Middle Jurassic) floras from the former coal-mining districts of Scania in southern Sweden are well known throughout the world owing to their fine preservation (many yielding cuticles) and their crucial role in the establishment of many key fossil genera and species during the ‘Golden Age of Palaeobotany’ (Cleal et al. 2005). The floras are very rich, and individual assemblages are diverse; they contain elements of all the major plant groups extant at that time (algae, mosses, sphenophytes, lycophytes, ferns, seed ferns, cycads, bennettitaleans, ginkgoaleans and conifers).
During a recent investigation of bennettitalean remains from these floras (Pott and McLoughlin 2009), one substantial collection of plant fossils was found in the Swedish Museum of Natural History (NRM), Stockholm, that had not been studied previously. The assemblage derives from strata close to the Triassic–Jurassic boundary; hence, it is important for understanding the composition of middle-latitude Laurasian floras immediately prior to the major extinction event at the close of the Triassic. Here, we thoroughly document the Rögla flora to complete the description of fossil plant assemblages from Scania that began almost 200 years ago (Nilsson 1820) and to verify past palaeoecological interpretations (Pott and McLoughlin 2009).
Material, methods and geological setting
The plant fossils derive from the Rögla area, the northernmost Mesozoic plant fossil locality in Scania. A total of 529 rock slabs have been collected from three sites in the open coal pit and from four sites in the collieries south of the property of Rögla (Text-fig. 1). Of these, 158 (= 30%) bear identifiable plant remains. The specimens were collected between 1945 and 1956 by E. Mohrén, E. Bölau and B. Lundblad. Eighty-five specimens apparently derive from the opencast pit at Rögla, whereas 73 specimens were collected from adjacent collieries. The shapes of several rock slabs indicate that they were parts of a drill core that was extracted during exploration of the Rögla coalfield (Text-fig. 1). Some rock slabs yield more than one fossil, so the total number of plant imprints examined is 167. The material from all sites except one of the open coal pit sites (named Site 3 in the NRM fossil database) is dated as Rhaetian (Triassic); Site 3 is listed as Liassic (Early Jurassic), according to the NRM fossil database. All specimens are stored in the palaeobotanical collection of the NRM, under accession numbers S038469, S077606–S078103, S078107–S078122 and S079533–S079547 (bore hole).
The Höganäs Formation in western Scania is subdivided into three units: the Vallåkra, Bjuv and Helsingborg members, in ascending stratigraphic order (Text-fig. 2). The lower Rhaetian Vallåkra Member consists of grey and variegated clay and greenish sandstone lenses deposited under transitional conditions between the arid continental deposits of the Kågeröd Formation and the humid deltaic settings represented by the overlying parts of the Höganäs Formation (Norling et al. 1993). The upper Rhaetian Bjuv Member contains carbonaceous and kaolinitic shale, quartzose sandstone and coal. The prominent coal seams at the top (‘seam A’) and base (‘seam B’) of this unit were mined for 400 years until production ceased in 1961 (Norling et al. 1993). This unit has yielded the majority of plant fossils collected from Scania. The overlying Hettangian Helsingborg Member consists of shale and siltstone interbedded with sheet-like and lenticular mature sandstone. The unit has yielded modest assemblages of plant fossils. Herringbone cross-stratification, flaser bedding and diverse ichnofossils attest to tidally influenced shallow marine setting particularly for the uppermost part of this unit (Ahlberg 1994). The Helsingborg Member is overlain by Sinemurian–Aalenian strata of the marine Rya Formation (Text-fig. 2), which lacks plant macrofossils (Ahlberg et al. 2003).
The Scanian plant fossils are mostly compressions with well-preserved cuticles. Cuticles were prepared according to procedures outlined by Kerp (1990) and Kerp and Krings (1999). Fossiliferous slabs were dissolved in hydrofluoric acid (HF) to remove the sediment. Cuticles were macerated using Schulze’s reagent (35% HNO3 with a few crystals of KClO3) and subsequently treated with a 5–10% potassium hydroxide (KOH) solution. Macerated cuticles were washed in distilled water, gently dehydrated in pure glycerine and finally mounted with glycerine-jelly on microscope slides. Hand specimens were photographed with a Nikon D80 digital camera using cross-polarized light to enhance contrast. Cuticles were analysed with an Olympus BX-51 microscope and photographed with an Olympus DP-71 digital camera. Synonymy lists include only references to studies of Scanian fossils and those incorporating major changes in nomenclature.
Institutional abbreviations. All specimens and slides (see figure captions) are housed in the palaeobotanical collection of the Swedish Museum of Natural History (NRM), Stockholm.
Description. One specimen (S077748) represents a poorly preserved, fragmentary, compressed stem fragment (Text-fig. 3A), 28 mm wide and 40 mm long. The axis bears regular longitudinal striations (5 mm wide).
1845 Calamites lehmannianus Göppert, p. 143 (no illustration).
1869 Schizoneura hoerensis (Hisinger); Schimper, p. 283 (no illustration).
1878a Schizoneura hoerensis; Nathorst, p. 9, pl. 1, figs 1–4.
non 1878a Schizoneura hoerensis; Nathorst, p. 40, pl. 1, fig. 5.
1878b Schizoneura hoerensis; Nathorst, p. 24, pl. 10, figs 6–8.
1908 Neocalamites hoerensis (Schimper); Halle, p. 6; pl. 1, figs 1–4; pl. 2, figs 1–11.
1922 Neocalamites hoerensis (Schimper) Halle; Johansson, p. 7; pl. 6, figs 1–2.
1931 Neocalamites hoerensis (Schimper) Halle; Harris, p. 22; text-fig. 4.
1950 Neocalamites hoerensis (Schimper) Halle; Lundblad, p. 14; no illustration.
1961 Neocalamites hoerensis (Schimper) Halle; Harris, p. 30; text-fig. 8.
1971 Neocalamites lehmannianus (Goeppert) Weber; Stanislavsky, p. 29 (cum syn.), pl. 1, figs 1–2; pl. 2, figs 1–3; pl. 3, figs 1–3; pl. 4, figs 1–2; pl. 5, figs 1–3; pl. 6, fig. 1; pl. 7, figs 1–3; pl. 8, figs 1–3; pl. 9, figs 1–2; text-figs 5–6.
