Darriwilian ( Middle Ordovician ) conodont biostratigraphy in NW Estonia

Darriwilian conodonts have been studied in numerous sections of Baltoscandia, but few data are available from the classical outcrop area in northern Estonia. In this paper we report the succession of Darriwilian conodonts and chitinozoans from the bed-by-bed sampled Uuga and Osmussaar cliffs, NW Estonia. Standard Baltoscandian conodont and chitinozoan zones and subzones were identified from the topmost Volkhov to Uhaku regional stages. The global DapingianDarriwilian boundary coincides with the VolkhovKunda boundary in NW Estonia. The Kunda Stage is represented by its middle or upper part only, and the Aseri Stage is probably missing in both sections studied. The lowermost part of the Lasnamägi Stage with the conodont Yangtzeplacognathus foliaceus is also very condensed. The rest of the Lasnamägi and Uhaku stages, corresponding to the main part of the Pygodus serra conodont Zone, is well represented and can be correlated across Estonia and Sweden. Subtle regional variations in this interval indicate very uniform depositional conditions over wide areas of the Baltoscandian palaeobasin. Our results suggest that the Y. foliaceus Subzone in Estonia needs further assessment. Stratigraphically well-constrained Y. protoramosus appears to be more common than previously thought, and is thus a valuable regional subzonal index. The integrated conodont and chitinozoan scale provides more than 20 biostratigraphically important levels for local and regional correlations, probably approaching temporal resolution in the order of 0.1 Ma for the late Darriwilian.


INTRODUCTION
Well-preserved and fossiliferous carbonate rocks of Darriwilian age, Middle Ordovician, are widespread in Baltoscandia and widely utilized locally as building material.Divided into the uppermost Volkhov, Kunda, Aseri, Lasnamägi and Uhaku regional stages, they bear evidence of rapid diversification of many groups of organisms (Hints et al. 2010 and references therein).Nowadays mostly microfossils are used to date and correlate these strata, whereas conodonts and chitinozoans are especially useful.The standard North Atlantic conodont biozonation (Bergström 1971), widely used in Baltoscandia and elsewhere, is based on several successive zones and subzones, some of which have clearly evolutionary significance (Bergström 1983;Zhang 1998).Darriwilian conodont zones and subzones have been identified in numerous sections in Scandinavia and Estonia.In recent years new information has been obtained particularly from drill cores (Bergström et al. 2011;Männik & Viira 2012, and references therein).However, very few data are still available from the classical outcrop area of northern Estonia, which serves as the standard for many Baltic regional stages.Therefore the original concept of regional stages has become difficult to follow.Furthermore, in drill cores the sampling density and sample size often limit the temporal resolution and complicate detailed comparison of biozonations based on different fossil groups.
Recently Tammekänd et al. (2010) analysed the succession of Darriwilian chitinozoans from bed-by-bed samples at the Uuga Cliff, Pakri Peninsula, NW Estonia.This enabled biostratigraphic dating of Darriwilian strata and highlighted problems of regional stratigraphy.A large collection of conodonts was also obtained from the same set of samples, which forms the basis of our present study.Previously Lower Ordovician conodonts from the same section have been reported by Löfgren et al. (2005).In addition, Mellgren et al. (2012) published Darriwilian conodonts from Osmussaar Island, NW Estonia, ca 40 km west of the Uuga Cliff.This motivated us to re-study a series of old samples from the Osmussaar Cliff and to extend the succession of Mellgren et al. (2012) into the Lasnamägi and Uhaku regional stages.
This contribution aims to (1) biostratigraphically characterize two Darriwilian reference sections in NW Estonia, (2) analyse mutual distribution of conodonts and chitinozoans, (3) test and refine regional correlation criteria, (4) assess temporal resolution of palaeontological time scales and (5) compare the conodont succession in NW Estonia with previously studied sections elsewhere in Estonia and Sweden.
The thickness of Darriwilian strata in NW Estonia is about 10 m, of which 7 m is exposed and sampled at the Uuga and Osmussaar cliffs.The Volkhov Stage is represented by glauconitic limestones of the Toila Formation, only the topmost part of which was studied here.
The 0.5-1.1 m thick Kunda Stage is represented by sandy limestones to calcareous sandstones of the Pakri Formation, which is distributed in NW Estonia only and includes the so-called Osmussaar Breccia (see Suuroja et al. 2003).The breccia is characterized by specific sedimentary structures, occurrence of chromite grains (Alwmark et al. 2010) and at least partly mixed fossil associations (Tinn et al. 2010).All these features point to an impact-related event in Kunda time (see also discussion by Alwmark et al. 2010 andMellgren et al. 2012).
A thin limestone layer rich in quartz sand and phosphatic ooids (0.1-0.2 m in the studied sections), representing the Kandle Formation, lies on top of the Pakri Formation.On Osmussaar this unit has been assigned either to the Aseri Stage (Orviku 1940;Rõõmusoks 1970;Hints 1997) or Lasnamägi Stage (Männil & Rõõmusoks 1984;Mellgren et al. 2012).
The Väo Formation (5.1 m in the Uuga Cliff section) is composed of relatively pure limestones (wacketo packstones), subdivided into the Rebala, Pae and Kostivere members.The dolomitic Pae Member is well defined at the Uuga Cliff (Tammekänd et al. 2010, fig.2), but on Osmussaar similar dolomitization has not occurred and recognition of the three members remains tentative.The Lasnamägi-Uhaku stage boundary falls within the Väo Formation and is defined by the appearance of the graptolite Gymnograptus linnarssoni and the conodont Baltoplacognathus robustus (Männil 1976;Männil & Meidla 1994).Numerous impregnated and commonly bioturbated discontinuity surfaces are found throughout the Väo Formation.Several individual beds and surfaces (described in detail from the Pakri Peninsula by Einasto & Rähni 2005) can be traced over wide areas in northern Estonia, denoting a very flat sea floor and uniform depositional conditions (Jaanusson 1976).
The Kõrgekallas Formation is represented by slightly more argillaceous limestones above the so-called sixfold discontinuity surface (Orviku 1940;Rõõmusoks 1970;Einasto & Rähni 2005).The lower part of the Kõrgekallas Formation is exposed and has been sampled in the Uuga Cliff, whereas the Osmussaar Cliff exhibits only the uppermost part of the Väo Formation.

