End-Wenlock terminal Mulde carbon isotope excursion in Gotland, Sweden: Integration of stratigraphy and taphonomy for correlations across restricted facies and specialized faunas

Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology journal homepage: www.elsevier.com/locate/palaeo

End-Wenlock terminal Mulde carbon isotope excursion in Gotland, Sweden: Integration of stratigraphy and taphonomy for correlations across restricted facies and specialized faunas Emilia Jarochowska a,⁎, Oskar Bremer b, Daniel Heidlas a, Stephanie Pröpster a, Thijs R.A. Vandenbroucke c,d, Axel Munnecke a a

GeoZentrum Nordbayern, Fachgruppe Paläoumwelt, University of Erlangen-Nuremberg, Loewenichstr. 28, 91054 Erlangen, Germany Subdepartment of Evolution and Development, Department of Organismal Biology, Uppsala University, Norbyvägen 18A, SE-752 36 Uppsala, Sweden c Department of Geology, Ghent University, Krijgslaan 281-S8, B-9000 Ghent, Belgium d Evo-Eco-Paléo UMR 8198, Université de Lille, Avenue Paul Langevin, Bâtiment SN5, 59655 Villeneuve d'Ascq Cedex, France b

a r t i c l e

i n f o

Article history: Received 24 April 2016 Received in revised form 12 June 2016 Accepted 19 June 2016 Available online 22 June 2016 Keywords: Vertebrate Fish Conodont Palynomorph Taphonomy Deformation

a b s t r a c t The decline of the upper peak of the Homerian Mulde carbon isotope excursion (CIE) is used in low- to midpaleolatitudes as a marker for the Wenlock/Ludlow boundary, which is otherwise difficult to constrain in carbonate successions. In the Midland Platform (England) the CIE ends just below the boundary or ranges through it, whereas in Baltic sections it has been placed substantially below the inferred base of the Ludlow Stage. Difficulties in correlating the Baltic sections are caused by widespread development of lagoons and sabkhas with specialized conodont and vertebrate faunas. We describe here a lagoonal section from Gothemshammar, eastern Gotland (Sweden), spanning the entire upper peak of the Mulde CIE. Based on integrated conodont, δ13Ccarb and sequence stratigraphy, a hardground present at the lowering limb of the CIE is correlated with a sequence boundary present across the Baltic Basin, in the Midland Platform, and the southern shelf of Laurentia. This sequence boundary corresponds to a global eustatic regression and can serve as a correlative horizon in restricted facies with depauperate or specialized fauna. The Wenlock-Ludlow boundary is placed in the transgressive strata overlying this boundary. Species richness and abundance of thelodonts, anaspids, and osteostracans at Gothemshammar represent one of the first diversity peaks of vertebrates in the Silurian. Associated conodonts are characteristic for late Wenlock marginal-marine environments and distinguished by large, robust elements. We quantitatively assess the conodont assemblages to evaluate to which degree the overrepresentation of large elements in these facies is produced by taphonomic processes. The taphonomic alteration differs between species and facies, but is lowest for the shallow-water specialist Ctenognathodus murchisoni. Regardless, the use of this species as an index taxon is discouraged based on its strong facies affinity. Instead, the integrated approach proposed here indicates that the Wenlock/Ludlow boundary is situated lower in the Baltic sections than previously identified. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Upper Wenlock stratigraphy and the position of the Wenlock/ Ludlow boundary in carbonate successions remain elusive for paleoecological reasons: during the late Homerian the tropical epicontinental carbonate platforms of Baltica and Laurentia developed vast expanses of shallow restricted flats, populated by specialized faunas disparate from their open-marine counterparts. These epeiric basins saw an unprecedented diversity of early vertebrates (Märss, 1989; Turner, 1999; Sansom, 2008) and specialized conodont assemblages (Viira, 1982a; ⁎ Corresponding author. E-mail address: [email protected] (E. Jarochowska).

http://dx.doi.org/10.1016/j.palaeo.2016.06.031 0031-0182/© 2016 Elsevier B.V. All rights reserved.

Barrick, 1997; Viira and Einasto, 2003; von Bitter et al., 2007; Jarochowska et al., 2016). Late Ordovician and early Silurian vertebrate faunas from the Baltic Basin and other areas generally show low diversities and are mainly represented by thelodonts (e.g. Märss, 1989; Turner, 1999). However, towards the Wenlock/Ludlow boundary diverse faunas containing many thelodont, osteostracan, and anaspid taxa have been found in several areas of the Baltic Basin (Fredholm, 1990; Blom et al., 2002; Märss et al., 2007, 2014). This diversity peak is more or less contemporaneous in both the Eastern Baltic region and Gotland (Sweden), but equally diverse faunas have been reported from younger sediments of the East Baltic (Blom et al., 2002; Märss et al., 2007, 2014). Thelodonts seem to have spread across a wide range of environments (Märss et al., 2007), but facies associations of

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

anaspids and osteostracans indicate that these groups preferred lagoon and shoal areas (Märss and Einasto, 1978; Sansom, 2008; Märss et al., 2014). Sansom (2008) suggested that our understanding of early osteostracan evolution is largely obscured by the taphonomic bias resulting from their seemingly specific environmental preferences during the Silurian. Deposits of Llandovery and early Wenlock age in Gotland have only produced a handful of thelodont scales, and only a few fragments of anaspids and osteostracans have been reported from the upper Slite Group (Fredholm, 1990), slightly before the diversity peak mentioned before. Limited understanding of the stratigraphic distribution, dispersal patterns and ecology of these groups is partly due to their occurrence in restricted facies which yield little faunal indicators of age and environmental conditions. The replacement of open-marine, high-diversity early Homerian conodont faunas with rare, specialized forms has been attributed to an ‘extinction event’ (e.g. Jeppsson and Calner, 2003; Cramer et al., 2012) and linked to perturbations in the carbon cycle, responsible for the Homerian ‘Mulde’ double-peaked carbon isotope excursion (e.g. Samtleben et al., 2000; Calner et al., 2004; Munnecke et al., 2010; Cramer et al., 2012). However, the conodont turnover of the time closely follows the transition from fully marine, platform-margin and slope facies to restricted, microbially-dominated, low-diversity intraplatform flats, to the point that individual highest occurrences fall at lithological and parasequence boundaries (Calner and Jeppsson, 2003; Viira and Einasto, 2003; Calner, 2005; Jarochowska et al., 2016). Such shallow facies dominated by the microbial carbonate factory are rich in hiatuses, often cryptic, and the paucity of fauna renders correlations almost impossible. These difficulties are reflected in the diversity of conodont zonations proposed for the middle Silurian (summarized in Corradini et al., 2015). What is more, specialist shallow-water conodonts are also poorly characterized morphologically and new taxa are commonly described based on small collections from isolated localities (e.g. Viira, 1983, 1994; Viira and Einasto, 2003; Jeppsson et al., 2006), or left in open nomenclature (e.g. Strömberg, 1997; Jarochowska et al., 2016). Peculiar morphologies of these specialist forms may result from sampling bias favoring normal-marine, easy to process lithologies (Lehnert et al., 2005), from adaptations to particular diet, or from taphonomic factors. In Devonian conodont biofacies, which have long been recognized and used for facies-specific zonations, shallow-water conodonts are typically the largest and most ‘robust’ (e.g. Sandberg and Dreesen, 1984; McGoff, 1991). Robustness is also invoked with respect to Wenlock lagoonal conodont faunas (Strömberg, 1997; Calner and Jeppsson, 2003). It has, however, not been investigated, to what extent

305

is this the result of taphonomic factors, such as entrainment of small, gracile elements by waves (Broadhead et al., 1990; McGoff, 1991) or their fracturing during diagenesis (von Bitter and Purnell, 2005). This study has three aims: (1) integrate all environmental and paleontological information to identify clues which may be used in correlating upper Wenlock lagoonal, intra-platform successions with each other and with their open-marine counterparts, (2) characterize the early vertebrate communities thriving in such environments, and (3) to evaluate whether the high proportion of ‘robust’ conodonts results from taphonomic bias. To achieve this, an upper Wenlock lagoonal section is examined in the Swedish island of Gotland, the Silurian succession by far best documented in terms of carbonate platform development and stratigraphy (Samtleben et al., 2000; Calner and Jeppsson, 2003; Calner et al., 2004; Jeppsson et al., 2006; Cramer et al., 2012). 2. Geological setting The investigated section (φ N 57°36′33.46″; λ E 18°47′46.32″) is located on the NE coast of the island of Gotland, E Sweden (Figs. 1–2). This area is part of a low-latitude carbonate platform, which stretched across the southern margin of the paleocontinent of Baltica during the early Paleozoic. The Silurian succession in Gotland spans the uppermost Llandovery through Ludlow (Calner et al., 2004). The strata dip towards the southeast and range from off-platform, deeper-water facies across the western coast to restricted, intra-platform facies in the east. In this eastern facies belt the middle Silurian deposits are poorly exposed, and, consequently, poorly studied. The section is located ca. 0.5 km WNW of the first of a series of coastal outcrops characterized as Gothemshammar 1–8 (Hede, 1928; Laufeld, 1974a) and corresponds (with an overlap) to strata immediately below those exposed at Gothemshammar 1. The interval has been made available for investigation by recent coastal erosion (Fig. 2). The new exposure is designated Gothemshammar 9, in order to distinguish it from previously described sections and preserve the stratigraphic order of numbering. The exposed interval is situated slightly below the boundary between the Halla and Klinteberg formations, placed by Hede (1928) at a hardground surface at Gothemshammar 1. The position of this boundary is revised further in the present study. The Halla Fm. in eastern Gotland consists, from bottom to top, of the Bara Oolite Mb. and the informally defined ‘Hörsne’ and ‘Gothemshammar’ members (Calner and Jeppsson, 2003). The ‘Gothemshammar mb’ is formed by oncoidal packstones interpreted as back-reef, lagoonal deposits Across the paleo-basinward transect,

C

A

B

Fig. 1. Map of the Gothemshammar 9 locality in Gotland, Sweden. A) Political map with Gotland marked in red. B) Position of the section in NE Gotland. Scale bar equals 20 km. C) Red square marks the position of the studied section at the coast. Scale bar equals 1 km. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

306

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

i.e. in the SW direction in present-day terms, the ‘Gothemshammar mb.’ corresponds to the Djupvik Mb., which forms a steep slope in the topography in the central part of the island and is represented by crossbedded pack- and grainstones (Calner and Jeppsson, 2003). The lower Klinteberg Fm. in western and central Gotland consists of crinoidal pack- to rudstones, bioherms and biostromes assigned to the Hunninge Mb. In eastern Gotland coeval strata are represented by skeletal grain- to rudstones with abundant well-rounded coated grains formed by cyanobacteria and stromatoporoids, overlain by a thin algal biostrome (Calner and Jeppsson, 2003). Based on topographic and biostratigraphic correlation, oncoidal floatstones with corals and stromatoporoids exposed along the coast at Gothemshammar are assigned to the lower part of the Klinteberg Fm. and interpreted as deposited in calm, back-reef or back-bank environment (Hede, 1928; Frykman, 1989). The ‘Gothemshammar mb.’ spans the uppermost part of the Homerian Ozarkodina bohemica longa conodont Zone, the entire Kockelella ortus absidata Zone, and the lowermost part of the Ctenognathodus murchisoni Zone in the regional zonation for Gotland (Jeppsson and Calner, 2003; Jeppsson et al., 2006). The lower part of the Klinteberg Fm. falls within the C. murchisoni Zone, i.e. below the Wenlock/Ludlow boundary (Jeppsson et al., 2006; Märss and Männik, 2013). Samtleben et al. (2000) identified elevated δ13Ccarb values (up to 2.28‰) in the Klinteberg Fm. exposed along the coast at Gothemshammar. These elevated values correspond to the upper peak of the double-peaked Mulde carbon isotope excursion (CIE), recorded worldwide across low- to mid-latitudes and allowing global correlation (Calner et al., 2004; Munnecke et al., 2010; Cramer et al., 2012; Jarochowska and Munnecke, 2016).