Description. The fossils are preserved as compressions or pith casts. All are incomplete; the largest shoot portion is c. 95 mm long (Text-fig. 3C). The shoots are cylindrical, regularly subdivided into nodes and internodes (Text-fig. 3C, E) and represented by three size categories: 5–6 mm wide (Text-fig. 3E), 13–15 mm wide and 40–60 mm wide in the middle of the internodes (Text-fig. 3C–D). All shoots are slightly wider at the nodes. The narrower forms probably represent lateral branches of the wider (primary and secondary) axes. Internodes are typically c. 25–30 mm long in the smaller shoots, but reach 63–90 mm in primary shoots. Nodal diaphragms are evident in two specimens (Text-fig. 3D). The outer surface of the internodes is smooth or bears broad longitudinal striae (Text-fig. 3D); the inner surface bears prominent, densely arranged, longitudinal striae (separating massive vascular bundles; Text-fig. 3C, E).
Remarks. First recognized as Schizoneura hoerensis by Schimper (1869) and Nathorst (1878a, b), this species was reassigned by Halle (1908) as the type species of Neocalamites. Weber (1968) outlined the taxonomic history and revised N. hoerensis. He convincingly determined that the species was a junior synonym of Calamites lehmannianus (Göppert 1845). The latter name was based on specimens from the Rhaetian–Early Jurassic of Franconia, Germany, that now are considered to belong to Neocalamites (Schenk 1867, 1888; Weber 1968). However, this placement has been disregarded by most subsequent authors (see synonymy list), probably due to the publication of Weber’s (1968) paper in the German language and in a minor German journal.
Neocalamites comprises many species of Carnian to Jurassic age (Pott et al. 2008; Harris 1961); the genus apparently disappeared in the Late Jurassic (Krassilov et al. 2007). Species similar to N. lehmannianus in morphology and size are N. merianii from Ladinian–Carnian deposits and N. carrerei from Rhaetian–Jurassic deposits of eastern Asia. Weber (1968) indicated that these three species can be distinguished only by their stratigraphic and geographical occurrence, which are unsatisfactory distinctions for morphotaxa. Hence, this group deserves taxonomic re-evaluation. At least N. merianii and N.lehmannianus are common in their respective deposits (cf. Pott et al. 2008).
Description. The single specimen consists of a 12 mm wide, 47 mm long, poorly preserved, weakly striate, compressed stem fragment (Text-fig. 3B).
Remarks. Because striations are ill-defined, and the specimen lacks evidence of nodes, specific allocation is impossible. Based on its size, it may represent an ill-preserved example of Neocalamites lehmannianus.
Material. Two specimens (part and counterpart: S077828, S0778679).
Description. The single pinna (39 mm long and 15 mm wide basally) bears eight pairs of sterile pinnules (Text-fig. 3F) lacking cuticle. Pinnules are nearly opposite and roughly ovate to lanceolate but asymmetrical (slightly curved towards the pinna apex) with slightly acute apices. Individual pinnules are 3–4 mm wide basally and 6–7 mm long. The proximal margin is slightly contracted and the distal margin slightly expanded. Veins arise from the basipetal area of the pinnule at 70 degrees and proceed to the margin with regular bifurcations (Text-fig. 3F). The central vein is only slightly more prominent than the secondary veins.
Remarks. Identification of this species is tentative as the Rögla specimens are small and Todites williamsonii has variable pinnule shapes (Harris 1961). The fragmentary specimens from Rögla match the morphology of those from the Jurassic of Yorkshire and the claystone facies at Höör reported by Antevs (1919). Harris (1961) stated that distinction between the Middle Jurassic T. williamsonii and the Rhaetian T. goeppertianus is generally not possible and so considered it better to treat these forms as conspecific, an interpretation that is followed herein. Weber (1968) placed similar material in T. williamsonii var. goeppertianus. Todites roessertii is also similar but differs from T. williamsonii in its venation pattern (Gothan 1914; Weber 1968).
1831 Pecopteris polyploides; Lindley and Hutton, p. 167, pl. 60, figs 1–2.
1867 Gutbiera angustiloba Presl ex parte; Schenk, p. 64, pl. 18, figs 6, 8, 10.
1867 Andriana baruthina Braun var. remota; Schenk, p. 87, pl. 24, fig. 1.
1867 Laccopteris goeperti ex parte; Schenk, p. 94, pl. 34, figs 2–3.
1867 Laccopteris muensteri; Schenk, p. 97, pl. 24, figs 6–10; pl. 25, figs 1–2.
1919 Gutbiera angustiloba Presl ex parte; Antevs, p. 16; pl. 1, figs 8–8a.
?1919 Laccopteris sp.; Antevs, p. 16, pl. 1, figs 1–3.
1931 Laccopteris brauni (Presl) Raciborski; Harris, p. 73, text-fig. 25A.
1936 Phlebopteris muensteri (Schenk); Hirmer and Hörhammer, p. 17 (cum syn.), pl. 3, figs 1–7; pl. 4, figs 1–6; pl. 5, figs 1–6; text-figs B, 5-2A, 5-2B.
1950 Phlebopteris sp.; Lundblad, p. 25, pl. 3, figs 8–9, text-fig. 25A.
1993 Phlebopteris muensteri (Schenk) Hirmer and Hoerhammer; Van Konijnenburg-van Cittert, p. 240 (cum syn).
Material. Four specimens (S077643, S077793, S077916, S077825).
Description. Phlebopteris muensteri occurs on four rock slabs in the Rögla collection. Three specimens belong to one sterile frond; the other is part of a fertile frond (Text-fig. 3G). The petiole of the sterile frond is 133 mm long bearing eight pinnae in one plane at its distal end (Text-fig. 3I). Only the proximal parts of the pinnae are preserved, the largest being 84 mm long. Pinnules are narrow but broadly attached to the rachis, 8–10 mm long, 2.5–3 mm wide, regularly and densely spaced but not in contact. Pinnules are entire-margined with rounded apices. A central vein gives off several densely spaced secondary veins that pass straight, and with sparse bifurcations, to the pinnule margin (Text-figs 3H, 4). Sporangia are arranged in densely spaced sori attached to the underside of pinnules (Text-fig. 3G). Cuticles are not preserved.
Remarks. The Rögla specimens closely match some specimens referred to Gutbiera angustiloba from the claystone facies at Höör (Antevs 1919) that were later assigned to Phlebopteris muensteri by Hirmer and Hörhammer (1936). We regard fragmentary specimens reported as Phlebopteris sp. from Billesholm by Lundblad (1950) as conspecific. Phlebopteris muensteri is similar to P. angustiloba but lacks the distinct mattress-like venation pattern of the latter.