MATERIAL AND METHODS
This study is based on 143 samples collected mostly bed-by-bed from the NW Estonian coastal cliffs on the Pakri Peninsula and Osmussaar Island.The Uuga Cliff section (Pakri Peninsula) was sampled in 2006 by Olle Hints and Jaak Nõlvak (Fig. 2; samples OM6-1 to OM6-62 from the Kandle to Kõrgekallas formations; chitinozoan data were published by Tammekänd et al. 2010).Subsequently conodonts were studied from the same residues and additional samples were collected from the Toila and Pakri formations (OM11-1 to OM11-4; Fig. 2).The Osmussaar section is a composite of two sampling sites (Fig. 1) from where nearly 77 samples were collected already in 1971 by Ralf Männil and Lembit Põlma (Fig. 3;.Stratigraphically the studied part of the Osmussaar Cliff is limited to the Kandle and Väo formations. Conodonts, chitinozoans and other acid-resistant microfossils (scolecodonts and graptolites) were extracted from ca 0.3 to 2 kg limestone samples using ca 10% acetic acid.The residues were cleaned from clay either by decanting or 0.05 mm sieve.Microfossils were handpicked from the residues and stored in standard microfossil cavity slides (conodonts) or in glycerine preparations (chitinozoans, scolecodonts).The Osmussaar samples, processed in the 1970s, were picked for conodonts only selectively, with emphasis on biostratigraphically important species.Conodonts were identified by Viive Viira and Olle Hints, and chitinozoans by Jaak Nõlvak.For a few graptolite finds unpublished identifications by Ralf Männil were used.Selected conodonts from the more complete Uuga Cliff were imaged with a Zeiss EVO MA15 scanning electron microscope.
The research was conducted at the Institute of Geology at Tallinn University of Technology (abbreviated as GIT), where the samples, residues and all recovered specimens are deposited.For the figured specimens the collection number 654 is allocated.More information on individual samples and specimens can be obtained from the on-line catalogue at http://sarv.gi.ee.