method of Jeppsson et al. (1999). The residue was sieved and the fraction between 63 μm and 1 mm was examined. Conodonts are housed in the collection of Institute of Paleontology, University of Erlangen-Nuremberg (repository numbers are the same as sample numbers). Thelodonts, anaspids and osteostracans are housed in the Palaeontological Collections, Museum of Evolution, Uppsala University, Sweden (PMU). Sample weights and counts are given in Supplementary Table 2. Taphonomic alteration of conodont assemblages is understood here as selective loss of elements with respect to their number in the moment of the animal's death. It was assessed using a modified approach based on Van den Boogaard and Kuhry (1979) and von Bitter and Purnell (2005). The number of P elements and S and M elements were counted and compared against expected frequencies calculated based on proportions of elements in ontogenetically stable 15-element apparatuses. Criteria for element identification followed the general reconstruction of ozarkodinid apparatus (Purnell et al., 2000) and reconstructions for individual species as follows: Wurmiella excavata – by Jeppsson (1969) and von Bitter and Purnell (2005), Oulodus excavatus – Jeppsson (1972), Ozarkodina confluens – Jeppsson (1969), and Ctenognathodus murchisoni – Strömberg (1997). Only for W. excavata direct evidence for an apparatus composed of 15 elements is available in the form of natural assemblages (von Bitter and Purnell, 2005). For the remaining species 15-element apparatuses are inferred based on topological and morphological homologies (Purnell and Donoghue, 1997). Grouping of elements into only two categories serves two purposes. The first is to reduce the bias resulting from difficulties in identifying homological elements. These difficulties differ substantially between species, e.g., P1 and P2 elements are morphologically similar in Wurmiella (Jones et al., 2009) or Oulodus. The second objective is to express quantitatively the overall modification of the elemental composition in a way allowing comparisons between different species, in which the mechanical robustness is distributed differently across types of elements. Regardless of these differences, P elements are robust, reflecting their food-processing function, whereas S and M elements – to different degrees depending on the species – are more gracile, which is associated with their food-grasping role (e.g. Purnell, 1993). Therefore we propose to use the proportion of P against S and M elements as an index summarizing the overall bias in the recovered assemblage. Statistical significance of this bias was evaluated using Pearson's χ2 goodness of fit test that the observed counts were derived from a population, in which P elements represented 26.67% and S + M elements – 73.33% of all elements, respectively. Expected proportions are based on a 15element apparatus composition. Based on the assumption that P elements are more likely to be preserved than S and M, the bias was also expressed as loss of S + M elements with respect to P elements, calculated as 1 – (number of expected S + M elements / (11 × (number of observed P elements) / 4)).

3. Methods

3.3. Palynology

3.1. Carbonate carbon isotopes Micritic areas in bulk samples were powdered using a hand drill. The powders were reacted with 100% phosphoric acid at 70 °C using a Gasbench II connected to a ThermoFinnigan Five Plus mass spectrometer. All values are reported in per mille relative to V-PDB by assigning δ13C and δ18O values of +1.95‰ and − 2.20‰ to NBS19 and −46.6‰ and 26.7‰ to LSVEC, respectively. Reproducibility and accuracy were monitored by replicate analysis of laboratory standards calibrated to NBS19 and LSVEC and were better than ±0.05‰ (1σ) and ±0.09‰ (1σ), respectively. Results are given in Supplementary Table 1 and in Fig. 2.

The samples (TVDB-13-410 and TVDB-14-207 through TVDB-14210, Fig. 3) were broken into pieces of c. 0.5 cm and decarbonated using 34% HCl. They were then agitated with c. 150 ml 48% HF over c. 24 h, and subsequently subject to a second 17% HCl treatment, over 32 h and at 60 °C. Samples were neutralized and residues sieved at 51 μm. Individual chitinozoans (together with some other representative palynomorphs) were handpicked from the organic residue larger than 51 μm using a stereomicroscope (at ×30–50). Other palynomorphs than chitinozoans or palynomorphs from the smaller fraction residue (below 51 μm) were not systematically studied. Detailed sample positions are given in Supplementary Table 2.

3.2. Vertebrate microfossils

3.4. Microfacies

Rock samples (EJ-14-407 through EJ-14-411, Fig. 3) cut into 5–8 cm long pieces were dissolved in 7% acetic acid buffered according to the

Thin sections were produced at GeoZentrum Nordbayern, University of Erlangen-Nuremberg. Rock samples were impregnated with

Fig. 2. Coastal exposure at Gothemshammar 9. A, B, C… are lithological units marked in Fig. 3 and described in the text.

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

307

Fig. 3. Lithological log of the Gothemshammar 9 section with fossil occurrences and summary of taphonomic alteration in recovered conodont assemblages (see text). EJ-… and TVDB… mark sample positions. HST – highstand systems tract, TST – transgressive systems tract, SB – sequence boundary.

synthetic resin, cut, ground with a series of corundum silicate powders, and glued on glass plates. Overview photos were produced using a slide scanner (Nikon Coolscan V ED and HP ScanJet G4050). Details were photographed using a Zeiss Axiophot petrographic microscope. 4. Results 4.1. Facies and depositional environment Based on 28 thin sections, six facies types (FT) have been distinguished (detailed characteristics in Supplementary Table 3), and the section divided into six lithological units (A–F) representing different depositional environments (Figs. 2–3). Unit A consists of limestone-marl alternations dominated by highabundance, low-diversity assemblages with monospecific pavements of the bivalve Palaeoneilo sp. (FT1, Fig. 4A) and mollusk- and oncoidrich beds (FT2–3, Fig. 4B). Some shells are filled with vadose silt. Normal grading, predominantly convex-up orientation of shells, and umbrella structures (Fig. 4A) suggest transport of shells by storms, but their extremely low diversity indicates that shell material was derived from restricted environment. This unit is interpreted as deposited in a lagoon below the fair-weather wave base, regularly agitated by storms which concentrated shell material and possibly transported it from shallower parts of the lagoon. Unit B is developed as limestone-marl alternations with sparse mollusks and oncoids (FT 6). Based on the presence of calcareous udoteacean algae (Dimorphosiphon? spp.) it is interpreted as deposited within the euphotic zone, but below the fair-weather wave base. Sediment homogeneity and ostracodes with smooth carapaces indicate soft-bottom, thoroughly bioturbated environment, and low-diversity fauna dominated by mollusks. Unit C is formed by limestone-marl alternations with bimodal sediment composition: bivalves and oncoid accumulations with peloidal grainstone matrix (FT2–4; Fig. 4C–H). Unlike in unit A, components are chaotically oriented with only subordinate coquinas formed by convex-up oriented bivalves. Towards the top of this unit oncoids become less fragmented and abraded, and gradually become thick and irregular, which is interpreted here as their transition from allochthonous

to autochthonous. Some beds have polished surfaces with truncated fossils and iron oxide coating, indicating longer periods of erosion and cementation on the sea floor. This is proposed to have been deposited near the fair-weather wave base, in water energy increasing upwards and higher on average than in underlying units, as suggested by the supply of diverse components bearing no traces of hydrodynamic sorting or long-distance transport. Unit D is similar to unit A (FT1 and 3), but interpreted to have been deposited in overall higher energy as the fine matrix is winnowed. Unit E is formed by peloidal wacke- to rudstones (FT4) grading upwards into peloidal grainstones (FT5), in which peloids are commonly pyritized. The orientation of the bioclasts changes from chaotic to horizontally aligned towards the top of the beds, and the degree of fragmentation increases. The top 10 cm are cross-bedded. This unit is interpreted as shallowing upwards from deposition below the fairweather wave base to above it, in a wave-agitated, but nevertheless restricted environment. Restriction is indicated by the high abundance of microbial carbonates (oncoids) and peloids. The origin of peloids has not been established, but intensive bioerosion is visible throughout the section, especially in mollusk shells (Supplementary Table 3) and points to micritized bioclasts being their likely source. The top of this unit is a hardground surface mineralized with iron oxide (Figs. 4I–J and 5A–B, E). The surface is penetrated by Trypaniteslike borings up to 1.5 cm deep (Fig. 5A–B). Cross bedding and bioclasts are truncated and show evidence of dissolution (Fig. 5E). The surface has a flat topography at the scale of the outcrop and has rare encrustations of small (b1 cm) bryozoans (Fig. 5C). Underneath, in the topmost part of unit E, bioclasts are rimmed with isopachous fibrous and bladed cements (Fig. 5G) indicating marine phreatic environment. Many bioclasts seem to have gravitational cements, as if formed in the vadose zone (Fig. 5F). However, they could be also interpreted as growing in the empty space sheltered underneath the bioclasts, whereas the upper surface of bioclasts is typically covered with micritic sediment preventing cement growth. Small-scale undulations (up to 2 cm amplitude) follow Thalassinoides-like horizontal burrows. The borings are either cemented with spar or filled with red-stained silt. Unit F is an oncoid floatstone forming amalgamated beds (Fig. 4I–J). The oncoids are porostromate, branching, with irregular outlines, and

308

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Fig. 4. Microfacies of the Gothemshammar 9 section. Positions are given relative to the hardground surface between units E and F. A–B) Bivalve dominated rud- to floatstone with peloidpackstone matrix, unit A; A) −180 cm; B) broken cortoids indicating that fracturing was syndepositional and predated filling of the bivalve molds; −210 cm. C–D) Mud- to wackestone with large gastropods and oncoids, unit C; C) −116 cm; d) broken and repeatedly encrusted bivalve shell; outer layers are formed by Rothpletzella spp.; −110 cm. E–G) Bivalve-pack- to floatstone with vertically stacked shells in the upper part, top of unit C, −92 cm; F) Erosional contact between packstone (left) and matrix of the bivalve-peloid-floatstone (right). Clearly visible are the truncated shells, partly overgrown by Rothpletzella spp. (arrow). G) Erosional contact between packstone (top) and matrix of the bivalve-peloid-floatstone (bottom). H) Peloidal wacke- to rudstone. Large components enriched in the lower part, the rest is characterized by bioturbation, unit C, −128 cm. Bivalve rudstone with peloidal grainstone matrix. Vadose silt fills both the stacked shells and interparticle pores, unit A, exact position. I–J) Hardground surface between units E and F. Peloidal grainstone in the lower part. In the middle: mineralized surface with abraded bioclasts and vertical borings originating from there, unit E. J) Thin section through the hardground with mineralized surface penetrated by sub-vertical borings.

have diameters up to 3 cm. Associated fauna includes abundant rhynchonellid brachiopods, ostracods and stromatoporoids. Some bioclasts in the lowermost part have mineralized iron oxide coatings (Fig. 5D). This unit is interpreted as deposited below the wave base, in calm, thoroughly bioturbated environment with low depositional rates. The abundance of cyanobacteria and low diversity indicate a restricted, probably lagoonal setting. The succession of encrustations and borings penetrating previously mineralized surface indicates a two-step development of the hardground surface, in which dissolution and mineralization took place first, under strongly oxidizing conditions in the absence of hard-substrate fauna. An abundance of pyrite under the hardground suggests that the mineralization postdates deposition and lithification. This, together with the presence of gravitational cements, suggests

subaerial exposure. The final step of the hardground formation is interpreted here as having taken place during initial flooding of this surface under low depositional rate, which would have provided suitable conditions for hard-bottom faunas to settle. Sediment starvation in this initial phase is supported by the presence of mineralized bioclasts (Fig. 5D). 4.2. Carbon isotopes Carbon isotope values rise across units A through B from 1.8‰ to 2.6‰ and remain in the range between 2.0‰ and 2.7‰ in unit C and the lower part of unit D below the bentonite bed (Fig. 3). In the upper part of unit D the δ13C values decline from 2.4‰ to 1.5‰ and continue declining to 1.0‰ in units E and F.