Antevs (1919) reported that Phlebopteris angustiloba including the specimens now assigned to P. muensteri, occurs at Pålsjö, Sofiero, Billesholm, Munka Tågarp and Rödalsberg amongst the Scanian assemblages; the first three sites are very close to Rögla. Generally, P. muensteri is considered to have a distribution confined to the Rhaetian–Early Jurassic of Greenland and Europe (Weber 1968; Van Konijnenburg-van Cittert 1993).
Description. The Rögla specimens are fragmentary pinnae up to 160 mm long and 36 mm wide. Pinna lobes vary from strongly arched and pointed to bluntly triangular (Text-fig. 5A–D). Fertile specimens (Text-fig. 5C) bear densely arranged sori with distinct sporangia on the lamina lobes. Each lobe has a prominent midvein that gives off a complex reticulum of subsidiary veins.
Remarks. The variation in pinnule form is consistent with the morphological continuum recognized by Nathorst (1876, 1906) in abundant material of this species from Bjuv. Dictyophyllum exile is one of the most common ferns in assemblages from Höganäs, Hyllinge, Bjuv, Billesholm and Höör in Scania. However, another relatively common species (i.e. D. nilssonii), reported from Pålsjö and Sofiero (Nathorst 1876; Chow 1924), has pinnae of similar shape. The latter is distinguished by its decurrent pinna base, which links adjacent pinnae; pinnae are free up to the base in D. exile. In addition, D. exile has pinnae of constant width (3–4 cm), except for the base and the apex, whereas the pinnae of D. nilssonii increase gradually in width and then decrease again and are usually much wider (up to 10 cm) in the central part than in D. exile. Pinnae in D. nilssonii are described as being more subulate in some cases. However, the latter is only restricted to a variety of D. nilssonii named ‘genuinum’ (Nathorst 1906). Fragmentary specimens lacking the pinna base are, therefore, difficult to assign to either species because the pinna margin and lobation are congruent. However, these species appear to be separated stratigraphically. Whereas D. exile is mainly restricted to Rhaetian deposits, D. nilssonii is more common in Lower Jurassic deposits. The Rögla specimens are considered to be of Rhaetian age and, therefore, likely belong to D. exile. This species appears to have a very variable lamina form. This is attributed to the source of fragments from a single frond because lamina lobe development differs greatly between proximal, medial and distal portions of a pinna (see restorations by Nathorst 1906). Dictyophyllum nathorstii is most similar to D. exile, but distinguished by pinnae that are free up to the base in D. exile and adnate basally in D. nathorstii (Schweitzer et al. 2009).
Dictyophyllum exile is the most abundant fern at Rögla. According to Schweitzer et al. (2009), D. exile was widely distributed across Laurasia in the Rhaetian–Early Jurassic (e.g. southern Germany, Greenland, Scania, Poland, Iran, central Asia and China), but it has not been reported from the Southern Hemisphere. We regard Dictyophyllum acutilobum reported by Möller (1902) from the Rhaetian of Bornholm to be conspecific with the Scanian material; indeed, Möller (1902) regarded the material from Höganäs and Helsingborg as being identical.
1878b Camptopteris spiralis; Nathorst, p. 33, pl. 2, fig. 8; pl. 3, fig. 1; pl. 4, figs 1–6; pl. 8, fig. 1.
1906 Camptopteris spiralis Nathorst; Nathorst, p. 14, pl. 6, figs 25–31; pl. 7, figs 12–14; text-fig. 4.
1922 Camptopteris spiralis Nathorst; Johansson, p. 11, pl. 5, fig. 57.
1950 Camptopteris spiralis Nathorst; Lundblad, p. 29 (no illustration).
2009 Camptopteris spiralis Nathorst; Schweitzer et al., p. 50 (cum syn.), pl. 10, figs 4–7; text-fig. 14.
Material. Three specimens (S038469, S077775, S077957).
Description. The largest of three sterile pinnae fragments in the Rögla assemblage is 69 mm long and 2–3.5 mm wide (Text-fig. 5E, I). These linear pinnae have serrate margins with a prominent vein passing from the midrib to each lamina tooth. Subsidiary veins form a complex reticulum but are indistinct.
Discussion. This species was described fully by Nathorst (1878a, b, 1886, 1906) from Bjuv, and the present specimens, although fragmentary, match his material very well. Based on the review of the family by Oishi and Yamasita (1936) and direct comparison with the fossils studied by Nathorst (1906), there is no doubt that the Rögla fossils are referable to Camptopteris spiralis. This species occurs in several Scanian assemblages of Rhaetian age (i.e. Billesholm, Bjuv, Bosarp, Hyllinge, Skromberga). According to Schweitzer et al. (2009), this taxon has a disjunct distribution, being confined to the upper Rhaetian of Scania and Iran. The Russian records (Kryshtofovich and Prynada 1932; Kryshtofovich, 1933; Schweitzer et al. 2009) are questionable and probably better identified as basal fragments of Dictyophyllum sp. pinnae (cf. Nathorst 1906).
Phylum PTERIDOSPERMOPHYTA (seed ferns) INCERTAE ORDINIS
Description. The fragmentary specimen (S078079; Text-fig. 5J) from Rögla is represented by one imperfectly preserved leaflet of a segmented frond, 19 mm long and 6 mm wide. Up to six parallel veins are visible. Both the abaxial and adaxial cuticles have costal and intercostal fields, the latter characterized by 3–4 rows of elongate epidermal cells (Text-fig. 6A–B). The difference is less distinct on the thicker adaxial cuticle. The isodiametric to polygonal epidermal cells have straight anticlinal walls. Some epidermal cells bear solid papillae or thickenings. Stomata are generally more numerous on the adaxial side of the leaf (cf. Kustatscher and Van Konijnenburg-van Cittert 2007) but always sunken and commonly small; ring-like structures surround the stomatal pit (Text-fig. 6C–D). Stomata occur in intercostal fields on the abaxial side. Each stoma is ringed by 5–7 subsidiary cells that locally form a thickened structure (Text-fig. 6C–D).
1879 Ptilozamites blasii (Brauns); Nathorst, p. 64, pl. 13, figs 4–8.
1914 Ptilozamites blasii (Brauns) Nathorst; Antevs, p. 15, pl. 1, figs 9–10; pl. 2, fig. 1; pl. 3, fig. 10.
1950 Ptilozamites blasii (Brauns) Nathorst; Lundblad, p. 49, pl. 1, fig. 7.
2007 Ptilozamites blasii (Brauns) Nathorst; Kustatscher and Van Konijnenburg-van Cittert, p. 76 (cum syn.), pl. 2.
Material. S077674, S077800.