Conodont biostratigraphy
Conodonts are generally very well preserved and abundant in the Darriwilian strata of the Uuga and Osmussaar cliffs (with the exception of the Pakri and Kandle formations).The conodont colour alteration index (CAI) is estimated to be 1, indicating burial temperatures well below 100 °C (Epstein et al. 1977).The taxonomic composition of the Darriwilian conodont fauna is similar to that of previously studied sections of Estonia.The assemblages are commonly dominated by species of Baltoniodus, Semiacontiodus and Panderodus.A particular feature of the studied sections seems to be the low abundance of Protopanderodus and absence of Costiconus (= Walliserodus).Biostratigraphically important species were found throughout the succession, enabling the standard North Atlantic conodont zones and subzones (Bergström 1971;Nõlvak et al. 2006) from the Paroistodus originalis to Pygodus serra zones (tentatively to the P. anserinus Zone) to be followed.The most useful correlative levels are provided by the successive subzones of the P. serra Zone, which include one to three evolutionary lineages of polyplacognathids (Bergström 1971(Bergström , 1983;;Dzik 1978;Zhang 1998).The genus level classification of that group of platform conodonts is still under discussion and different generic names have been used by different authors.Herein we follow Zhang (1998), who divided Eoplacognathus sensu lato into Eoplacognathus, Baltoplacognathus and Yangtzeplacognathus.The distribution of individual taxa in the Uuga and Osmussaar sections is shown in Figs 2 and 3 and selected conodont taxa are illustrated in Figs 4-6.The succession of conodonts with emphasis on stratigraphically useful taxa is characterized below.
The Toila Formation is represented by a single sample from the Uuga Cliff (OM11-1).The most abundant species in this sample are Baltoniodus navis (Lindström) and Microzarkodina flabellum Lindström.The occurrence of Paroistodus originalis together with Triangulodus brevibasis (Sergeeva) indicates the Paroistodus originalis Zone (Löfgren 1995), which corresponds to the middle part of the Volkhov Stage (Nõlvak et al. 2006;Männik & Viira 2012) and the upper part of the Dapingian (Cooper & Sadler 2012).The absence of Baltoniodus Fig. 2. Distribution of conodonts in the Uuga Cliff section.For a full range chart of chitinozoans, general rock composition and stable carbon isotope record from the same samples see Tammekänd et al. (2010) The Pakri Formation, studied in the Uuga Cliff, yielded mainly broken and worn conodonts of the genera Scolopodus, Paltodus, Acodus and Cordylodus (note that they are excluded from Fig. 2).These genera are typical of the Hunneberg and Billingen stages of  boundary of the zone cannot be identified in either section, but most likely the entire Pakri Formation corresponds to the E. pseudoplanus Zone in the study area.
The lowermost sample from the Kandle Formation in the Uuga Cliff section (OM6-1) contains no zonal species, but Panderodus sulcatus (Fåhraeus) and Sagittodontina sp.indicate the Eoplacognathus suecicus Zone or younger strata.The next samples (OM6-2 and OM6-3) yield characteristic elements of Yangtzeplacognathus foliaceus (Fåhraeus), the index species of the Y. foliaceus Subzone of the Pygodus serra Zone.Pygodus serra is very rare in both sections and was identified in only three samples from the Väo Formation (Figs 2, 3).Interestingly, sample OM6-3 also contained a few conodont elements that resemble E. suecicus and Baltoplacognathus reclinatus (Fig. 4I-K).It is possible that other polyplacognathid species co-occur with Y. foliaceus or the specimens in question indicate reworking from under-or overlying beds, but no firm conclusions can be drawn due to the limited number of specimens currently available.Lithologically similar strata of the Osmussaar Cliff contained no diagnostic taxa in our sample (M-5851).Mellgren et al. (2012) identified Y. foliaceus (= Eoplacognathus foliaceus) from approximately the same level at Osmussaar (sample Est09-OS3), however, their illustrated specimens (Mellgren et al. 2012, fig.6C,  G, Q) are rather different from typical Y. foliaceus, which are found at Uuga (Fig. 4A-H).Regardless of that, the Y. foliaceus Subzone, if present, is very thin in NW Estonia.The oolithic limestone of the Kandle Formation appears to be of Lasnamägi rather than Aseri age in the studied sections, confirming the interpretation of Männil & Rõõmusoks (1984) and Mellgren et al. (2012), but disagreeing with, e.g., Orviku (1940), Rõõmusoks (1970) and Hints (1997).
The first specimens of Baltoplacognathus reclinatus (Fåhraeus) appear in the Väo Formation, marking the base of the succeeding B. reclinatus Subzone.Characteristic specimens of this species are found in samples OM6-4 to OM6-24 at the Uuga Cliff and M-5854 to M-5860 at the Osmussaar Cliff.Samples from OM6-25 to OM6-28 from the Uuga Cliff contain few and/or fragmentary specimens, some of which appear to be intermediate between B. reclinatus and the morphologically closely connected successor species B. robustus.A similar unclear interval is observed in the Osmussaar succession from samples M-5865 to M-5874a.In this study the base of the Baltoplacognathus robustus Subzone is drawn at the level where the first typical mature Pb elements of the index species appear (sample OM6-29 at Uuga and M-5877 on Osmussaar).This level is slightly above the marked discontinuity surface within the so-called 'Raudsüda' bed (the lithological boundary criterion for the Uhaku Stage).The entire B. robustus Subzone is 0.65-0.8m thick in the studied sections.
The base of the Yangtzeplacognathus protoramosus Subzone is drawn at the first appearance of the nominal species.It is significant that the oldest specimens of Y. protoramosus co-occur with the youngest recorded specimens of B. robustus in just one sample in both sections (OM6-35 and M-5891).Possibly these samples span across a sharp zonal boundary and/or the boundary interval is time-averaged.A more plausible explanation is that the two species represent separate evolutionary lineages as suggested by Zhang (1998) and hence their mutual occurrence would be expected.In any case the lower and upper boundaries of the Y. protoramosus Subzone are sharp in both sections and no interval of morphological transition could be observed.The Y. protoramosus Subzone is 0.4-0.5 m thick in the studied sections and embraces a distinct lithological marker (the so-called double discontinuity surface) in its lower part (Figs 2, 3).
The Eoplacognathus lindstroemi Subzone occurs in both sections just above the Y. protoramosus Subzone and no range overlap of the nominal species was recorded.At the Uuga Cliff, E. lindstroemi was identified in successive samples from OM6-40 to OM6-61, being most abundant in samples OM6-46 and OM6-47.The upper boundary of the subzone is defined by the appearance of Pygodus anserinus, which has not been found in the studied sections.However, the uppermost part of the Uuga Cliff section yielded rare specimens of Sagittodontina kielcensis (Dzik), which are used here to tentatively draw the base of the Pygodus anserinus Zone and its lower S. kielcensis Subzone.The succession studied on Osmussaar ends within the middle part of the E. lindstroemi Subzone in the uppermost part of the Väo Formation.
In addition to the above-mentioned index species, few other taxa are stratigraphically useful and worth mentioning.