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

309

Fig. 5. Two hardground surfaces at Gothemshammar 9 (A–G) and Gothemshammar 1 (H–I). A) Top view with iron oxide mineralization and pits marking borings. B) Side view with Trypanites-like borings (arrows). C) Bryozoan growin on the hardground surface. D) Brachiopod shell coated with iron oxide in the basal Klinteberg Fm. E) Evidence for dissolution of bioclasts at the top surface of the hardground. Arrow points to a truncated echinoderm. f) Bladed cements with gravitational orientation (arrow) below the hardground surface. G) Bladed cements on a brachiopod shell. H) Surface view of the hardground surface with truncated oncoids and Trypanites-like borings, with patches of overlying lithology. I) Polished slab through the hardground surface, the arrow points to the upper edge of a truncated stromatoporoid, bored and overlain with gray mudstone.

4.3. Vertebrates All the investigated samples are rich in scales and other vertebrate fragments. Although fragmentary, the remains appear to not have undergone much taphonomic abrasion. The samples have been systematically picked using a stereo microscope, but not all scales and fragments were selected and picked out. Thelodont scales represent the most abundant vertebrate remains. The other two groups, osteostracans and anaspids, generally occur in equal amounts, but the osteostracans show the highest overall taxonomic diversity. There are some differences between individual samples regarding vertebrate content (Fig. 3), but at this stage it is not possible to discern whether these are primary signals or merely the result of sampling. For instance, the taxonomic composition does not change substantially throughout all examined samples. Furthermore, there are a number of remains with osteostracan and anaspid affinities that are too worn or too fragmentary to be identified. The majority of the vertebrates found in this study have already been described from Gothemshammar or figured in open nomenclature (Martinsson, 1966; Gross, 1968; Fredholm, 1990; Blom et al., 2002; Märss et al., 2014), and our descriptions focus on the new scale types and sculpture patterns belonging to taxa not reported from Gothemshammar before.

4.3.1. Thelodonts The thelodont scales in the samples can be divided into two distinct, roughly equally occurring scale groups (Fig. 6) referable to the scale sets presented in Märss et al. (2007) for Paralogania martinssoni (Gross, 1967) sensu stricto and Thelodus laevis Pander (1856). Märss (2003) referred a wider variety of head scales and transitional scales to Par. martinssoni, and similar ones (Fig. 6A, I) are also found in these samples, albeit less frequently. A few scales (Fig. 7A) found in each sample have

ridged crowns hosting lateral spines and enlarged bases with the pulp cavity reduced to a minute pore at the posterior end (Fig. 7B), the latter being typical for mature thelodont scales (Märss et al., 2007). These have not been reported from Gothemshammar previously, but their morphology suggests affinity to Par. martinssoni and they are reminiscent of some scales in Märss (2003). Scales referable to Oeselia mosaica Märss (2005) are also present, and we may have underestimated their number as their typical ultrastructure is only identifiable in the SEM. Par. martinssoni and T. laevis have been reported previously from Gothemshammar in the Halla Formation (Martinsson, 1966; Gross, 1968; Fredholm, 1990) and the presence of O. mosaica has been inferred (Märss, 2005). It is worth noting that the T. laevis scales found in these samples sometimes host spines radiating from the lateral sides of the crown (Fig. 6KK, MM). This variant was reported by Fredholm (1992) from the similarly aged Möllbos 1 locality on Gotland, and putatively called ‘Thelodus schmidti - type B’ in an unpublished manuscript, but no formal description was ever made. When scales identified as T. laevis are viewed in the SEM, fine striations and pits typical for this taxon (Märss, 2006) are visible on several scales (e.g., Fig. 6JJ). Some of the scales however, display an ultrasculpture of irregular grooves radiating over an otherwise smooth crown surface (Fig. 7C–D) or a faint, irregular polygonal pattern (Fig. 7E). This is diagnostic for O. mosaica from the contemporary Viita and Vesiku beds of Saaremaa, Estonia (Märss, 2005), although the ultrasculpture of the scales presented here appears to be slightly finer and not as prominent (compare Fig. 7D to Märss, 2005: plate 1: 5). Scales of O. mosaica have an overall morphology similar to the younger Thelodus parvidens (Agassiz, 1839) according to Märss (2005), but they also overlap with the T. laevis morphotypes. Other typical morphological features for scales with O. mosaica-like ultrasculpture found in this study appear to be small sizes with smooth, rounded crowns that in some scales host slightly crenulated anterior edges. However, similar

310

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Fig. 6. Morphological scale sets of Paralogania martinssoni (A–O) and Thelodus laevis (AA–MM) following Märss et al. (2007). A–D) head scales (PMU 29836–29,839); E–N) Trunk and transitional scales (PMU 29840–29849); O) ?fin scale; AA–BB) anterior head scales (PMU 29851–29852); CC–DD) cephalo-pectoral scales (PMU 29853–29854); EE–MM) postpectoral and precaudal scales (PMU 29855–29863). Scales in crown view except A in lateral view. A and EE from sample EJ-14-411, remaining scales from EJ-14-410. Scale bar equals 200 μm.

fine scales lacking the polygonal ultrasculpture also occur in the samples. According to Märss (2005), the ultrasculpture may be possible to discern under a regular light-microscope using low-angle light, but it is otherwise difficult to assess how common this taxon is in the samples. With the aid of the SEM, O. mosaica has been confirmed in samples EJ14-408, EJ-14-410, and EJ-14-411. 4.3.2. Anaspids Fredholm (1990) figured fragments of anaspids from the Halla Fm., but kept them under open nomenclature. These specimens were to some extent revised by Blom et al. (2002), which will be further commented on below. The samples treated here contain scales from several anaspid species, all of which have been reported from Gotland before (Fredholm, 1990; Blom et al., 2002) and belong to the order Birkeniida Berg, 1937. There are also a number of other fragments, such as spines (Fig. 8A), but they are generally hard to identify to the species or even genus level when found disarticulated (Blom et al., 2002).

Pterygolepis nitida Kiær (1911) (Fig. 8B) and Rytidolepis quenstedtii Pander (1856) (Fig. 8C) are present in all the investigated samples. According to Blom et al. (2002), who also reported both these taxa from Gothemshammar, the former was figured by Fredholm (1990: Fig. 7E). The specimens in Fredholm (1990: Fig. 7F–G) are here identified as R. quenstedtii. Many of the scales displaying a sculpture of ridges typical for Pt. nitida have evenly spaced openings that are evident using the SEM. These openings form canals that extend all the way through the scale, and open on the visceral side as well. The canals appear not to be part of the general scale-vascularization which is otherwise typical for more porous scales, such as those among rhyncholepidids (Blom et al., 2002). Pterygolepis nitida usually has dense non-vascularized scales, so the presence of these openings is unusual (compare Blom et al., 2002). They could represent special scales allowing for lateralline system, which is otherwise poorly understood in anaspids. Schidiosteus mustelensis Pander (1856) (Fig. 8D) was found in three samples. This taxon was not reported by Fredholm (1990), but was identified in samples from Gothemshammar by Blom et al. (2002)

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

311

Fig. 7. Special Paralogania martinssoni scales in crown (A; PMU 29864) and basal view (B; PMU 29865), and Oeselia mosaica scales (C–E; PMU 29866–29868) in crown view. A, C and E from EJ-14-410, B from EJ-14-411, and D from EJ-14-408. Scale bar equals 100 μm.

who also assigned a specimen figured by Gross (1968: Fig. 14A) to Sch. mustelensis. The taxa Rhyncholepis butriangula Blom et al. (2002) (Fig. 8E) and the closely related Rhyncholepis parvula Kiær (1911) (Fig. 8F) show some differences in their occurrences. Rhyncholepis parvula scales are found in all but the oldest sample (EJ-14-409), while Rh. butriangula scales have been definitely identified in samples EJ-14-407, EJ-14-408, and EJ-14-411. Besides reporting both Rh. butriangula and Rh. parvula from Gothemshammar, Blom et al. (2002) also revised scales figured as Birkeniida sp. C by Fredholm (1990: Fig. 7A–B) to Rh. butriangula. Blom et al. (2002) did the same with one of the anaspid scales presented by Gross (1968: Fig. 14B), referred to as Birkeniida sp. C by Fredholm (1990). The remaining anaspid specimens figured by Gross (1968: Fig. 14C–D) are harder to identify. Gross' (1968: Fig. 14C) specimen, which Fredholm (1990) referred to Birkeniida sp. C as well, is here suggested to be more similar to Pt. nitida. However, we remain cautious as the overall morphology of this fragment suggests a specialized position on the body. One specimen illustrated by Gross (1968: fig. 14D) could belong to Rh. butriangula, but its fragmentary state hampers a definite assignment. 4.3.3. Osteostracans As mentioned before, the osteostracans are the taxonomically most diverse vertebrate group in this section, the majority of them belonging to the order Tremataspidiformes (Berg, 1937). In turn, the most commonly occurring osteostracan has been identified as Thyestes verrucosus (Eichwald, 1854), Thyestidae (Rohon, 1892) (Fig. 8G), numerous fragments of which have been found in all the samples. Fredholm (1990) putatively assigned some of her specimens to Procephalaspis oeselensis?

(Robertson, 1939), but the specimens from Gothemshammar (Fredholm, 1990: Fig. 8A–C) are here revised as Th. verrucosus based on the features of their sculpture. Furthermore, Fredholm (1990) referred fragments figured as Thyestes sp. by Gross (1968: Fig. 11B–D) to P. oeselensis? mainly based on morphological features, but also on histological similarities (Gross, 1968: Fig. 13A–E) following the suggestion by Janvier (1985). However, we do not follow this revision, based the more recent work by Märss et al. (2014). The specimens in Gross (1968: Figs. 11A–E, ?12F–G, 13A–E) are instead referred to Th. verrucosus. Three genera and four species in the family Tremataspididae (Woodward, 1891) have been found in the samples, although only one of them has definitively been identified in EJ-14-410 (see Fig. 3 for occurrences). Saaremaaspis mickwitzi Rohon (1892) (Fig. 8H) was reported as Oeselaspis Robertson, 1935 sp. from Gothemshammar by Fredholm (1990: Fig. 8H–I). She also referred fragments from Gothemshammar, originally presented as Oeselaspis sp. indet. in Gross (1968: Figs. 9A & C, 10A), to this genus, but we here revise all these except one to S. mickwitzi. The genus Oeselaspis is present however, represented by Oeselaspis pustulata Patten (1931) (Fig. 8I). This taxon was reported from Gotland by Märss et al. (2014), who referred to a fragment figured by Gross (1968: Fig. 9A) as belonging to this species. Two taxa of the genus Tremataspis (Schmidt, 1866) have been identified in the samples, namely Tremataspis schmidti Rohon (1892) (Fig. 8J) and Tremataspis milleri Patten (1931) (Fig. 8K). Only Tr. schmidti has been reported from Gothemshammar before (Märss et al., 2014). It was figured as Tremataspis sp. indet. in Gross (1968: Figs. 6C, 7A & C) and later referred to Tremataspis sp. in Fredholm (1990) who also presented

312

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Fig. 8. Anaspid (A–F) and osteostracan (G–S) fragments from the Gothemshammar 9 section: anaspid spine (A; PMU 29869), Pterygolepis nitida (B; PMU 29870), Rytidolepis quenstedtii (C; PMU 29871), Schidiosteus mustelensis (D; PMU 29872), Rhyncholepis butriangula (E; PMU 29873), Rhyncholepis parvula (F; PMU 29874), Thyestes verrucosus (G; PMU 29875), Saaremaaspis mickwitzi (H; PMU 29876), Oeselaspis pustulata (I; PMU 29877), Tremataspis schmidti (J; PMU 29878), Tremataspis milleri (K; PMU 29879), Meelaidaspis gennadii (L; PMU 29880), Osteostraci gen. et sp. indet. (M–S; PMU 29881–29887). All are in external view, except A which is in side view. Scale bar equals 1 mm.

scales of the same taxon (Fredholm, 1990: Fig. 8K–L). All these were subsequently revised as Tr. schmidti by Märss et al. (2014), which is corroborated here. A few specimens in samples EJ-14-407 and EJ-14-408 were identified as Meelaidaspis gennadii (Märss et al., 2014) (Fig. 8L), family incertae sedis of the order? Cephalaspidiformes Berg, 1937. This osteostracan has not been reported from Gotland before, but is rather common at just one level in Elda Cliff, Kuusnõmme Beds of Rootsiküla Stage on Saaremaa Island, Estonia (Märss 2016, pers. comm.). A number of osteostracan fragments were not possible to refer to any existing taxa (Fig. 8M–S). These probably represent at least one new taxon, but they are here referred to as Osteostraci gen. et sp. indet. Interestingly, some of these fragments (Fig. 8M; PMU 29881) are remarkably similar to those presented in open nomenclature by