Description. Specimen S077800 consists of a basal fragment of a leaf incorporating four pairs of up to 33-mm-long and 19-mm-wide segments (Text-fig. 5K). The segments are almost rectangular, tapering slightly, with up to 27 basally bifurcating parallel veins entering the leaf segments perpendicularly. Specimen S077674 (Text-fig. 5F) consists of only one triangular to rhombic leaf fragment, 92 mm long and 41 mm wide. Parallel veins enter at a point that is interpreted to be the base.
Cuticles are very thick but poorly preserved. Epidermal cells are isodiametric to rectangular and are more elongate over the veins. On the adaxial cuticle (Text-fig. 6E), stomata occur sporadically both on and between the elongate cell bands over the veins. The stomata are sunken and surrounded by up to seven subsidiary cells, forming a strong ring up to 28 μm in diameter around the pit. Vein courses are indistinct on the abaxial cuticle, but stomata are scattered more or less regularly in the intercostal fields. Stomata are very small, deeply sunken, cryptic and surrounded by a ring of subsidiary cells (Text-fig. 6F). The pit is small, commonly circular and usually has a very thick wall along the sides of the aperture. Guard cells are deeply sunken below the surface but are visible in some cases. Abaxial epidermal cells are isodiametric polygonal.
1959 Ginkgoites troedsonii; Lundblad, p. 10, pl. 1, figs 1–12; pl. 2, figs 1–13; text-figs 1–4.
Description. The specimen (Text-fig. 5G) consists of a slender (6 mm wide, 38 mm long) parallel-sided, slightly curved leaf fragment with almost no distinctive characters except some small, irregularly scattered, punctiform structures. However, it yielded excellently preserved thick and leathery cuticles (Text-figs 7A–B). Stomatiferous costal and nonstomatiferous intercostal fields are well defined. Epidermal cells are isodiametric, polygonal or rectangular. Anticlinal cell walls are straight and outer periclinal walls generally smooth, bearing only faint idiocuticular striae (Text-fig. 7B). Stomata are distributed regularly in the costal fields; stomatal pores are randomly orientated. Individual stomatal complexes are separated by one to several normal epidermal cells. However, they are usually interconnected by idiocuticular striae. The stomatal complex is oval to circular and 40–55 μm in diameter (Text-fig. 7D). The guard cells are sunken and possess prominent circum-poral thickenings. Apertures are typically slit-like and shorter than the length of the pit mouth. The guard cells are surrounded by 5–7 subsidiary cells, which are regular in shape and size and are more heavily cutinized than normal epidermal cells. A robust and solid papilla extends from each subsidiary cell and overarches the pit mouth (Text-figs 7C–F).
Remarks. Cuticle features reveal that the specimen best matches Ginkgoites troedsonii as described and illustrated by Lundblad (1959). Macromorphologically, the assignment is difficult, because only part of the leaf is preserved. The specimen from Rögla may, therefore, represent a part of a G. troedsonii leaf segment. The small punctiform structures resemble G.troedsonii resin bodies (Lundblad 1959). The specimen has cuticular features similar to G. marginatus, but resin bodies have never been described in that species from Scania (Lundblad 1959).
According to Lundblad (1959), Ginkgoites troedsonii is restricted to Billesholm, whilst G. marginatus is widely distributed amongst the Scanian localities (Höör, Sofiero, Dompäng, Helsingborg, Pålsjö). However, fossils from these localities derive from Hettangian strata including the type material (Lundblad 1959), whereas the Rögla specimen is considered to be Rhaetian in age. The record of G. marginatus from the Rhaetian of Skromberga/Stabbarp should be considered cautiously. Ginkgoites troedsonii and G. marginatus are similar in external shape and variation (Lundblad 1959) and are only well distinguished by the absence of stomata in large portions of the upper epidermis in G. marginatus (Lundblad 1959). We consider this to be a rather weak character as shown by Lundblad (1959) herself: she ignored this character when assigning the material from the Rhaetian of Stabbarp and Skromberga to G. marginatus. Johansson (1922, p. 44) noted that there ‘are no larger differences between upper and lower cuticle’. The same applies to some of the Greenland specimens (Harris 1935, p. 15), which was also ignored by Lundblad (1959). However, if this is taken into account, then our specimen matches G. troedsonii and not G. marginatus. Significantly, the Skromberga/Stabbarp specimens would also better fit in G. troedsonii than in G.marginatus as probably would the material from the Rhaetian of Jameson Land and Cape Stewart, Greenland (Hartz 1896; Harris 1935). Thus, a regional taxonomic re-evaluation is required.
Genus SPHENOBAIERA Florin, 1936
Type species. Sphenobaiera spectabilis (Nathorst) Florin, 1936.
Material. Six specimens (S077764, S077946, S077966, S077967, S077978, S078083).
Description. These parallel-sided leaf fragments reach 52 mm long and 2.1–2.4 mm wide (Text-fig. 9A–B); veins are not evident. The leaves were robust on the basis of their very thick cuticles (Text-fig. 7G). Epidermal cells are isodiametric to polygonal with straight anticlinal walls. Stomata are distributed regularly in rows interpreted as the intercostal fields; stomatal pores are randomly orientated. The stomatal complex is oval to circular and 35–45 μm in diameter. The guard cells are deeply sunken and surrounded (almost superimposed) by 5–6 subsidiary cells, which are more heavily cutinized than normal epidermal cells (Text-fig. 7H). A distinct and solid papilla extends from each subsidiary cell and overarches the pit mouth (Text-fig. 7H).
Remarks. Based on epidermal anatomy (morphology and distribution of stomata, lack of papillae), these fossil remains are best regarded as fragments of Sphenobaiera leaf segments. They most likely belong to S. spectabilis, originally described by Nathorst (1906) as Baiera spectabilis (cf. Harris 1935; Florin 1936a, b; Lundblad 1959). To date, Sphenobaiera spectabilis is known only from Stabbarp in Scania (Nathorst 1906; Johansson 1922; Florin 1936b) and from Jameson Land, Greenland (Harris 1926, 1935). Alternatively, the leaf remains may be assignable to Pseudotorellia (cf. Lundblad 1957). However, this is somewhat uncertain because the leaf remains are fragmentary, not tapering to either the apex or base, and vein courses are indistinct. Czekanowskia species are also similar; however, leaflets of Czekanowskia from Scania are always thinner, with more prominent vein courses and strongly elongate epidermal cells over the veins (Nathorst 1906; Florin 1936a).