Notes on chitinozoans and graptolites
The succession of chitinozoans at the Uuga Cliff was published by Tammekänd et al. (2010).Additional samples (OM11-1 to OM11-4) revealed the presence of the Conochitina cucumis and Cyathochitina regnelli zones, which correspond to the Volkhov and Kunda stages, respectively (Fig. 2; for a full range chart see Tammekänd et al. 2010, fig. 5).Chitinozoan data from the Osmussaar Cliff are unpublished and comments are included here.The preservation of material is good to excellent, similar to that at the Uuga Cliff (see Tammekänd et al. 2010).Altogether 36 species have been identified.The distribution of selected species with reference to the Baltoscandian chitinozoan zonation (Nõlvak & Grahn 1993;Nõlvak et al. 2006) is shown in Fig. 3.
The entire succession studied corresponds to the Laufeldochitina striata Zone, with the nominal species occurring throughout the section.The disappearance of Tanuchitina tallinnensis Grahn, Belonechitina crinita Grahn, Cyathochitina sebyensis Grahn and Belonechitina pellifera (Eisenack) in the lower part of the Väo Formation is followed on Osmussaar in the same sequence as at Uuga.Lagenochitina tumida Umnova ranges slightly higher in the Osmussaar section, whereas the stratigraphically restricted forms of Conochitina with specific soft brownish covering range slightly lower than at Uuga.In the middle and upper parts of the Väo Formation, the first abundant occurrence of Belonechitina cactacea (Eisenack) and Conochitina aff.dolosa Eisenack, and the appearance of Conochitina aff.tuberculata Eisenack are recorded in the same order in both sections.The occurrence of C. aff.tuberculata in the topmost 0.3 m of the Osmussaar Cliff section suggests that there most of the Väo Formation is present.
Two species of graptolites recovered from the Osmussaar Cliff are of special interest.Gymnograptus linnarssoni (Moberg) was identified in a number of samples within the B. robustus and Y. protoramosus zones and the basal part of the E. lindstroemi Zone (Fig. 3).This species is used to define and correlate the base of the Uhaku Stage (Männil 1976;Männil & Meidla 1994), but in many sections, for example at the Uuga Cliff, it has not been recovered or occurs sporadically.The occurrence in the Osmussaar Cliff section, even if not representing the full range of the species in Baltoscandia, enables integration with conodont and chitinozoan zonations.The conodont succession suggests that the single find of G. linnarssoni at the Uuga Cliff (Tammekänd et al. 2010, fig. 5) represents most likely the middle part of its range.Another useful graptolite, Gymnograptus cf.retioloides Wiman, is recorded in two samples (M-5914, M-5915; identification by Ralf Männil) near the top of the Osmussaar Cliff slightly above the appearance of Conochitina aff.tuberculata, with the same order of occurrence and position as at the Uuga Cliff.