Martinsson (1966: Fig. 1A) and Gross (1968: Fig. 6A), and show similarities to the genus Tahulaspis (Märss et al., 2014). Fragments with a more regular porous network (Fig. 8n–o; PMU 29882–29883) that resemble the later stages of development in Tahulaspis Märss et al., 2014 are also present in the samples. These Tahulaspis-like fragments are found in all samples except EJ-14-410. All of the samples contain fragments that share features with the genus Dartmuthia (Patten, 1931) (Fig. 8PQ; PMU 29884–29885), while fragments with parallel, smooth ridges similar to Procephalaspis (Denison, 1951) (Fig. 8R; PMU 29886) only occur in EJ-14-407 and EJ-14-409. However, the latter lack the small ridgelets in between the larger ridges that are present in the East Baltic specimens figured in Märss et al. (2014: figs. 34L, M). Samples EJ-14407, EJ-14-408, and EJ-14-409 contained rare, fragmentary remains of unknown affinity with smooth surfaces punctured by widely separated,

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

large holes (Fig. 8S; PMU 29887). The fragmentary state of the specimens discussed above and the dissimilarities to previously described taxa render a definitive assignment to any of the mentioned genera

313

difficult. Tahulaspis, Dartmuthia, and Procephalaspis have all been reported from other parts of the Baltic Basin, but only in younger strata (Märss et al., 2014). The remains described here may therefore

Fig. 9. Conodonts from the Gothemshammar 9 section. A–K) Ctenognathodus murchisoni (Pander, 1856); A–D) P1 elements; A–B) sample EJ-14-407; C–D) sample EJ-14-410; E) S1 element, sample EJ-14-409; F) S2 element, sample EJ-14-409; G–H) P2 elements, sample EJ-14-407; I) S2 element, sample EJ-14-410; J) M element, sample EJ-14-410; K) S0 element, sample EJ-14407. L–O, T–U and X) Oulodus excavatus (Branson and Mehl, 1933); L) S3/4 element, sample EJ-14-411; M) S0 element, sample EJ-14-407; N) P2 element, sample EJ-14-411; O) P1 element, sample EJ-14-407; T) S1/2 element, sample EJ-14-411; U) M element, sample EJ-14-407; X) P2 element, sample EJ-14-411. P–S and BB–FF) Ozarkodina ex gr. confluens; P, R and S) P1 elements, sample EJ-14-407; Q) S0 element, sample EJ-14-411; BB) M element, sample EJ-14-411, CC–FF) P1 elements; CC, EE) sample EJ-14-410; DD, FF) sample EJ-14-408. V–W) Wurmiella excavata (Branson and Mehl, 1933), sample EJ-14-407; V) P1? element; W) P2P1? element. Y) Panderodus unicostatus (Branson and Mehl, 1933), sample EJ-14-411. Z) Oulodus siluricus (Branson and Mehl, 1933), P2? element, sample EJ-14-411. AA) Ozarkodina wimani (Jeppsson, 1974), P1 element, sample EJ-14-407. Scale bar 500 μm. Arrows indicate deformations discussed in the text.

314

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

represent earlier forms of these genera, but more material is needed before accurate descriptions are possible. 4.3.4. Conodonts The total of 36.25 kg dissolved for this study yielded 4630 identifiable elements, i.e., an average yield between 22 elements × kg− 1 (EJ-14-409) and 347 elements × kg− 1 (EJ-14-411). All samples contained, in decreasing abundance, Oulodus excavatus (Branson and Mehl, 1933) (Fig. 9L–O, T–U, X), Ozarkodina bohemica (Walliser, 1964) (Fig. 10F–R), Oz. ex gr. confluens Branson and Mehl (1933) (Fig. 9P–S, BB–FF), Wurmiella excavata (Branson and Mehl, 1933) (Fig. 9V–W), Ctenognathodus murchisoni (Pander, 1856) (Fig. 9A–K), and Oulodus siluricus (Branson and Mehl, 1933) (Fig. 9Z). Individuals assigned to the Ozarkodina confluens group included Oz. confluens cf. densidentata (Viira, 1983) and Ozarkodina anika (Viira and Einasto, 2003) (Fig. 9R, DD–FF) because transitional forms were present and a rigorous morphometric analysis seems warranted in order to distinguish these taxa. Many elements belonging to the genus Ctenognathodus could not be assigned to species or were tentatively identified as Ctenognathodus cf. sp. P (Viira and Einasto, 2003) based on P1 elements present in these samples, but as Viira and Einasto (2003) noted, their S and M elements cannot be currently distinguished from similar species such as C. sp. S until larger collections are available. The upper part of the section yielded also single elements of Ozarkodina wimani (Jeppsson, 1974) (EJ-14-407, Fig. 9AA) and Panderodus unicostatus (Branson and Mehl, 1933) (EJ-14-411, Fig. 9Y) found immediately below the hardground. Several ozarkodinins tentatively identified as Ozarkodina spp. and Kockelella spp.? (based on P2 elements) and Ozarkodina snajdri? (Walliser, 1964) (P1) were also recovered. A number of specimens showed deformations which are likely to have developed early in the ontogeny, or, if they developed late, they did not affect the occlusal surfaces (e.g. O. confluens in Fig. 9P, compare with the limited occlusal surface for individuals of this species figured by Donoghue and Purnell (1999) and Purnell (1995)). The most common deformations observed were, following the terminology of Weddige (1990), ‘jugatio’, i.e. edged flanks of the basal cavity and cusp, observed in P1 elements of Oz. bohemica (Fig. 10H) with a frequency of ca. 0.5%, and in C. murchisoni with a frequency of ca. 4% (Fig. 9D). The frequencies are calculated against all identifiable (i.e. with preserved cusp or at least its base) elements of a given type (here e.g. P1) belonging to the same species. They are minimum estimates since the shape of all elements shows natural variability (e.g. the outline and curvature of the basal pit in Oz. bohemica) and only unambiguous deformations (as shown in Figs. 9–10) were counted. Likewise, some deformations spanned several of Weddige's (1990) classes, e.g., flanks of the basal pit in Oz. bohemica could also resemble incipient duplicated processes (Fig. 10L), i.e., ‘duplicatio’. The second most common deformation was ‘deflectio’, i.e., deflection of processes. In Oz. bohemica and

C. murchisoni only the dorsal processes of P1 elements were affected with frequencies of 0.7% and 8%, respectively. In Ou. excavatus the natural variability of curvature in both processes was so high that a clear threshold for pathological deflection could not be defined without measurements in three planes. A different type of deflection was observed in one P1 element of Oz. bohemica (Fig. 10I), in which the process appeared to have been broken and re-grown at an angle in the dorso-ventral plane. Protuberances were observed in one P1 element of Oz. ex gr. confluens in the form of overgrowths of hyaline tissue on the lateral surface of the blade, without connection to the basal cavity (Fig. 9P, frequency 0.08%), and in S0 elements of C. murchisoni as pathologically expanded caudal? apical lip (Fig. 9K, frequency 5.7%). These deformations roughly correspond to ‘accessio’, i.e., accessory characters without grave influence on the natural shape, although originally Weddige's (1990) classification was focused on platform-equipped P1 elements and did not discuss if analogous deformations could be identified in S elements. Two cases of ontogenetic deformation affecting only denticles of non-P elements in C. murchisoni have been observed (Fig. 10B, E). In these cases the denticles are partly fused because their growth starts too close (Fig. 10B) or because they grow towards each other (Fig. 10E). Biases in the recovery of different types of elements systematically differed between species and facies (Table 1 and Fig. 3). No significant bias could be detected for C. murchisoni and most samples with W. excavata (except for EJ-14-407 and EJ-14-409, which showed opposite biases). In Ou. excavatus S and M elements were consistently overrepresented in all samples, whereas in Oz. bohemica these elements were consistently underrepresented. Assemblages of Oz. confluens were biased towards P elements significantly except for sample EJ-14410. The relative loss of S and M elements in Oz. bohemica was lowest in the lower part of the section (EJ-14-409 – 40.61%; EJ-14-408 – 37.08%; Fig. 3, Table 1), and for Oz. ex gr. confluens – in the upper part (EJ-14-410 – −4.65%; EJ-14-411 – 20.86%). Among assemblages of Ou. excavatus the most strongly altered was recovered from sample EJ-14408 with a relative gain of S and M elements amounting to 520%. In the least altered Ou. excavatus assemblage S and M elements were overrepresented by 28.21% (EJ-14-411). 4.4. Palynology Sample 14-207 (collected 35 cm below the base of the log in Fig. 3, see Supplementary Table 2) yielded various scolecodonts and other Small Carbonaceous Fossils (SCFs; Butterfield and Harvey, 2011) that we did not identify further. Sample TVDB-13-410 from the same levels has similar organic walled microfossils, one of them identified as elements of the eurypterid respiratory system (Manning and Dunlop, 1995). These ‘Kiemenplatten’ (Fig. 11I, J) are suggested to have assisted in aerial respiration during eurypterid's terrestrial excursions (Manning and Dunlop, 1995). Many of the other SCFs probably represent parts of

Fig. 10. Conodonts from the Gothemshammar 9 section. A–E) Ctenognathodus sp. P? (Viira and Einasto, 2003); A) P1 element, sample EJ-14-410; B) fragment of an unidentified S or M element, sample EJ-14-408; C–D) M element; C) sample EJ-14-411; D) sample EJ-14-408; E1) fragment of an unidentified element with denticles growing into each other, sample EJ14-410; E2) drawing of an outline of E1. F–R) Ozarkodina bohemica (Walliser, 1964); F–H, O–R) P1 elements; F) sample EJ-14-407; G–H) sample EJ-14-410; O, Q) sample EJ-14-407; P, R) sample EJ-14-410; I) fragment of a deformed P1 element, sample EJ-14-407; J) S0 element, sample EJ-14-410; K, M) P2 elements, sample EJ-14-411; L) fragment of a P1 element with initial an additional process, sample EJ-14-407; N) fragment of an S1/2 element, sample EJ-14-411. Scale bar 500 μm. Arrows indicate deformations discussed in the text.