Remarks. Desmiophyllum is a morphogenus for isolated ribbon-shaped Mesozoic leaves of unknown affinity. Farr and Zijlstra (2010) treated it as of questionable ginkgoalean affinity, whilst Lundblad (1959) placed it within the descriptions of cycadalean species. Their possible affinity with Cordaites was discussed and dismissed by Florin (1936b).
1959 Desmiophyllum cyclostomum Lundblad, p. 50, pl. 10, figs 11–15; text-fig. 17.
Description. This slender, parallel-sided leaf fragment is 116 mm long, 29 mm wide, with fine parallel veins (Text-fig. 5H). Only the abaxial cuticle could be prepared (Text-fig. 6G). Epidermal cells are isodiametric to variably polygonal. Vein courses are not recognizable in the preserved cuticles, and stomata are randomly scattered on the epidermis. Stomata are typically incompletely dicyclic, with two polar and up to four lateral subsidiary cells. Subsidiary cells form the stomatal pit, the margin of which is conspicuously raised as a continuous circular ridge (Text-fig. 6H), although it is locally weaker over the polar subsidiary cells. Guard cells are slightly sunken.
Remarks. Assignment to Desmiophyllum cyclostomum is based mainly on agreement in epidermal anatomy with the specimens figured by Lundblad (1959). Our specimen is almost twice as wide as the one reported by Lundblad (1959), but the epidermal anatomy matches exactly. The cuticle also resembles Pseudoctenis florinii (Lundblad 1959) but differs in stomatal morphology and arrangement. Lundblad (1959) compared D. cyclostomum with Bjuvia simplex (Florin 1933) from Bjuv, but refrained from assigning the Hyllinge specimen to Bjuvia because of its external shape and the clear assignment of Bjuvia to the Cycadales. We include the specimen from Rögla in D. cyclostomum and tentatively assign it to Cycadales based on its stomatal type (i.e. haplocheilic). The specimen’s external morphology is poorly known, and assignment to Bjuvia is not an option. The similar epidermal anatomy of D. cyclostomum and B. simplex indicates possible cycadalean affinity of the leaves. Apparently, D. cyclostomum was restricted previously to Hyllinge (Lundblad 1959).
1825 Nilssonia? aequalis; Brongniart, p. 219, pl. 12, fig. 6.
1878 Pterophyllum aequale Brongniart; Nathorst, p. 18, pl. 2, fig. 13; p. 48; pl. 6, figs 8–11.
1879 Pterophyllum aequale Brongniart; Nathorst, p. 67, pl. 15, figs 6–11a.
1922 Pterophyllum aequale (Brongniart) Nathorst; Johansson, p. 32, pl. 1, figs 14–16; pl. 8, figs 20–21.
1932b Pterophyllum schenkii Zeiller; Harris, p. 49, pl. 6, figs 1–2; text-figs 22–24.
1950 Pterophyllum compressum; Lundblad, p. 56, pl. 9 figs 9–13; pl. 10, figs 1–6; text-figs 20–22.
2003 Pterophyllum cf. aequale (Brongniart) Nathorst; Schweitzer and Kirchner, p. 60, pl. 12, figs 4–5; text-fig. 20.
2009 Pterophyllum aequale (Brongniart) Nathorst, emend. Pott and McLoughlin, p. 125 (cum syn.), pl. 2, figs 1–12; pl. 3, figs 1–8; text-fig. 4.
Description. The Rögla specimen is fragmentary, up to 14 mm long and 33 mm wide, with at least five partly preserved, laterally inserted leaf segments (Text-fig. 9C). The parallel-sided leaf segments reach 6 mm wide and 31 mm long; they are slightly expanded proximally and are inserted at c. 90 degrees (Text-fig. 9C). Up to eight basally bifurcating veins enter the leaf segments and run straight to the apices. The leaf is hypostomatic and yields well-preserved cuticles (Text-fig. 8A–D); costal and intercostal fields are distinguishable on both sides of the leaf (Text-fig. 8A–B). Stomata are restricted to intercostal fields. In the adaxial cuticle, the cells of intercostal areas are mostly elongate, rectangular to isodiametric; those over the veins are longer and more slender. Anticlinal cell walls are straight, and periclinal cell walls are smooth with a small central cuticular thickening or solid papilla. The costal fields of the abaxial epidermis contain narrowly rectangular cells. Anticlinal cell walls are mostly straight, but a few have weak, irregular undulations. Periclinal walls are smooth, but each bears a small central thickening or solid papilla. Intercostal fields of the abaxial surface contain elongate, rectangular to isodiametric cells with mostly straight anticlinal cell walls. Stomata are regularly distributed, brachyparacytic and 25–30 μm in diameter (Text-fig. 8D). The pits are orientated perpendicularly to the veins (Text-fig. 8C). The diacytic stomata possess two rectangular subsidiary cells, each bearing a small papilla that overhangs the pit mouth (Text-fig. 8D). Guard cells are slightly sunken.
Remarks. Pott and McLoughlin (2009) gave a detailed description of this species based on 106 specimens from Scania. The only example of Pterophyllum aequale from Rögla is fragmentary, but yields well-preserved cuticles allowing confident allocation to this species. Pterophyllum aequale is the most widespread bennettitalean species within Scania (represented at 10 major localities), and it is recognized with confidence only from uppermost Triassic strata (Pott and McLoughlin 2009). This stratigraphic distribution is consistent with other records of the species apart from a single Early Jurassic record in Iran (Schweitzer and Kirchner 2003).
1896 Pterophyllum subaequale Hartz, p. 236, pl. 15, figs 1, 3.
1922 Pterophyllum andraeanum Schimper; Johansson, p. 33, pl. 5, figs 15, 17; pl. 8, figs 22–23.
1932a Pteropyhllum subaequale Hartz; Harris, p. 96, text-fig. 38A–C.
1932b Pteropyhllum subaequale Hartz; Harris, p. 75, pl. 6, figs 8–14; text-figs 39–42.
2009 Pterophyllum subaequale Hartz emend. Pott and McLoughlin, p. 132 (cum syn.), pl. 5, figs 1–7; pl. 6, figs 1–9; text-fig. 4.
Material. S077677, S077680.
Description. The largest available leaf fragment is 119 mm long and 61 mm wide with at least seven pairs of laterally inserted leaf segments (Text-fig. 9E). Ensiform leaf segments reach 7 mm wide and 31 mm long; they are slightly contracted proximally and are inserted at c. 80 degrees (Text-fig. 9D). Up to eight basally bifurcating veins enter the leaf segments and run straight to the apices. Cuticles could not be extracted.