Integrated biozonation and correlation of the Uuga and Osmussaar sections
Correlation of the Uuga and Osmussaar sections based on integrated conodont and chitinozoan biostratigraphy is shown in Fig. 8. Nearly 20 correlative levels can be recognized in the Väo Formation (Lasnamägi and lower Uhaku stages) and both sections display a markedly similar sequence of events, with no cross-correlations between chitinozoan and conodont data.A scatter-plot representation of the correlation is presented in Fig. 9, showing that the chitinozoans seem to have slightly higher dispersion (i.e.variation between the sections).
Multiple options for correlation are evident only in the lowermost part of the succession, where Yangtzeplacognathus foliaceus and chitinozoans Baltochitina delicata and Desmochitina grandicolla indicate a more complete section at Uuga, whereas the appearance of Baltoplacognathus reclinatus and Conochitina sp.n. 1 of Tammekänd et al. (2010) allow of slightly different interpretation (Fig. 8).This discrepancy cannot be resolved without additional sampling efforts.Other conodont subzone boundaries correspond quite well with chitinozoan levels.For instance, Baltoplacognathus robustus appears 10-20 cm above the disappearance of Belonechitina pellifera and E. lindstroemi appears very close to the LAD of Conochitina clavaherculi (Fig. 8) in both sections.The latter level in the Osmussaar Cliff section coincides also with the LAD of Gymnograptus linnarssoni.
Both chitinozoans and conodonts show that the ageequivalent strata from the base of the Lasnamägi Stage to the uppermost common marker level (FAD of Conochitina aff.tuberculata) are about 1 m thicker at the Uuga Cliff.This agrees well with lithological and macrofossil data, and follows the westward thinning of the Darriwilian succession in northern Estonia (e.g., Orviku 1940;Hints 1997).
Integrated conodont and chitinozoan data show that the difference in thickness is not distributed equally across the sequence.The lower part of the Väo Formation (Rebala Member), correlated by the LAD or Belonechitina crinita, is of nearly the same thickness in both sections, if not slightly thicker at Osmussaar.In contrast, the interval between the LAD of Belonechitina crinita and FAD of Baltoplacognathus robustus is ca 1.4 m Uuga, but only 0.75 m at Osmussaar.This interval contains a number of discontinuity surfaces including the one used as a lithological marker for the lower boundary of the Uhaku Stage in northern Estonia (Einasto et al. 1996;Einasto 2002).As the B. robustus Subzone is of nearly equal thickness in both sections, the missing strata are probably on account of the topmost Lasnamägi Stage rather than the basal Uhaku Stage.Whether there is just one gap, or a set of smaller ones, remains beyond the current biostratigraphical resolution.The Y. protoramosus Subzone and common part of the E. lindstroemi Subzone are only slightly thinner in the Osmussaar section.