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

315

Table 1 Proportions of conodont elements recovered across the Gothemshammar 9 section. Asterisks indicate results of the goodness of fit test significant at α = 0.05, and overrepresented groups of elements are marked in bold. Number of elements Species

Sample

Ctenognathodus murchisoni

EJ-14-407 EJ-14-411 EJ-14-410

Ozarkodina bohemica

EJ-14-407 EJ-14-411 EJ-14-410 EJ-14-408 EJ-14-409

Ozarkodina confluens

EJ-14-407 EJ-14-411 EJ-14-410 EJ-14-408

Oulodus excavatus

EJ-14-407 EJ-14-411 EJ-14-410 EJ-14-408 EJ-14-409

Wurmiella excavata

EJ-14-407 EJ-14-411 EJ-14-410 EJ-14-408 EJ-14-409

Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected

P

S+M

20 16 5 4.3 4 7.2 403 127.2 385 108.8 133 46.1 64 40.3 22 13.1 179 102.4 62 49.1 53 55.5 86 50.1 60 140.8 78 100 28 78.4 4 24.8 5 12.8 59 26.4 17 11.7 30 26.1 12 9.1 5 11.2

40 44 11 11.7 23 19.8 74 349.8 23 299.2 40 126.9 87 110.7 27 35.9 205 281.6 122 134.9 155 152.5 102 137.9 468 387.2 297 275 266 215.6 89 68.2 43 35.2 40 72.6 27 32.3 68 71.9 22 24.9 37 30.8

arthropod exoskeletons, including eurypterid cuticle fragments (Fig. 11H). The sample also includes large acritarchs that can be observed in the organic residue larger than 51 μm (the bulk of the acritarchs is normally studied in the fraction that filters through a 51 μm sieve) such as Hoegklintia corallina (Fig. 11G). Sample TVDB-14-208 has scolecodonts, SCFs and acritarchs, including abundant Hoegklintia corallina although we remind the reader that we have not attempted an exhaustive study of the acritarch assemblages (see methods). Sample TVDB-14-209 has abundant scolecodonts and SCFs, including fragments of eurypterid cuticle. Sample TVDB-14-210 has the following chitinozoan fauna: Ancyrochitina ansarviensis, Sphaerochitina spp. and 2 specimens of Conochitina. A. ansarviensis is defined by Laufeld (1974b) from Ansarve 2 in Gotland, and the specimens from Gothemshammar 9 are almost identical to the holotype. Typical are the distinctly granular surface, short appendices and chamber/neck length ratio. As in the holotype assemblage, the appendices may split distally. However, in one specimen (Fig. 11A), the spines appear to be multi-rooted, which has not been described before, but may to reflect what (Laufeld, 1974b, p. 41) described as “In some specimens two appendices are attached closely …”. Specimens with a similar silhouette were placed in Sphaerochitina spp. (in

χ2 test statistic

P-value

Loss of S and M elements

1.4

0.243

20%

0.2

0.678

15%

1.9

0.164

−80%

815.5

0.000*

68%

956.1

0.000*

72%

223.0

0.000*

65%

19.1

0.000*

37%

8.3

0.004*

41%

78.1

0.000*

43%

4.6

0.031*

21%

0.1

0.699

−5%

35.0

0.000*

42%

63.2

0.000*

−135%

6.6

0.010*

−28%

44.2

0.000*

−180%

23.8

0.000*

−520%

6.5

0.011*

−156%

54.9

0.000*

55%

3.2

0.073

31%

0.8

0.377

13%

1.3

0.255

24%

4.7

0.031*

−124%

line with Laufeld, 1974a,b; one could also build a case to assign these to Fungochitina spp.; cf. Paris et al., 1999). These specimens were not identified to the species: their chambers are mostly collapsed (e.g. Fig. 11D) and this hampers their identification to the species level, which hinges on the shape of the chamber. Alternatively, the specimens of Sphaerochitina spp. could also be interpreted as poorly preserved specimens of Ancyrochitina ansarviensis, where the fragile appendices were not preserved, which are the sole discriminating feature between these two taxa. The sample also contains large acritarchs (Polygonium polygonale), scolecodonts and SCFs. 5. Discussion 5.1. Stratigraphy The Gothemshammar 9 sections spans the entire upper peak of the Mulde CIE. This peak starts near the base of the C. murchisoni Zone of Jeppsson et al. (2006), i.e. slightly below the boundary between the Halla and Klinteberg formations in western Gotland (Calner et al., 2004). This correlation is supported by the presence of C. murchisoni across the studied interval. The upper boundary of the CIE is variously

316

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Fig. 11. Chitinozoans, acritarchs and eurypteridsPalynomorphs from the Gothemshammar 9 section. A) Ancyrochitina ansarviensis, sample TVDB 14-210, scale bar 0.05 mm; B) Ancyrochitina ansarviensis, sample TVDB 14-210, scale bar 0.05 mm; C) Ancyrochitina ansarviensis, sample TVDB 14-210, scale bar 0.05 mm; D) Sphaerochitina sp., sample TVDB 14210, scale bar 0.05 mm; E) Sphaerochitina sp., sample TVDB 14–210, scale bar 0.05 mm; F) Polygonium polygonale, sample TVDB 14-210, scale bar 0.05 mm; G) Hoegklintia corallina, sample TVDB 14-208, scale bar 0.02 mm; H) eurypterid cuticle fragment, sample TVDB 14–209, scale bar 0.5 mm; I) Eurypterid Kiemenplatten, sample TVDB 13-410, scale bar 0.2 mm; J) Eurypterid Kiemenplatten (detail of ‘I’), sample TVDB 13-410, scale bar 0.02 mm.

placed in different sections, reflecting uncertainties in the position of the Wenlock/Ludlow boundary and substantially variable thicknesses of the uppermost Homerian deposits and possibly also some variability of the carbon isotope values (Blain et al., 2016; Fig. 12). In Gotland the Mulde CIE and also the C. murchisoni Zone is reported as ending within the Klinteberg Fm., and the overlying strata that are below the Wenlock/ Ludlow boundary are unzoned (Calner et al., 2004; Jeppsson et al., 2006). This differs from the δ13Ccarb record in Avalonian sections, where the CIE ends just below the Wenlock/Ludlow boundary (West Midlands) or ranges through it (Wenlock Edge; Marshall et al., 2012; Blain et al., 2016). Continuation of the CIE into the basal Gorstian is also reported from Arctic Canada (Noble et al., 2005; Fig. 12). The case of the Gothemshammar 9 section exemplifies the problem of resolving the position of the Mulde CIE in carbonate successions: the index organisms strongly co-vary with specific facies. C. murchisoni is long known to be a shallow-water specialist (Viira, 1982b; Strömberg, 1997; von Bitter et al., 2007) and its corresponding Zone has originally been defined only in the ‘shelf’ (as opposed to the basin) areas (Viira, 1982a). What is more, the species is absent from areas such as the Prague Basin (Slavík, 2014) or Midland Platform, and has recently been reported from older strata, near the onset of the first peak of the Mulde CIE (Radzevičius et al., 2016). Given all these constraints, the C. murchisoni Zone is mostly misleading in correlations and can only be used locally at best.

5.1.1. Vertebrates The vertebrate remains found in this study generally are in good agreement with previous reports from the Gothemshammar localities on Gotland. Fredholm (1990) reported the same thelodont taxa, with the exception of O. mosaica, but a few more morphological variants are reported in this study. All of the identified anaspids listed here are the same as those listed for Gotland by Blom et al. (2002: Fig. 9), who also revised some of the scales figured in Fredholm (1990). The taxonomy of the osteostracans reported from the Halla Fm., and from Gotland in general, has been subjected to change since Fredholm's (1990) work. Märss et al. (2014) indirectly revised some of the specimens figured by Fredholm (1990) in their review of Silurian East Baltic osteostracans, but the samples presented here further help to elucidate the diversity and distribution of osteostracans in the Baltic Basin during this time. As stated before, the taxa identified here are found in contemporary beds of the East Baltic (see Märss et al., 2014: Fig. 3), but there are also potentially new taxa that may be early representatives of younger genera. However, additional material is needed before any definitive assignments or formal descriptions of these remains are possible. The combined faunal composition, i.e., the presence of Oz. bohemica and Par. martinssoni allows assigning the entire studied interval to the Par. martinssoni vertebrate Zone (Karatajūtė-Talimaa, 1978; Märss and

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

317

Fig. 12. Correlation of Gothemshammar 9 with upper Homerian sections selected to illustrate the variability of the position of the decline of the Mulde CIE with respect to the WenlockLudlow boundary. Lea South Quarry – redrawn from Blain et al. (2016), Wren's Nest Hill – from Ray et al. (2010), Ledai-179 drill core – redrawn from Radzevičius et al. (2014), and Twilight Creek – redrawn from Noble et al. (2005). For explanations of lithological symbols, see the respective articles.

Männik, 2013) and Kockelella ortus absidata conodont Zone sensu Melchin et al. (2012). 5.1.2. Palynomorphs Laufeld, in his extensive study of chitinozoans from Gotland, recovered Ancyrochitina ansarviensis only from the Högklint Beds in Ansarve 2, and suggested this could be an ‘excellent index fossil, because of its short stratigraphic range’ (Laufeld, 1974b, p. 41). Since then, however, the species has been recovered from several stratigraphic levels in the Llandovery, Wenlock and Ludlow of the Baltic region, e.g., by Nestor (2007, 2009); Loydell and Nestor (2005) and Nestor et al. (2002), and from the type Ludlow area of the UK by Sutherland (1994). Therefore, its presence here, in the upper Wenlock Klinteberg Formation, is not necessarily an abnormal occurrence. Nevertheless, it must be noted that many of the occurrences outside of the type stratum concern atypical morphologies (often published in open nomenclature, e.g., Ancyrochitina ansarviensis? in Sutherland, 1994) and the absence of the taxa from strata between the Högklint Fm. at Ansarve (Laufeld, 1974b) and those of the Klinteberg Formation studied here in Gothemshammar 9 is remarkable. Unfortunately, the chitinozoan fauna does not allow us to correlate accurately with assemblages of the same group elsewhere (e.g. see patterns described in Steeman et al.,

2015 for a recent account of the chitinozoan biostratigraphy through the Mulde CIE in the Welsh Basin). The recovered acritarchs have wide stratigraphical distributions on Gotland: Hoegklintia corallina has been recovered from the Wenlock and Ludlow (Upper and Lower Visby, Högklint, and Eke formations and the Slite and Hemse groups) and Polygonium polygonale has been found from Ordovician (subsurface) and entire Silurian on the Island (Le Hérissé, 1989). 5.1.3. Sequence stratigraphy Sequence stratigraphy has proven to be an effective tool in integrating Silurian bio-, chemo- and litho-stratigraphic data (e.g. Calner and Jeppsson, 2003; Ray et al., 2010; Cramer et al., 2015; Blain et al., 2016; Davies et al., 2016; Jarochowska et al., 2016). The position of the hardground in the upper part of the section corresponds to a highorder sea-level fall in the interval of declining δ13C values of the Mulde CIE. In the reef-tract and platform-margin settings in England this sea-level fall is reflected in the falling-stage systems tract corresponding to the upper part of the Nodular Beds Mb. and the basal part of the Upper Quarried Limestone Mb. of the Much Wenlock Limestone Fm. (respectively parasequences 10 and 11 of Ray et al., 2010; Blain et al., 2016; Fig. 12). In intraplatform environments represented e.g. in

318

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Estonia and Ukraine it is manifested in a subaerial unconformity resulting in erosion of a large part of the underlying strata recording the CIE (hiatus between the Rootsiküla and Paadla formations, Märss and Männik, 2013; boundary between depositional sequences 3 and 4 in Jarochowska et al., 2016). Gothemshammar 9 is similar to the latter settings in that it lies within the lagoonal facies belt, spread over substantial areas in the low-relief epicratonic platform. These restricted, lowdiversity facies have naturally low resolution in recording sea-level changes. However, the integration of lithological evidence, i.e. the consistent prograding and emergence contributing to the formation of the hardground, with its position in the lowering limb of the Mulde CIE, supports the correlation of this surface with the sequence boundary traced across the Baltic Basin and in the Midland Platform (Fig. 12). This widely recorded latest Wenlock sea-level fall is followed by a transgression which in the Midland Platform led to the deposition of the major part of the Upper Quarried Limestone Mb. and the Lower Elton Fm. (Ray et al., 2010). The Wenlock/Ludlow boundary is placed at the base of the latter unit in the type Ludlow area (Melchin et al., 2012). At Gothemshammar 9, unit E (Fig. 3) correlates to the basal part of these transgressive strata (i.e., to the middle part of the Upper Quarried Limestone Mb.). The hardground level at Gothemshammar 9 corresponds to a change in sedimentary regime from progradational to retrogradational at the boundary between the Halla and Klinteberg formations in western Gotland (Fig. 3). Although the traditional understanding of these units dating back to Hede (1928) is more topo- than lithostratigraphic, the sequence boundary provides a correlative surface traceable across all facies belts and expressed in ubiquitous changes to less restricted lithologies enriched in skeletal components. The Halla/Klinteberg boundary was placed by Hede (1928) at a hardground surface with truncated and polished fauna, exposed at Gothemshammar 1 (Laufeld, 1974a; Fig. 5h–i). The position of this surface would favor interpretation as the surface of maximum sediment starvation or maximum flooding surface, but uniform truncation of both micritic matrix and skeletal grains is not likely to occur through submarine dissolution and probably represents another episode of emergence. Calner and Jeppsson (2003) pointed to the hardground described here, located ca. 3 m below, as the level of most distinct facies change, and proposed it as a candidate for the boundary between the two formations. Based on the sequence-stratigraphic correlation above, this opinion is supported here. 5.2. Environmental and taphonomic controls on faunal composition The overall low diversity of benthic fauna with monospecific bivalve accumulations and high proportion of microbial carbonates in the uppermost Halla Fm. at Gothemshammar 9 point to radically marginalmarine, restricted conditions. Temporary emergence and desiccation can be ruled out as restricting factors, because – except for the unit E (Figs. 2–3) – the deposition took place below the wave base, with common storm influence. No lithological indicators of elevated salinity have been found. Also the presence of eurypterid fragments cannot be used as evidence for hypersaline conditions, as the common occurrence of these organisms in hypersaline facies seems to result from a combination of their behavior and taphonomy, but not actual physiological tolerance (e.g. Braddy, 2001; Vrazo et al., 2016). Potential restricting factors include therefore fluctuating salinity, oxygen depletion, brackish influence, or elevated temperature, all of which are typical of shallow lagoonal environments with limited water mixing. The presence of at least two species of Ctenognathodus at Gothemshammar 9, virtual absence of coniform taxa recorded in coeval platform-margin settings (e.g. Barrick, 1997; Jarochowska and Munnecke, 2016) and low frequency of W. excavata all are characteristic for marginal-marine, shallow environments (Viira, 1982a,b; Viira and Einasto, 2003; von Bitter et al., 2007; Männik, 2010; Jarochowska et al., 2016). The diverse vertebrate fauna is also indicative of this environment, especially the high abundance and diversity of osteostracans (Märss et al., 2014).