Remarks. Pott and McLoughlin (2009) provided a detailed description of this species based on 25 specimens from predominantly Hettangian strata at Stabbarp, Scania. The Rögla specimens have characters equivalent to those from Stabbarp, but are more fragmentary. Despite their fragmentation and lack of cuticle, Pterophyllum subaequale from Rögla can be clearly distinguished from P. aequale by the finer details of its gross morphology. The epidermal characters distinguishing these species were outlined by Pott and McLoughlin (2009). Although confidently dated records of Pterophyllum subaequale are restricted to Hettangian strata in Scania, Poland, and possibly south-west Russia (Pott and McLoughlin 2009), the species has also been recorded from Rhaetian strata in Greenland (Hartz 1896; Harris 1932a).
Material. Four specimens including two counterparts (S077607, S077653, S078038–S078039).
Description. These isolated leaf segments are best assigned to Pterophyllum. The parallel-sided leaflets reach 41 mm long and 7 mm wide (Text-fig. 9F). Eight to nine parallel veins are present in each leaflet; bifurcations are not evident within the preserved portions of the leaflets.
Remarks. The specimens are too fragmentary and poorly preserved to be identified to species level, but they superficially resemble P. aequale.
Description. The leaves are fragmentary to almost complete (Text-fig. 9G–J) with lengths up to 123 mm long and widths to 24 mm. The lamina is regularly segmented in the middle and distal parts of the leaf but entire-margined in the proximal portion. Leaflets are oppositely to sub-oppositely positioned, densely arranged and free up to the base. Leaflets are attached by their whole base and are more or less falcate with the acroscopic margin being slightly concave and the basiscopic being strongly convex; apices are rounded. Up to 14 basally bifurcating veins enter the leaf segments and run straight to the apices.
The leaves are hypostomatic; costal and intercostal fields are only distinct on the abaxial leaf surface (Text-fig. 8E–F). Cuticles are robust. Epidermal cells in the adaxial cuticle are generally elongate, rectangular with broadly undulate anticlinal walls. Stomata are absent on the adaxial leaf surface. The abaxial epidermis is clearly differentiated into costal and intercostal fields. Costal fields host elongate and roughly rectangular epidermal cells with delicate and broadly undulate anticlinal walls and smooth periclinal walls; stomata are absent. Individual cells of the intercostal fields are polygonal and isodiametric. Anticlinal walls are widely undulate. Stomata are regularly distributed within the intercostal fields, brachyparacytic and 20–25 μm in diameter (Text-fig. 8G–H). The stomatal pores are orientated arbitrarily, and the two rectangular subsidiary cells create, by overarching the pit mouth, a slightly sunken stoma.
Remarks. The Rögla specimens closely match examples from other Scanian assemblages. Apart from Dictyophyllum exile, Anomozamites angustifolius is the most common fossil at Rögla. It is also one of the most common gymnosperm fossils at several other Scanian localities. Assignment to A. angustifolius is based on both macromorphological and epidermal characters (cf. Pott and McLoughlin 2009); the species was recently erected and distinguished based on material from several Scanian localities (except Rögla) that had previously been assigned mostly to Anomozamites minor (see Pott and McLoughlin 2009).
Remarks. The systematic position of Mesozoic conifer-like genera such as Palissya, Stachyotaxus and Cyparissidium has been discussed extensively in recent decades. The architecture of the cone-like structures assigned to Palissya remains particularly controversial (Parris et al. 1995; Schweitzer and Kirchner 1996), and no concensus has been reached on their phylogenetic affinities. Although affiliation with the Coniferales has been commonly invoked, this placement is by no means certain (Florin 1958; Harris 1979; Parris et al. 1995; Arndt 2002).
The difficulties of assigning sterile leafy conifer or conifer-like twigs to clearly demarcated morphogenera were discussed in detail by Harris (1979). He used criteria such as leaf proportions, leaf base and tip shapes, leaf divergence and manner of leaf insertion to differentiate genera. He favoured the assignment of sterile conifer shoots bearing divergent, elongate, dorsiventrally flattened, univeined leaves to Elatocladus in agreement with an earlier proposal by Berry (1924). Rees and Cleal (2004) adopted a similar strategy, and in accordance with their study, we follow Harris’ (1979) diagnosis of Elatocladus herein. However, we note that very similar or conspecific leafy axes have been inferentially linked, although never convincingly found attached, to Palissya cones by several previous workers (Nathorst 1908; Florin 1958; Parris et al. 1995) except for an unpublished record (Hill, 1974; cf. Schweitzer and Kirchner, 1996). Although Elatocladus has generally been used for younger (Jurassic–Cretaceous) shoots and leaves than those described below (latest Triassic), temporal separation is not a strong basis for differentiation of morphotaxa, and we note that Harris (1935) also recognized several species of this genus from Rhaetian deposits of Greenland.
Stachyotaxus includes both cones and shoots and is known only from Rhaetian strata. Shoots are dimorphic with a proximal part covered with small, scale-like leaves and a distal part bearing longer ensiform leaves that are confined (distichously) in one level (Nathorst 1908; Harris 1935). Several authors have recognized two species (S. elegans and S. septentrionalis) but recent investigations by Arndt (2002) suggest that these taxa are conspecific. The cones consist of loose, spirally arranged bract-scale complexes each bearing a single seed inserted within a cup-like structure on the adaxial surface.
Cyparissidium includes very small shoots with spirally arranged, very small, tapering leaves that are appressed to the shoot (Nathorst 1879; Harris 1979). This genus is apparently more common in Middle Jurassic to Cretaceous deposits (Harris 1979). Separation of the many morphologically similar Mesozoic conifer genera would ideally be aided by cuticular data and attached reproductive structures, but these are seldom available.
Material. Six specimens (S077888, S077889, S077892, S078015, S078016, S078069).
Description. The largest of the six sterile shoots is 43 mm long with leaves reaching 14 mm long and 1.5 mm wide (Text-fig. 9M). Leaves are persistent, bifacial, spirally arranged, of moderate density, spreading in all directions at 30–70 degrees, narrowly linear to lanceolate, acuminate, slightly contracted at the base but not petiolate, broadly decurrent, univeined, flat, keeled on their abaxial side and in some cases slightly vaulted on their adaxial side. Cuticles are not available.
Remarks. Equivalent leafy twigs from Scanian assemblages have been assigned to Palissya by some previous authors, and these are difficult to separate from Torreya (Harris 1935, 1979). However, mid-Mesozoic examples of the latter have leaves with short flattened petioles (Harris 1979). No examples of Palissya have been reported from coeval strata of East Greenland, since Harris (1935) considered it wiser to retain all sterile conifer twigs with small needles in Elatocladus. Unfortunately, the Rögla specimens are too badly preserved to yield cuticles and lack reproductive structures so their affinity with Palissya sensu stricto remains uncertain. For synonymy of proper Palissya sphenolepis, see Florin (1958), who also included P. brauniiEndlicher, 1847 in P. sphenolepis.