Regional comparison of conodont biozonation
As shown above, the conodont distribution patterns in both studied sections are very similar and provide a good basis for characterization of the NW Estonian Darriwilian succession.However, it is also of interest to explore how our new data compare with those available from other parts of Estonia as well as Scandinavia.A comparison of the thicknesses of conodont zones and subzones in selected sections is further discussed below and illustrated in Table 1.
Only two earlier studies of the northern Estonian outcrop area can be used for comparison (Männil 1986;Viira et al. 2001), both from sections at Lasnamägi, Tallinn.Männil (1986, fig. 2.1.1)provided generalized ranges of upper Darriwilian index species from Y. foliaceus to E. lindstroemi.Viira et al. (2001), on the other hand, included material from the lower Darriwilian, up to the Y. foliaceus Zone.The combined section at Lasnamägi appears to be more complete, especially in the Kunda to Aseri interval, where the Baltoniodus norrlandicus and Eoplacognathus suecicus zones are well defined.In the Lasnamägi and Uhaku stages (i.e.Pygodus serra Zone), the same successive index taxa have been observed both in NW Estonia and the Tallinn area.These include the short-ranging Y. protoramosus, which was first recorded in Estonia by Männil (1986) under an informal name Eoplacognathus bergstroemi.As in NW Estonia, the zonal index P. serra is very rare in the Tallinn area.
A notable difference between our data and earlier studies in the Tallinn area concern the Y. foliaceus Subzone.Viira et al. (2001) listed this taxon from their uppermost sample (Mä95-51), defining the base of the corresponding subzone.However, these authors illustrated a sinistral Pb element (Viira et al. 2001, fig.8U), which diverges considerably from the typical specimens of Viira (1974, pl. 8:10-11) and those from the Uuga Cliff (Fig. 4).According to Männil (1986, fig. 2.1.1),the Y. foliaceus Subzone extends to the top of the Pae Member in Tallinn, which strongly disagrees with our data from the Uuga Cliff.Although the Pae Member may be diachronous and the Väo Formation is more complete in the Tallinn area, a number of common chitinozoan species in the lower part of the Väo Formation suggest a different explanation; it is likely that different taxonomic concepts of Y. foliaceus have been applied.Viira (1974, fig. 8) has nevertheless shown that the Y. foliaceus Subzone is present in the Tallinn area, but its boundaries obviously need re-evaluation.Since the original collection of Männil (1986) cannot be located at present, this problem can only be resolved by resampling the Lasnamägi type section.
The conodont succession in other parts of Estonia has recently been studied in the Kerguta (Viira et al. 2006b), Taga-Roostoja (Viira & Männik 1999), Mehikoorma (Männik & Viira 2005), Valga (Männik 2001) and Ruhnu cores (Männik 2003).Additionally, Viira (2011) showed the distribution of Darriwilian conodonts in a number of core sections in southern Estonia and adjacent Russia.In summary, the pre-serra Darriwilian strata are notably thicker and more complete in most of these sections (Table 1).For instance, the lower Darriwilian conodont succession from the B. norrlandicus to E. suecicus zones is ca 20 m thick in the Mehikoorma core in SE Estonia (Männik & Viira 2005).This interval corresponds to about 1 m at the Uuga Cliff, represented by the E. pseudoplanus Zone only.
In contrast, the P. serra Zone is much less variable across Estonia.All five subzones identified in NW Estonia have been recognized in other sections in similar or only slightly thicker units (Table 1), suggesting that unlike in the early Darriwilian, the depositional conditions were Table 1.Comparison of thicknesses of conodont zones and subzones in the studied sections and localities in Estonia and Sweden.The data come from the following sources: Kerguta core (Viira et al. 2006b), Taga-Roostoja core (Viira & Männik 1999), Mehikoorma core (Männik & Viira 2005), Valga core (Männik 2001) and Ruhnu core (Männik 2003).The Lasnamägi section is compilation from an outcrop (Viira et al. 2001) and the unpublished Lasnamägi 198 drill core.Swedish data from Bergström (2007) and Bergström et al. (2011).For locality map see Fig. 1.Note that the Y. protoramosus Subzone in the Taga-Roostoja section was corrected by restudying the original collection.In the Valga core, a sample with questionably identified Y. protoramosus is omitted.See further explanation in text  (Bergström 2007;Bergström et al. 2011).
The E. suecicus Zone and Y. foliaceus Subzone are relatively thick at Kårgärde, 5 and 2 m, respectively, and correspond to a gap in the NW Estonian succession.The B. reclinatus and B. robustus subzones are thicker at Kårgärde (4 and 2 m, respectively), but otherwise directly comparable to those in NW Estonia.Interestingly, the ca 0.5 m range of Y. protoramosus is reported to overlap entirely with the E. lindstroemi Subzone (Bergström 2007; but note a confusion regarding this identification: Y. protoramosus is mentioned in the text on p. 32, but fig.6 lists the range of Eoplacognathus jiangeensis instead).This is not surprising as Y. protoramosus likely represents a different evolutionary lineage from E. lindstroemi (Zhang 1998).However, this is in contrast to the pattern at the Uuga and Osmussaar cliffs and elsewhere in Estonia.Further study is needed to see if any other sections display a similar overlap of ranges of these species.Until then, in most sections, the short range of Y. protoramosus could be used as a stratigraphically well-constrained regional subzone.The E. lindstroemi Subzone in its type section is approximately 2 m thick, which is only slightly more than recorded here at the Uuga Cliff.
The succession in the Smedsby Gård core starts in the Furudal Limestone, corresponding to the P. serra and P. anserinus zones.These index species are common throughout their ranges.Only the upper part of the B. reclinatus Subzone is represented there and is 1.6 m thick.The full B. robustus Subzone in Smedsby Gård is about twice as thick as in NW Estonia.Bergström et al. (2011) also identified Y. protoramosus just above B. robustus, but included it in the E. lindstroemi Subzone.
In summary, our data confirm that the lower Darriwilian succession in NW Estonia is stratigraphically incomplete, with missing biozones and reduced thickness in the Kunda-Aseri interval.However, during the Pygodus serra Zone, remarkably uniform depositional conditions prevailed all over the Baltoscandian area, with only minor changes in sequence thickness and without major hiatuses.There the studied sections in NW Estonia can be directly compared with those elsewhere in Estonia and Sweden.