The hypothesis that an environmental gradient exists in terms of conodont ‘robustness’ has been raised in different ways by various authors, e.g. Johnston (1986) observed correlation between the size and ‘robustness’ and grain size of processed rocks. Davies et al. (1993) noted a similar relationship and suggested that the proportion of “robust” elements of Cavusgnathus might be taphonomically enhanced by reworking in wave-agitated settings. The latter authors noted also the overwhelming dominance of P elements over ramiforms in carbonate platform deposits. The role of hydrodynamics in conodont taphonomy was demonstrated by experiments on various morphologies (Broadhead et al., 1990; McGoff, 1991). These studies indicate that element size shows a strong positive relationship with settling velocity and may therefore contribute to postmortem bias in conodont assemblage compositions. For ramiform elements settling velocities were estimated to be up to 10 times lower, possibly resulting in transport over proportionally longer distances. What is more, the concept of conodont ‘robustness’ invokes elements of comparable length, but larger volume of biomineralized tissue. The overrepresentation of robust elements in specific environments may therefore reflect their taxonomic composition (e.g., in the present study it would correspond to the higher proportion of the typically large genus Ctenognathodus in shallow-water facies) or higher proportion of robust ‘individuals’ from the same species (e.g. thicker processes in the widespread species Ou. excavatus and Oz. ex gr. confluens). This overrepresentation may thus be produced by taphonomic processes or reflect original composition of a living conodont community. In disarticulated conodont assemblages it is not possible to reconstruct the composition of the original community. Instead, the degree of taphonomic alteration incurred on the taphocoenosis can be a posteriori estimated through the departure from initial element proportions in complete apparatuses. Biases affecting observed proportions of elements have been discussed by von Bitter and Purnell (2005), who quantified the effects of post-depositional alteration of recovered assemblages of W. excavata (Table 2). These authors used the loss of different types of elements with respect to theoretical frequencies as a proxy for the post-depositional bias. To allow comparison with the present study, previous results were transformed into observed and expected counts of P and S + M elements and evaluated using the χ2 goodness of fit test (Table 2). This approach did not affect the conclusion of von Bitter and Purnell (2005) that P elements were overrepresented in W. excavata assemblages in both carbonate and shale lithologies, but only in the shales this overrepresentation was statistically significant. Similarly, significant overrepresentation in P elements was also reported by Klapper and Murphy (1974); Jeppsson (1974) and Simpson and Talent (1995) from a mixture of various lithologies (Table 2). In contrast, assemblages recovered by Helfrich (1980) from shales had element proportions nearly identical with expected (estimated loss of S + M elements amounting to 2%, compared to 24%–58% in other studies, Table 2). At Gothemshammar 9 only one out of five samples showed significant overrepresentation in P elements, and one showed significant depletion (Table 1). In order to evaluate how this is affected by facies and species-specific apparatus construction, the same test was performed for other abundant taxa. The difference in ‘robustness’ within the apparatus seems to be most pronounced in Oz. bohemica, in which S and M elements are typically very thin and flattened (e.g. Fig. 10J, N). In this species the degree of taphonomic alteration appears – at least in this section – to be related to facies, e.g., assemblages from the lower part of the section showed relatively smaller depletion in ramiform elements (Table 1, Fig. 3). In Oz. ex gr. confluens, belonging to the same genus, but with relatively larger S and M elements, the taphonomic alteration was comparable or smaller than in Oz. bohemica. With respect to this species group, the alteration of all assemblages at Gothemshammar was substantially lower than in all previous studies (Tables 1–2). Comparison between the three species above shows that the bias in conodont recovery is controlled both by facies and morphology of different elements within the apparatus. For C. murchisoni counts

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

319

Table 2 Conodont element proportions reported in previous studies evaluated here using the χ2 goodness of fit test against expected proportions of P and S + M elements. Asterisks indicate results of the test significant at α = 0.05, and overrepresented groups of elements are marked in bold. Reference

Species

Counts

P

S+M

χ2 test statistic

P-value

Loss of S and M elements

Klapper and Murphy (1974)

W. excavata

Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected Observed Expected

2536 1058.7 673 198.1 717 537.7 757 373.9 143 130.1 248 188.3 348 177.3 54 25.3 125 122.9 36 17.3 84 50.9 103 72 102 74.9

1434 2911.3 70 544.9 1298 1477.7 645 1028.1 345 357.9 458 517.7 317 487.7 41 69.7 336 338.1 29 47.7 107 140.1 167 198 179 206.1

2811.2

0.000*

58%

1552

0.000*

71%

0.000*

25%

0.000*

51%

1.7

0.188

9%

25.8

0.000*

24%

0.000*

49%

0.000*

53%

0.828

2%

27.4

0.000*

52%

29.3

0.000*

39%

18.2

0.000*

30%

13.3

0.000*

27%

O. confluens Jeppsson (1974)

W. excavata O. confluens

von Bitter and Purnell (2005)

W. excavata carbonate W. excavata shale

Simpson and Talent (1995)

W. excavata O. confluens

Helfrich (1980)

W. excavata

Viira and Einasto (2003)

O. confluens Ctenognathodus sp. P Ctenognathodus sp. S

Strömberg (1997)

C. murchisoni

of individual elements are only available from Strömberg (1997), who – based on samples pooled from various shallow-water lithologies – reported an average S + M element depletion of 27% (Table 2). In assemblages of other Ctenognathodus species from around the Wenlock/ Ludlow boundary in Estonia, P elements were also significantly overrepresented, but the loss of S + M elements was in the range of 30%– 39%. All the literature reports show higher degrees of alteration than observed here (Table 1). At Gothemshammar 9, the inability to detect significant alteration of the C. murchisoni assemblages seems to result from its overall low abundance, but it opens also the possibility that this species – distinguished by ‘robust’ elements in all positions – is less prone to taphonomic losses than species with more gracile S and M elements, such as Oz. bohemica. This difference in durability of different elements between species may be an important factor in determining the composition of conodont taphocoenoses. E.g., the elevated frequency of C. murchisoni in shallow-water disarticulated assemblages (Viira, 1982b; Viira and Einasto, 2003; Jarochowska et al., 2016) may result from taphonomically produced overrepresentation of ‘robust’ elements. The above test did not allow us to distinguish between biostratinomic, syn-sedimentary processes such as hydrodynamic sorting, predation and scavenging (biases 2–3 of von Bitter and Purnell, 2005) and postsedimentary processes such as reworking in the sediment (e.g. through bioturbation), dissolution and fracturing. It reduced, however, to a minimum the role of differences in element identifiability within and between species. Variations in abundance of studied species did not allow a statistical evaluation of taphonomic effects between shallow-water facies differing in water energy. Unexpectedly, the only sample representing deposition above the wave base, and therefore constant reworking (EJ14-411, Fig. 3), showed low levels of alteration to all assemblages except for Oz. bohemica, which may suggest that post-depositional processes lead to higher alterations than e.g. hydrodynamic sorting, as observed also by von Bitter and Purnell (2005). 5.3. Prevalence of conodont deformations Damage and repair have already been observed by Pander (1856, p. 8) and are inherent to conodonts, reflecting their food-grasping and foodprocessing function (Donoghue and Purnell, 1999). Early ontogenetic

81.628 535.4

224 44.2 0.047

deformations such as duplication or deflection of processes are comparably less understood. Currently there is hardly any data on their prevalence (Weddige, 1990; Corradini et al., 1995) to which findings of the present study could be compared. Nevertheless, such deformations are repeatedly observed (e.g. Walliser, 1964; Klapper and Murphy, 1974; Strömberg, 1997; Slavík et al., 2010) and possibly not even recognized as such (e.g. Sanz-López et al., 2006, Fig. 5.1-4, compare Fig. 10H, L here). Corradini et al. (1995) were the only authors so far to provide frequencies, falling within the range (0.3% to 7.8%) close to that observed at Gothemshammar 9 (0.08% for hyaline overgrowths to 8% for process deflection). Their study emphasized deformations of ramiform elements predominantly from the genus Oulodus (there: Lonchodina). This is in striking contrast to observations from Gothemshammar 9, where among all 1553 examined elements of Ou. excavatus and Ou. siluricus (Supplementary Table 2) not a single comparable deformation has been observed. Based on the illustrations in Corradini et al. (1995), most affected elements were approximately extensiform digyrate, i.e. corresponded to P elements of Oulodus. Lack of deformations visible in Gothemshammar 9 Oulodus assemblages may therefore result from a low percentage of Oulodus P elements being preserved (Fig. 3, Table 1). Corradini et al. (1995) discussed the possibility that deformations in conodonts might be more frequent in specific time slices during the Silurian, associated with e.g. nutrient availability. However, studies discussed earlier and the present study all indicate that facies (i.e. syn- and post-depositional conditions) exert fundamental control on both the composition of living conodont communities and the composition of their taphocoenoses. Therefore systematic studies are needed to establish background levels of specific types of deformations and only then their external controls can be identified. 6. Conclusions 1. The lagoonal succession at Gothemshammar 9 spans the entire upper peak of the upper Homerian Mulde carbon isotope excursion, reaching values up to +2.7‰. 2. A hardground has been identified in the upper part of the succession, correlated here with a sequence boundary present across the Baltic Basin, in the Midland Platform (UK) and the southern shelf of Laurentia. In those sections the Wenlock/Ludlow boundary falls

320

3.

4.

5.

6.

7.

8.

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

within the early phase of the transgressive systems tract of the overlying sequence. The propostition of Calner and Jeppsson (2003) to place the boundary between the Halla and Klinteberg formations at the level of the hardground is supported here based on new evidence. Species richness and abundance of anaspids and osteostracans at Gothemshammar 9 are among the highest reported in the Silurian. The thelodont Oeselia mosaica is reported from Gothemshammar for the first time, along with new occurrences of the osteostracans Tremataspis milleri, Meelaidaspis gennadii, and possible representatives of Tahulaspis, Dartmuthia, and Procephalaspis. Conodont assemblages are dominated by large elements: S and M of Oulodus excavatus, P of Ozarkodina bohemica and Oz. confluens, and nearly balanced assemblages of Ctenognathodus murchisoni. The taxonomic composition and large size of elements are characteristic for late Wenlock marginal-marine environments. Lithological indicators, low diversity of benthic fauna and high of vertebrates, as well as the lack of chitinozoans in the Halla Fm., support this interpretation. Based on a review of previous reports and data presented here, further use of the Ctenognathodus murchisoni zone is discouraged due this conodont's strong affinity to shallow-water facies. Proportions of recovered elements differ between conodont species and may show differences between facies. Ou. excavatus assemblages are consistently depleted in P elements, Oz. bohemica and Oz. confluens show different degrees of S and M element depletion, whereas C. murchisoni assemblages do not deviate significantly from expected element proportions in a 15-element apparatus. Conodonts show early ontogenetic deformations or deformations developed outside occlusal surfaces (frequencies up to 8%).