Leafy axes of this type (assigned to Palissya sphenolepis) were first reported from Bjuv by Nathorst (1908) and later from Stabbarp and Skromberga by Johansson (1922). However, the full distribution of this taxon remains uncertain owing to the different taxonomic approaches to the identification of such remains by past researchers. Further studies of Palissya cones and associated foliage are clearly warranted, because their higher taxonomic affinities remain obscure and even the ovuliferous versus microsporangiate character of the reproductive organs remains in dispute (Parris et al. 1995; Schweitzer and Kirchner 1996).
Genus STACHYOTAXUS Nathorst, 1886
Type species. Stachyotaxus septentrionalis (Agardh) Nathorst, 1886.
Description. Four of the available specimens represent cones or parts of cones, and the remaining 32 are leafy shoots. One additional specimen representing a seed is probably affiliated with a cone of S. septentrionalis. Leafy shoots are up to 83 mm long and consist of a proximal part with small, scale-like leaves helically arranged around the axis and a distal part bearing longer ensiform leaves that are spirally inserted, but twisted basally to be confined (distichously) to one plane (Text-fig. 9K, P–Q). Leaves of the proximal part are 4–5 mm long (Text-fig. 9K), those of the distal part reach 6–9 mm long. Leaves are univeined, keeled and decurrent at the basiscopic margin. Cones, interpreted as ovuliferous structures, reach 44 mm long (Text-fig. 9L) incorporating a central axis bearing loose perpendicularly inserted bracts. The bracts consist of a proximal short (2.5–3 mm long) petiole widening into a triangular, seed-bearing, abaxially keeled, slightly expanded, median portion (shield) that bears a small apical spine. The complete scale is incurved such that the terminal spine points distally or, in some cases, towards the axis. The cone scales of the Rögla specimens appear to have borne two seeds adaxially (Text-fig. 9L, arrow). Seeds are ovate, c. 2.5–3 mm long. Cuticles could not be extracted.
The taxonomic placement of Stachyotaxus septentrionalis has a confused history. The generic assignment has been changed several times, and some authors have recognized two discrete species (mainly S. elegans and S. septentrionalis). However, we agree with recent investigations by Arndt (2002) that suggest these forms are conspecific and fall comfortably within the generic circumscription. Leaves from the Carnian of Lunz, Lower Austria, that were assigned to S. lipoldii by Kräusel (1949) are incompletely studied, and their placement in the genus is questionable. However, if further research confirms their attribution to Stachyotaxus, then the generic range would be extended to the Carnian of central Europe.
1879 Cyparissidium nilssonianus; Nathorst, p. 103, pl. 22, figs 12–19.
non 1879 Cyparissidium nilssonianus; Nathorst, p. 103, pl. 22, figs 24–32.
Material. Four specimens (S077876, S077973, S077983, S078058).
Description. Very small, branched, sterile shoots up to 89 mm long and <2 mm wide. Axes 0.5 mm wide with dense, spirally arranged, appressed, minute, flattened, scale-like leaves, up to 2.5 mm long and 0.5 mm wide (Text-fig. 9N–O). Free part of the leaf contracting gradually from its base to apex. Diverging shoots are as thick as mother axes. Cuticle was not recoverable.
Remarks. No studies of Cyparissidium nilssonianum have been undertaken since Nathorst (1879) first described the species. However, it is a fairly common component of Rhaetian assemblages from Bjuv and Höganäs. The specimens from Rögla are very similar to those from Bjuv and Höganäs. Several other species from Scania have been assigned to Cyparissidium but were later transferred to other genera (e.g. Stachyotaxus and Palissya; Nathorst, 1878a). Although Cyparissidium is more common in Middle Jurassic–Cretaceous strata (Heer 1874; Harris 1979; Florin 1958), the specimens at hand have consistent generic characters in common with material described and illustrated from the Middle Jurassic of Yorkshire (Harris 1979), apart from the leaves of the former being slightly less dense and longer than the latter. The identity of specimens assigned to C. nilssonianum by Schenk (1887) and Zeiller (1905) from the Rhaetian of Iran was questioned by Harris (1979), and those specimens were formally transferred to C. rudlandicum by Schweitzer and Kirchner (1996).
The Rhaetian–Hettangian floras of Scania have been described in detail over the past 130 years. It is, therefore, surprising that a major assemblage of plants from the northernmost fossil deposit of this age in Scania has not previously been examined in detail. Our study completes the survey of major plant fossil assemblages from Scania. Large assemblages are unlikely to be forthcoming from Scania in the immediate future owing to the cessation of coal mining in the 1960s, and increasingly restricted access to the remaining clay pits.
Comparison to other localities in Scania and stratigraphic position
The composition of the Rögla flora is strikingly similar to those from Bjuv, Höganäs, Billesholm and Hyllinge (cf. Nathorst 1878a, b, 1879, 1886; Johansson 1922; Lundblad 1950, 1959), but this is not surprising given their close proximity and apparently similar age (Text-figs 1, 2). Taxa occurring in all floras include Dictyophyllum exile, Camptopteris spiralis, Ptilozamites nilssonii, Pterophyllum aequale, Anomozamites angustifolius and Stachyotaxus elegans (Table 1), whereas Neocalamites lehmannianus and Cyparissidium nilssonianum occur in four to five of the floras. Rögla and Bjuv share a few taxa that have not been reported from Höganäs, Billesholm or Hyllinge (i.e. Ptilozamites blasii, Elatocladus sp.), whilst Höganäs, Billesholm and Rögla share Todites williamsonianii, which has not been found at Bjuv and Hyllinge (Table 1). Few taxa are shared only between Rögla and Billesholm or Hyllinge (i.e. Phlebopteris muensteri, Ginkgoites troedsonii, Desmiophyllum cyclosstomum) that are not reported from Bjuv and Höganäs. The fossil assemblage from the Rhaetian strata at Stabbarp (Johansson 1922) is also very similar (Table 1). The strong similarity between all these floras favours a common age. Based on the close association with coal seams in the Rögla pits, and the presence of biostratigraphically significant taxa (Ptilozamites sp. cf. P. nilssoni, P. blasii, Stachyotaxus septentrionalis), in the absence of Thaumatopteris schenkii, an upper Rhaetian age is proposed for the Rögla assemblage and it is interpreted to be derived from the Bjuv Member of the Höganäs Formation. Floras from Pålsjö and surroundings (Nathorst 1878a; Chow 1924), derived from the Helsingborg Member, have a rather distinct composition and are slightly younger (Hettangian; Pott and McLoughlin 2009).