Updates to the integrated biostratigraphical scheme and comments on stage boundaries
The new high-resolution data from the Uuga and Osmussaar cliffs allow updates to the current correlation scheme between conodont and chitinozoan biozonations summarized by Nõlvak et al. (2006).The new scheme, as well as levels that remain uncertain at present, are shown in Fig. 10.
In the previous compilation (Nõlvak et al. 2006), the Y. foliaceus Subzone was considered to overlap with the Conochitina clavaherculi Subzone to some extent.According to our data, it seems more likely that B. reclinatus appeared slightly earlier than C. clavaherculi and consequently the Y. foliaceus Subzone must entirely fall within the Cyathochitina sebyensis Subzone.Second, the LAD of C. clavaherculi virtually coincides with the LAD of Y. protoramosus and FAD of E. lindstroemi in both sections studied, as well as in the Lasnamägi section in Tallinn (Männil 1986).Thus it would be more appropriate to correlate the base of the E. lindstroemi Subzone with the base of the Lower C. tuberculata Subzone (see discussion on the definition of the latter subzone in Tammekänd et al. 2010).Previously the Y. protoramosus Subzone was considered to correspond to the Lower C. tuberculata Subzone (Nõlvak et al. 2006).The Aseri, Lasnamägi and Uhaku stages were introduced based on lithological and macrofossil evidence, with type sections located in northern Estonia (Orviku 1940(Orviku , 1960;;Rõõmusoks 1970).Outside the stratotype area conodonts are nowadays among the most commonly used fossil groups for tracing boundaries of these stages in Baltoscandia (e.g., Bergström 1983;Zhang & Sturkell 1998;Zhang 1999;Löfgren & Zhang 2003).However, the exact correspondence between the original stage concepts and stratotypes and conodont (and chitinozoan) biozonation has remained poorly known.Thus at present the lower boundaries of the Aseri and Lasnamägi stages can only tentatively be correlated with the E. suecicus and P. serra zones (Fig. 10).
The lower boundary of the Uhaku Stage is drawn by the appearance of Gymnograptus linnarssoni and B. robustus (Männil & Meidla 1994).In the Tallinn area this boundary coincides with a discontinuity surface within a bed locally known as 'Raudsüda' (Einasto et al. 1996).As hitherto only a generalized faunal log of the Lasnamägi section is available to characterize this level (Männil 1986, fig. 2.1.1),our study provides valuable additional data.In both the Uuga and Osmussaar sections the lithological marker appears to coincide with the LAD of Belonechitina pellifera (Fig. 7).Confidently identified B. robustus appears 10-20 cm above this level, followed by G. linnarssoni 0.35 cm above the same discontinuity surface on Osmussaar.All these markers can be considered virtually coincident for regional correlations, providing multiple options for identifying the base of the Uhaku Stage in Baltoscandia.In northern Estonia, however, conodonts and chitinozoans appear to be more useful than the relatively rare graptolite faunas.Bergström et al. (2011) provided an interesting discussion on the geochronologic resolution of conodont biostratigraphy as well as on net accumulation rates in the Baltic Ordovician.Using time scales of Cooper & Sadler (2004), Webby et al. (2004) and Sadler et al. (2009), they estimated the duration of Uhaku conodont subzones to be on average less than 0.5 Ma.Following a similar approach, we focus here on the P. serra Zone.According to Cooper & Sadler (2004), the P. serra Zone corresponds to about 1.2 Ma.Webby et al. (2004), as well as the latest Geologic Time Scale (Cooper & Sadler 2012), infer about 2 Ma.Based on these time scales, it would at present be reasonable to assume that the temporal range of the P. serra Zone was most likely somewhere between 1.0 and 2.5 Ma.

Temporal resolution of Darriwilian biostratigraphy
The five subzones recognized in our study would therefore represent on average 0.2-0.5 Ma, which is close to the estimation by Bergström et al. (2011).Some of the individual subzones, especially the Y. protoramosus Subzone, are relatively thinner across Baltoscandia, and their temporal content is likely smaller.Moreover, when using well-constrained conodont zonal boundaries and adding the chitinozoan correlative levels, biostratigraphic resolution below 0.1 Ma could probably be achieved.This is similar to what has been demonstrated by Cramer et al. (2010) from the Llandovery-Wenlock boundary interval.Bergström et al. (2011) also estimated net accumulation rates for the late Darriwilian of Baltoscandia, suggesting extremely low values of 3-4 mm ka -1 .In the Uuga section the entire P. serra Zone is roughly 5.5 m thick, inferring a similar net accumulation rate of 2.2-5.5 mm ka -1 (based on the above mentioned duration of the P. serra Zone).Indeed, the deposition has most likely been episodic rather than continuous and single beds may have been formed rather quickly, as mentioned also by Bergström et al. (2011).The duration of periods of non-deposition remains mostly beyond biostratigraphic resolution, but at least some gaps can be traced by quantitative microfossil data, as shown by Tammekänd et al. (2010) and by comparing adjacent sections as done in this study.

CONCLUDING REMARKS
The data set from the Uuga and Osmussaar cliff sections is among the most detailed ones for Darriwilian conodont and chitinozoan distributions hitherto available.Altogether, more than 20 biostratigraphically important levels were recognized, probably approaching temporal resolution of 0.1 Ma in the upper Darriwilian.Some lithological marker horizons fit perfectly with the biostratigraphical data and indicate potential for future improvements in the temporal framework at least in northern Estonia.
The base of the Darriwilian coincides with the regional Volkhov-Kunda boundary in the study area, and the Aseri Stage is most probably missing in both the Uuga and Osmussaar sections.The lowermost part of the Lasnamägi Stage, corresponding to the Y. foliaceus Subzone, is also very condensed or missing in NE Estonia.The rest of the Lasnamägi and Uhaku stages is well dated by successive conodont subzones of the Pygodus serra Zone, which can be identified all over Estonia and Sweden.The zonal species P. serra and P. anserinus are too rare in most of the Estonian sections to be used for high-resolution biostratigraphy.
Our data and regional comparison suggest that Y. protoramosus is more common than previously thought, has a short range and is stratigraphically well-constrained in Estonia and Sweden.Thus, unlike Bergström et al. (2011), we consider it a valuable regional subzonal index.Our results also suggest that some previous identifications of the Y. foliaceus Subzone in Estonia are contradictory and this unit awaits further assessment.
The results point to the need for obtaining new highresolution data from the historical stratotype sections of the Aseri, Lasnamägi and Uhaku stages, to reconcile the original stage concepts with the current biostratigraphic resolution available in Baltoscandia.