Acknowledgements The authors thank H. Blom, P. Männik, S. Turner, T. Märss, P. Van Roy and T. Harvey for suggestions and helpful comments, F. Nitsch, J. Velleman and L. Soens for the help in the field, and two anonymous reviewers for constructive comments. EJ, DH, SP and AM acknowledge funding from the Deutsche Forschungsgemeinschaft (Mu 2352/3). EJ received support from SYNTHESYS, which is financed by the European Community – Research Infrastructure Action under the Seventh Framework Programme (FP7/2007-2013). This article is a contribution to the International Geoscience Programme (IGCP) Project 591 – The Early to Middle Paleozoic Revolution.

Appendix A. New locality GOTHEMSHAMMAR 9, 6391090 1678730, ca. 5.1 km NNE of Gothem church. Topographic map 6J Roma NV & NO. Geol. map Aa 169 Slite. Shallow cliff section on the beach. Turn eastward immediately south of the bridge of road 146 crossing the Gothemå, turn first field road to the left and follow the road through the small village Åminne into a forest area. The outcrop is below a short sideway on the left side, just a few meters north of the field road. Halla-Klinteberg fms boundary. Map. KMZ file containing the Google map of the most important areas described in this article.

Appendix B. Supplementary data Supplementary data associated with this article can be found in the online version, at http://dx.doi.org/10.1016/j.palaeo.2016.06.031. These data include the Google map of the most important areas described in this article.

References Agassiz, L., 1839. Fishes of the Old Red Sandstone. In: Murchisson, R.I. (Ed.), The Silurian System. John Murray, London, pp. 587–602. Barrick, J.E., 1997. Wenlock (Silurian) depositional sequences, eustatic events, and biotic change on the southern shelf of North America. Geol. Soc. Am. Spec. Pap. 321, 47–65. http://dx.doi.org/10.1130/0-8137-2321-3. Berg, L.S., 1937. A classification of fish-like vertebrates. Bull. Acad. Sci. URSS 1277–1280. Blain, J.A., Ray, D.C., Wheeley, J.R., 2016. Carbon isotope (δ13Ccarb) and facies variability at the Wenlock-Ludlow boundary (Silurian) of the Midland Platform, UK. Can. J. Earth Sci. http://dx.doi.org/10.1139/cjes-2015-0194. Blom, H., Märss, T., Miller, C.G., 2002. Silurian and earliest Devonian birkeniid anaspids from the Northern Hemisphere. Trans. R. Soc. Edinb. Earth Sci. 92, 263–323. http:// dx.doi.org/10.1017/S0263593300000250. Braddy, S.J., 2001. Eurypterid palaeoecology: palaeobiological, ichnological and comparative evidence for a “mass-moult-mate” hypothesis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 172, 115–132. Branson, E.B., Mehl, E.G., 1933. Conodonts from the Bainbridge (Silurian) of Missouri. Mo. Univ. Stud. 8, 39–52. Broadhead, T.W., Driese, S.G., Harvey, J.L., 1990. Gravitational settling of conodont elements: implications for paleoecologic interpretations of conodont assemblages. Geology 18, 850–853. Butterfield, N.J., Harvey, T.H.P., 2011. Small carbonaceous fossils (SCFs): a new measure of early Paleozoic paleobiology. Geology 40, 71–74. http://dx.doi.org/10.1130/G32580.1. Calner, M., 2005. Silurian carbonate platforms and extinction events—ecosystem changes exemplified from Gotland, Sweden. Facies 51, 584–591. http://dx.doi.org/10.1007/ s10347-005-0050-0. Calner, M., Jeppsson, L., 2003. Carbonate platform evolution and conodont stratigraphy during the middle Silurian Mulde Event, Gotland, Sweden. Geol. Mag. 140, 173–203. http://dx.doi.org/10.1017/S0016756802007070. Calner, M., Jeppsson, L., Munnecke, A., 2004. The Silurian of Gotland — part I : review of the stratigraphic framework, event stratigraphy, and stable carbon and oxygen isotope development. Erlanger Geol. Abh. 5, 113–131. Corradini, C., Olivieri, R., Serpagli, E., 1995. Possible relationships between anomalous conodonts and Silurian oceanic episodes. Neues Jb. Geol. Paläontol. Monat. 12, 737–746. Corradini, C., Corriga, M.G., Männik, P., Schönlaub, H.P., 2015. Revised conodont stratigraphy of the Cellon section (Silurian, Carnic Alps). Lethaia 48, 56–71. http://dx.doi.org/ 10.1111/let.12087. Cramer, B.D., Condon, D.J., Soderlund, U., Marshall, C., Worton, G.J., Thomas, A.T., Calner, M., Ray, D.C., Perrier, V., Boomer, I., Patchett, P.J., Jeppsson, L., 2012. U-Pb (zircon) age constraints on the timing and duration of Wenlock (Silurian) paleocommunity collapse and recovery during the “big crisis”. Geol. Soc. Am. Bull. 124, 1841–1857. http://dx.doi.org/10.1130/B30642.1. Cramer, B.D., Vandenbroucke, T.R.A., Ludvigson, G.A., 2015. High-resolution event stratigraphy (HiRES) and the quantification of stratigraphic uncertainty: Silurian examples of the quest for precision in stratigraphy. Earth Sci. Rev. 141, 136–153. http://dx.doi. org/10.1016/j.earscirev.2014.11.011. Davies, R.B., Austin, R.L., Moore, D., 1993. Environmental controls of Brigantian conodontdistribution: evidence from the Gayle Limestone of the Yoredale Group in Northern England. Ann. Soc Géol. Belg. 116, 221–241. Davies, J.R., Waters, R.A., Molyneux, S.G., Williams, M., Zalasiewicz, J.A., Vandenbroucke, T.R.A., 2016. Gauging the impact of glacioeustasy on a mid-latitude early Silurian basin margin, mid Wales, UK. Earth Sci. Rev. 156, 82–107. http://dx.doi.org/10. 1016/j.earscirev.2016.02.004. Denison, R.H., 1951. Evolution and classification of the Osteostraci. Fieldiana Geol. 11, 155–196. Donoghue, P.C.J., Purnell, M.A., 1999. Growth, function, and the conodont fossil record. Geology 27, 251–254. http://dx.doi.org/10.1130/0091-7613(1999)027. Eichwald, E.v., 1854. Die Grauwackenschichten von Liv- und Esthland. Bull. Soc. Imp. Nat. Moscou 27, 1–111. Fredholm, D., 1990. Agnathan vertebrates in the Lower Silurian of Gotland, Sweden. Geol. Fören. Stockh. Förh. 112, 61–80. http://dx.doi.org/10.1080/11035899009453162. Fredholm, D., 1992. Possible environmental influences on the morphology of fossil vertebrate scales. In: Mark-Kurik, E. (Ed.), Fossil Fishes as Living Animals. Academy of Sciences of Estonia, Tallinn, p. 49. Frykman, P., 1989. Carbonate ramp facies of the Klinteberg Formation, WenlockLudlow transition on Gotland, Sweden. Sver. Geol. Unders. Avh. Uppsats C 820, 1–79. Gross, W., 1967. Über Thelodontier-Schuppen. Palaeontogr. Abt. A 127, 1–67. Gross, W., 1968. Fragliche Actinopterygier-Schuppen aus dem Silur Gotlands. Lethaia 1, 184–218. http://dx.doi.org/10.1111/j.1502-3931.1968.tb01736.x. Hede, J.E., 1928. Berggrunden (Silursystemet). In: Munthe, H., Hede, J.E., Lundquist, G. (Eds.), Beskrivning Till Kartbladet Slite. Sveriges Geologiska Undersökning, Ser. Aa Vol. 169, pp. 13–65. Helfrich, C.T., 1980. Late Llandovery–Early Wenlock conodonts from the upper part of the Rose Hill and the basal part of the Mifflintown formations, Virginia, West Virginia, and Maryland. J. Paleontol. 54, 557–569. Janvier, P., 1985. Les thyestidiens (Osteostraci) du Silurien de Saaremaa (Estonie). Première partie: morphologie et anatomie. Ann. Paléontol. 71, 83–147. Jarochowska, E., Munnecke, A., 2016. Late Wenlock carbon isotope excursions and associated conodont fauna in the Podlasie Depression, eastern Poland: a not-so-big crisis? Geol. J. http://dx.doi.org/10.1002/gj.2674. Jarochowska, E., Munnecke, A., Frisch, K., Ray, D.C., Castagner, A., 2016. Faunal and facies changes through the mid Homerian (late Wenlock, Silurian) positive carbon isotope