Table 1. Distribution of species identified in this study amongst key Rhaetian floras from Scania.
X indicates presence; (X) indicates tentative identification; indicates possible presence, checked by observation of the collections, but not published records.
Todites sp. cf. T. williamsonianus
Ptilozamites sp. cf. P. nilssonii
Sphenobaiera sp. (cf. S. spectabilis)
Two taxa found in the Rögla flora potentially cast doubt on the upper Rhaetian age (Table 1): Pterophyllum subaequale was considered by Pott and McLoughlin (2009) to be a predominantly Hettangian species in Scania, but two specimens from Rögla confidently identified on gross morphological criteria suggest that the range of this taxon may have initiated in the upper Rhaetian. These specimens derive from ‘Site 3’ (of uncertain specific location) at Rögla, and it is alternatively possible that these specimens come from basal Hettangian strata immediately overlying the uppermost coal seam (‘seam A’) in the Rögla pit. Indeed, the age of ‘Site 3’ is listed (by an anonymous authority) as Hettangian without justification in the NRM fossil database. This age assignment could be correct because all species from that site (except Stachyotaxus septentrionalis) have been reported from both upper Rhaetian and Hettangian strata. The second species, Phlebopteris muensteri, is reported from the Hettangian Höör, Pålsjö and Sofiero floras within Scania. However, it is also convincingly described from the Rhaetian of Billesholm and Hyllinge. The P. muensteri specimens from Rögla derive from ‘Gruvan 1’ and ‘Site 4’ where they occur with typical Rhaetian taxa; hence, we favour a Rhaetian age for these assemblages.
One striking difference between the Bjuv/Höganäs and the Rögla/Billesholm/Hyllinge floras is the great scarcity of cycad foliage and especially the absence of Nilssonia at Rögla. Only two specimens attributable to cycads (viz. Desmiophyllum cyclostomum) have been found at Rögla. This impoverishment in cycads may reflect strong local ecological controls on the floras.
Comparisons between the Scanian Rhaetian–Hettangian floras and others across the Northern Hemisphere were provided by Pott and McLoughlin (2009). In broad terms, the Scanian floras (including Rögla) show strong affinities at species level with the Jameson Land flora of Greenland, but only modest affinities with other roughly coeval European floras (e.g. from Poland, Germany, Austria, Italy) and relatively tenuous links with floras further afield (Iran, Afghanistan, China, Japan, Vietnam). Southern Hemisphere floras are notably different, even at generic level. Dobruskina (1994) included the Scanian assemblages within the ‘Greenland-Japan belt’ of the ‘European-Sinian palaeofloral area’ based on the presence of Dipteridaceae, Czekanowskiales and Ginkgoales. She distinguished this belt from the more southerly ‘Iran-Vietnam belt’, which includes assemblages from Germany and the UK where Dictyopteridiaceae are also common, but Czekanowskiales and Ginkgoales are practically absent. Whilst the Rögla flora falls within Dobruskina’s (1994) Greenland-Japan belt, the Greenland/Scania floras share such a large range of taxa that they appear to constitute a distinct North Atlantic sub-province (Text-fig. 10) during the Rhaetian. Further, the distinct floristic signature of this region persisted into the Early Jurassic.
Palaeogeographic reconstructions suggest that the Scanian late Rhaetian fossil assemblages accumulated in lowland settings adjacent to the eastern shore of a large but shallow marine embayment along the southern margin of the Fennoscandian Shield (Norling et al. 1993; Stampfli and Borel 2002; Blakey 2005). The plant-bearing beds are generally interpreted to have been deposited in meandering river floodplains within deltaic or coastal plain settings (Pott and McLoughlin 2009 and references therein).
Abundant osmundaceous, matoniaceous and dipteridaceous ferns and sphenophytes in the flora point to consistently humid habitats as these groups are strongly dependent on moist conditions for reproduction. However, the seed ferns, cycads, ginkgoaleans and particularly the bennettitaleans have cuticular/epidermal characteristics (e.g. thick cuticles, sunken stomates and overarching papillae) that are generally regarded as xeromorphic features arising as adaptations in plants facing some sort of physiological drought. However, such stresses can be caused by various environmental conditions, e.g. true aridity, high salinity and habitats with sustained winds. Environments with low-nutrient osmotic soils or reduced pH values, such as peat-mires, despite their high water tables, generate drought-like physiological conditions because of negative osmotic potentials. The extensive coals in the late Rhaetian of Scania (Text-fig. 2) attest to consistently moist environments with thick, long-lived, peat-forming mires (Moore 1989).
Although not definitive, the presence of several putative conifer species with small, scale-like leaves in tight helices also suggests conditions of physiological drought, although their apparent lack of robust cuticle may imply otherwise. A traditional interpretation might invoke a hinterland, dry-soil source for these plants, and their preservation in the floodplain deposits only after substantial transport. On the other hand, as the affiliation of these plants with conifers is not conclusive, their ecological requirements remain cryptic. There is no conclusive evidence to indicate their preference for drier habits as generally assumed for conifers and the abundance of Stachyotaxus septentrionalis in the assemblage suggests that these remains have not been transported far. Indeed, even a confirmed coniferalean affinity for these plants does not preclude their growth in very wet habitats (cf. for instance Taxodium distichum (Linnæus) Richard, 1810 and Retrophyllum minor (Carrière) Page, 1989).
Nineteen species were recorded in the Rögla flora belonging to seven plant orders. Whilst gymnosperms dominate the flora in species diversity (12 species), one fern, one bennettitalean and one conifer species (Dictyophyllum exile, Anomozamites angustifolius and Stachyotaxus septentrionalis) codominate the assemblage in terms of fossil abundance (Text-fig. 11). The composition of the flora broadly suggests a multi-storey parent vegetation with numerical dominance by ferns in the understorey and conifers in the upper storey and with a moderate diversity (nine species) of midstorey non-conifer gymnosperms (e.g. bennettitaleans).
Acknowledgements. The authors wish to thank the Swedish Research Council (Vetenskapsrådet) and the Friends of the Swedish Museum of Natural History and Malaises Foundation for providing financial support for this study. Comments from Han van Konijnenburg-van Cittert and an anonymous reviewer to improve the manuscript are greatly acknowledged.