Fig. 1 .
Fig. 1.Locality map, showing sampling sites on Osmussaar Island, the Pakri Peninsula and other sections, where Darriwilian conodonts have recently been studied and are compared here.
For example, Triangulodus alatus Dzik is characteristic of the lower part of the Väo Formation in both sections (Figs 2, 3).Ansella serrata Fåhraeus & Hunter is typical of the middle part of the Väo Formation.Due to relative rarity of this species, however, its range ends are at slightly different levels in the Uuga and Osmussaar sections.Another relatively rare species, Phragmodus polonicus Dzik, originally described from the Moijcza section of Poland (Dzik 1978), appears in the middle of the B. robustus Subzone (Uuga Cliff) or just above it (Osmussaar Cliff).In other Estonian sections P. polonicus has previously been identified only in the Kukruse Stage, in the basal part of the Amorphognathus tvaerensis Zone (Viira et al. 2006a).Conodont biostratigraphy in the Uuga and Osmussaar sections is summarized in Fig. 7.

Fig. 7 .
Fig. 7. Summary of conodont biostratigraphy in NW Estonian sections.Note that the precise temporal content of indicated gaps remains hypothetical.Regular time scale and correlation with regional stages are based on the latest international geological time scale (Cooper & Sadler 2012).GS, Global stage; P. a., Pygodus anserinus.
The formal subzone boundaries are defined by the appearance of Conochitina clavaherculi Eisenack, which marks the top of the Cyathochitina sebyensis Subzone and the base of the C. clavaherculi Subzone in the lower part of the Väo Formation.The disappearance of C. clavaherculi in the upper part of the Väo Formation defines the base of the succeeding Lower Conochitina tuberculata Subzone.Tammekänd et al. (2010) additionally distinguished several informal, but biostratigraphically useful levels at the Uuga Cliff.Most of these were also identified in the Osmussaar succession, with the exception of Baltochitina delicata Nõlvak (= Baltochitina sp.n. 1 in Tammekänd et al. 2010) and Desmochitina grandicolla Eisenack.

Fig. 8 .
Fig. 8. Correlation of conodont and chitinozoan bioevents between the Uuga and Osmussaar cliff sections.FAD, first appearance datum; LAD, last appearance datum; LAO, last abundant occurrence; FAO, first abundant occurrence.These notations are used in case of single specimens outside the main continuous range.See further explanation in the text.The LAD of Eoplacognathus pseudoplanus at Osmussaar is based on data of Mellgren et al. (2012).Baltoc, Baltochitina.

Fig. 9 .
Fig. 9. Scatter-plot graphical representation of the correlation between the Uuga and Osmussaar cliff sections.The line of correlation was fitted empirically.

Fig. 10 .
Fig.10.Updated correlation between conodont and chitinozoan biozonations and Baltic regional stages.Note that exact correspondence between the lower and upper boundaries of the Aseri Stage and conodont biozones awaits further high-resolution data from the type sections.GS, Global stage; Gr, graptolites; Pyg.anser., Pygodus anserinus.
. DS, discontinuity surface; ps., E. pseudoplanus Zone; fol., Y. foliaceus Subzone; protoram., Y. protoramosus Subzone; anser., P. anserinus Zone; kielce., S. kielcensis Subzone.Distribution of selected conodonts, chitinozoans and graptolites in the Osmussaar Cliff section.All species of Desmochitina and Rhabdochitina are omitted.For lithological legend see Fig. 2. The rock composition is shown after unpublished carbonate and insoluble residue analyses by Lembit Põlma from the 1970s.Due to the lack of a stratigraphically constrained dolomitic layer, distinction of the Pae Member is tentative.Protora., Y. protoramosus Subzone; Gym.cf.retiol., Gymnograptus cf.retioloides; Lasnamägi and early Uhaku times in the Estonian area.It is worth mentioning that only in the Ruhnu core, which probably represents the deepest shelf settings in the Estonian area, P. serra and P. anserinus were relatively common, just as they are in Swedish sections