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322 excursion in Podolia, western Ukraine. Lethaia 49, 170–198. http://dx.doi.org/10. 1111/let.12137. Jeppsson, L., 1969. Notes on some upper Silurian multielement conodonts. Geol. Fören. Stockh. Förh. 91, 12–24. http://dx.doi.org/10.1080/11035896909448421. Jeppsson, L., 1972. Some Silurian conodont apparatuses and possible conodont dimorphism. Geol. Palaeontol. 6, 51–69. Jeppsson, L., 1974. Aspects of late Silurian conodonts. Fossils Strata 6, 1–54. Jeppsson, L., Calner, M., 2003. The Silurian Mulde event and a scenario for secundo– secundo events. Trans. R. Soc. Edinb. Earth Sci. 93, 135–154. http://dx.doi.org/10. 1017/S0263593300000377. Jeppsson, L., Anehus, R., Fredholm, D., 1999. The optimal acetate buffered acetic acid technique for extracting phosphatic fossils. J. Paleontol. 73, 964–972. Jeppsson, L., Eriksson, M.E., Calner, M., 2006. A latest Llandovery to latest Ludlow highresolution biostratigraphy based on the Silurian of Gotland — a summary. GFF 128, 109–114. Johnston, D.I., 1986. Early Ordovician (Arenig) Conodonts from St. Pauls Inlet and Martin Point, Cow Head Group, Western Newfoundland. Memorial University of Newfoundland. Jones, D., Purnell, M.A., von Bitter, P.H., 2009. Morphological criteria for recognising homology in isolated skeletal elements: comparison of traditional and morphometric approaches in conodonts. Palaeontology 52, 1243–1256. http://dx.doi.org/10.1111/j. 1475-4983.2009.00915.x. Karatajūtė-Talimaa, V., 1978. Thelodonts from the Silurian and Devonian of the USSR and Spitsbergen. Mokslas, Vilnius. Kiær, J., 1911. A New Downtonian Fauna in the Sandstone Series of the Kristiania Area. A Preliminary Report. Vidensk. Skr. I. Mat. E Klasse Vol. 7 pp. 1–22. Klapper, G., Murphy, M.A., 1974. Silurian–lower Devonian conodont sequence in the Roberts Mountains Formation of central Nevada. Geol. Sci. Univ. Calif. Publ. 111, 1–62. Laufeld, S., 1974a. Reference localities for palaeontology and geology in the Silurian of Gotland. Sver. Geol. Unders. Avh. Uppsats C 75 (68), 1–172. Laufeld, S., 1974b. Silurian Chitinozoa from Gotland. Fossils Strata 5, 1–130. Le Hérissé, A., 1989. Acritarches et Kystes d'algues Prasinophycees de l'Ordovicien Superieur et du Silurien de Gotland. Université de Bretagne Occidentale, Brest, France. Lehnert, O., Miller, J.F., Leslie, S.A., Repetski, J.E., Ethington, R.E., 2005. Cambro-ordovician sea-level fluctuations and sequence boundaries: the missing record and the evolution of new taxa. Spec. Pap. Palaeontol. 73, 117–134. Loydell, D.K., Nestor, V., 2005. Integrated graptolite and chitinozoan biostratigraphy of the upper Telychian (Llandovery, Silurian) of the Ventspils D-3 core, Latvia. Geol. Mag. 142, 369. http://dx.doi.org/10.1017/S0016756805000531. Männik, P., 2010. Distribution of Ordovician and Silurian conodonts. Est. Geol. Sect. Bull. 10, 21–24. Manning, P.L., Dunlop, J.A., 1995. The respiratory organs of eurypterids. Palaeontology 38, 287–298. Marshall, C., Thomas, A.T., Boomer, I., Ray, D.C., 2012. High resolution δ13C stratigraphy of the Homerian (Wenlock) of the English Midlands and Wenlock Edge. Bull. Geosci. 87, 669–679. Märss, T., 1989. Vertebrates. In: Holland, C.H., Bassett, M.G. (Eds.), A Global Standard for the Silurian System. International Union of Geological Sciences, Cardiff, pp. 284–289. Märss, T., 2003. Paralogania from the Rootsiküla (Wenlock) and Paadla (Ludlow) stages of Estonia. Proc. Est. Acad. Sci. Geol. 52, 98–112. Märss, T., 2005. Thelodont Oeselia mosaica gen. et sp. nov. from the Wenlock and Ludlow of the East Baltic. Proc. Est. Acad. Sci. Geol. 54, 181–190. Märss, T., 2006. Exoskeletal ultrasculpture of early vertebrates. J. Vertebr. Paleontol. 26, 235–252. Märss, T., Einasto, R., 1978. Distribution of vertebrates in deposits of various facies in the North Baltic Silurian. Eesti NSV Tead. Akad. Toim. Geol. 27, 16–22. Märss, T., Männik, P., 2013. Revision of Silurian vertebrate biozones and their correlation with the conodont succession. Est. J. Earth Sci. 62, 181–204. Märss, T., Turner, S., Karatajūtė-Talimaa, V., 2007. Handbook of Paleoichthyology. Verlag Dr Friedrich Pfeil, Munich. Märss, T., Afanassieva, O., Blom, H., 2014. Biodiversity of the Silurian osteostracans of the East Baltic. Trans. R. Soc. Edinb. Earth Environ. Sci. 105, 73–148. http://dx.doi.org/10. 1017/S1755691014000218. Martinsson, A., 1966. Beyrichiacean ostracodes associated with the earliest Silurian vertebrates from Gotland. Geol. Fören. Stockh. Förh. 88, 327–339. http://dx.doi.org/10. 1080/11035896609448929. McGoff, H.J., 1991. The hydrodynamics of conodont elements. Lethaia 24, 235–247. http:// dx.doi.org/10.1111/j.1502-3931.1991.tb01472.x. Melchin, M.J., Sadler, P.M., Cramer, B.D., 2012. The Silurian Period. In: Gradstein, F.M., Ogg, J.G., Schmitz, M., Ogg, G. (Eds.), The Geologic Time Scale. Elsevier B.V., Amsterdam, pp. 525–558. Munnecke, A., Calner, M., Harper, D.A.T., Servais, T., 2010. Ordovician and Silurian seawater chemistry, sea level, and climate: a synopsis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 296, 389–413. http://dx.doi.org/10.1016/j.palaeo.2010.08.001. Nestor, V., 2007. Chitinozoans in the Wenlock-Ludlow boundary beds of the East Baltic. Est. J. Earth Sci. 56, 109–128. Nestor, V., 2009. Biostratigraphy of the Ludlow chitinozoans from East Baltic drill cores. Est. J. Earth Sci. 58, 170–184. Nestor, V., Einasto, R., Loydell, D.K., 2002. Chitinozoan biostratigraphy and lithological characteristics of the Lower and Upper Visby boundary beds in the Ireviken 3 section, Northwest Gotland. Proc. Est. Acad. Sci. Geol. 51, 215–226. Noble, P.J., Zimmerman, M.K., Holmden, C., Lenz, A.C., 2005. Early Silurian (Wenlockian) δ13C profiles from the Cape Phillips formation, Arctic Canada and their relation to biotic events. Can. J. Earth Sci. 42, 1419–1430.

321

Pander, C.H., 1856. Monographie der fossilen Fische des Silurischen Systems des RussischBaltischen Gouvernements. Akademie der Wissenschaften, St. Petersburg. Paris, F., Grahn, Y., Nestor, V., Lakova, I., 1999. A revised chitinozoan classification. J. Paleontol. 73, 549–570. Patten, W., 1931. New ostracoderms from Oesel. Science 73, 671–673. Purnell, M.A., 1993. Feeding mechanisms in conodonts and the function of the earliest vertebrate hard tissues. Geology 21, 375–377. http://dx.doi.org/10.1130/00917613(1993)021. Purnell, M.A., 1995. Microwear on conodont elements and macrophagy in the first vertebrates. Nature 374, 798–800. http://dx.doi.org/10.1038/374798a0. Purnell, M.A., Donoghue, P.C.J., 1997. Architecture and functional morphology of the skeletal apparatus of ozarkodinid conodonts. Philos. Trans. R. Soc. B Biol. Sci. 352, 1545–1564. http://dx.doi.org/10.1098/rstb.1997.0141. Purnell, M.A., Donoghue, P.C.J., Aldridge, R.J., 2000. Orientation and anatomical notation in conodonts. J. Paleontol. 74, 113–122. http://dx.doi.org/10.1017/S0022336000031292. Radzevičius, S., Spiridonov, A., Brazauskas, A., Norkus, A., Meidla, T., Ainsaar, L., 2014. Upper Wenlock δ13C chemostratigraphy, conodont biostratigraphy and palaeoecological dynamics in the Ledai-179 drill core (Eastern Lithuania). Est. J. Earth Sci. 63, 293–299. Radzevičius, S., Spiridonov, A., Brazauskas, A., Dankina, D., Rimkus, A., Bičkauskas, G., Kaminskas, D., Meidla, T., Ainsaar, L., 2016. Integrated stratigraphy, conodont turnover and palaeoenvironments of the upper Wenlock and Ludlow in the shallow marine succession of the Vilkaviškis-134 core (Lithuania). Newsl. Stratigr. 49, 321–336. Ray, D.C., Brett, C.E., Thomas, A.T., Collings, A.V.J., 2010. Late Wenlock sequence stratigraphy in central England. Geol. Mag. 147, 123–144. http://dx.doi.org/10.1017/ S0016756809990197. Robertson, G.M., 1935. Oeselaspis, a new genus of ostracoderm. Am. J. Sci. 29, 453–461. Robertson, G.M., 1939. An Upper Silurian vertebrate horizon with description of a new species, Cephalaspis oeselensis. Trans. Kans. Acad. Sci. 42, 357–363. Rohon, J.V., 1892. Die obersilurischen Fische von Oesel. I Teil. Thyestidae und Tremataspidae. Mém. Acad. Imp. Sci. St. Pétersbg. 38, 1–88. Samtleben, C., Munnecke, A., Bickert, T., 2000. Development of facies and C/O-isotopes in transects through the Ludlow of Gotland: evidence for global and local influences on a shallow-marine environment. Facies 43, 1–38. Sandberg, C.A., Dreesen, R., 1984. Conodont Biofacies and Provincialism. Geological Society of America Special PapersGeological Society of America http://dx.doi.org/10. 1130/SPE196. Sansom, R.S., 2008. The origin and early evolution of the Osteostraci (Vertebrata): a phylogeny for the Thyestiida. J. Syst. Palaeontol. 6, 317–332. Sanz-López, J., Perret, M.-F., Vachard, D., 2006. Silurian to Mississippian series of the eastern Catalan Pyrenees (Spain), updated by conodonts, foraminifers and algae. Geobios 39, 709–725. http://dx.doi.org/10.1016/j.geobios.2005.06.003. Schmidt, F., 1866. Über Thyestes verrucosus Eichwald und Cephalaspis schrenkii Pander, nebst einer Einleitung über das vorkommen silurischen Fischreste auf der Insel Oesel. Ver. Kais. Russ. Mineral. Ges. St. Petersbg. 2, 217–250. Simpson, A.J., Talent, J.A., 1995. Silurian conodonts from the headwaters of the Indi (upper Murray) and Buchan rivers, southeastern Australia, and their implications. Cour. Forschungsinstitut Senckenberg. 182, 79–215. Slavík, L., 2014. Revision of the conodont zonation of the Wenlock-Ludlow boundary in the Prague Synform. Est. J. Earth Sci. 63, 305–311. Slavík, L., Kříž, J., Carls, P., 2010. Reflection of the mid-Ludfordian Lau event in conodont faunas of Bohemia. Bull. Geosci. 85, 395–414. Steeman, T., Vandenbroucke, T.R.A., Williams, M., Verniers, J., Perrier, V., Siveter, D.J., Wilkinson, J., Zalasiewicz, J., Emsbo, P., 2015. Chitinozoan biostratigraphy of the Silurian Wenlock–Ludlow boundary succession of the Long Mountain, Powys, Wales. Geol. Mag. 153, 95–109. http://dx.doi.org/10.1017/S0016756815000266. Strömberg, C., 1997. The Conodont Genus Ctenognathodus in the Silurian of Gotland. Lund University, Sweden. Sutherland, S.J.E., 1994. Ludlow chitinozoans from the type area and adjacent regions. Palaeontogr. Soc. Monogr. 148, 1–104. Turner, S., 1999. Early Silurian to Early Devonian thelodont assemblages and their possible ecological significance. In: Boucot, A.J., Lawson, J. (Eds.), Palaeocommunities: A Case Study from the Silurian and Lower Devonian — International Geological Correlation Programme 53, Project Ecostratigraphy, Final Report. Cambridge University Press, Cambridge, pp. 42–78. Van den Boogaard, M., Kuhry, B., 1979. Statistical reconstruction of the Palmatolepis apparatus (Late Devonian conodontophorids) at the generic, subgeneric and specific level. Scr. Geol. 49, 1–57. Viira, V., 1982a. Late Silurian shallow and deep water conodonts of the East Baltic. In: Kaljo, D., Klaamann, E. (Eds.), Ecostratigraphy of the East Baltic Silurian. Valgus, Tallinn, pp. 79–88. Viira, V., 1982b. Shallow-water conodont Ctenognathodus murchisoni (Late Wenlock, Estonia). In: Kaljo, D., Klaamann, E. (Eds.), Communities and Biozones in the Baltic Silurian. Valgus, Tallinn, pp. 63–83. Viira, V., 1983. Upper Silurian Spathognathodus (conodonts) from Estonia. In: Klaamann, E. (Ed.), Paleontology of Early Paleozoic of the East Baltic and Podolia. Academy of Sciences of the Estonian Institute of Geology, Tallinn, pp. 41–68. Viira, V., 1994. A new upper Silurian conodont species from Estonia. Proc. Est. Acad. Sci. Geol. 43, 32–37. Viira, V., Einasto, R., 2003. Wenlock–Ludlow boundary beds and conodonts of Saaremaa Island, Estonia. Proc. Est. Acad. Sci. Geol. 52, 213–238. von Bitter, P.H., Purnell, M.A., 2005. An experimental investigation of post-depositional taphonomic bias in conodonts. Spec. Pap. Palaeontol. 73, 39–56. von Bitter, P.H., Purnell, M.A., Tetreault, D.K., Stott, C.A., 2007. Eramosa Lagerstätte — exceptionally preserved soft-bodied biotas with shallow-marine shelly and bioturbating organisms (Silurian, Ontario, Canada). Geology 35, 879–882.

322

E. Jarochowska et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 457 (2016) 304–322

Vrazo, M.B., Brett, C.E., Ciurca, S.J., 2016. Buried or brined? Eurypterids and evaporites in the Silurian Appalachian basin. Palaeogeogr. Palaeoclimatol. Palaeoecol. 444, 48–59. http://dx.doi.org/10.1016/j.palaeo.2015.12.011. Walliser, O.H., 1964. Conodonten des Silurs. Abh. Hess. Landesamtes Bodenforsch. Wiesbad. 41, 1–106.

Weddige, K., 1990. Pathological conodonts. Cour. Forschungsinstitut Senckenberg. 118, 563–589. Woodward, A.S., 1891. Catalogue of the Fossil Fishes in the British Museum (Natural History). British Museum, London.