Refined conodont stratigraphy at Martenberg (Rhenish Massif, Germany) as base for a formal middle/upper Frasnian substage boundary

The famous Martenberg section of the eastern Rhenish Massif, Germany, type-section of classical Frasnian goniatite and conodont zonations, has been restudied in order to document the microfacies development and to refine the conodont stratigraphy around the global semichatovae Event/Transgression, the proposed level to define a future upper Frasnian substage. More than 8.000 platform elements were identified and include new taxa. Palmatolepis jamieae is subdivided into the subspecies Pa. jamieae jamieae, Pa. jamieae savagei n. ssp., Pa. jamieae rosa n. ssp., and Pa. jamieae ssp. δ. Another new species, Pa. adorfensis n. sp., was previously partly identified as Pa. jamieae, while Pa. descendens n. sp. has previously been described in open nomenclature from Inner Mongolia. Morphotypes are defined in Icriodus symmetricus, Pa. ljaschenkoae, and Pa. proversa. A global literature survey shows that the eustatic semichatovae Event can be recognised in more than 20 regions of all continents with (sub)tropical Upper Devonian outcrops. At Martenberg, the transgression is preceded by a thin but distinctive interval with unconformities, microbial mats, sheet cracks, and currents that brought in the regionally youngest volcaniclastics. The new conodont data confirm that no typical Pa. jamieae (sensu the holotype) occur in the two beds originally supposed to represent the jamieae Zone in its reference section. We fully support the conclusion of Ovnatanova and Kononova (2020) that the jamieae Zone should be abandoned. Early Pa. jamieae subspecies and the related new taxa enter at Martenberg and in a few other regions in the globally easily recognisable Frasnian Zone 10 (= plana Zone). Frasnian Zone 11 (feisti Zone) is subdivided into subzones FZ 11a (= feisti Subzone) and FZ 11b (= nasuta Subzone). The base of the latter coincides with the semichatovae Transgression, the semichatovae Subzone of more shallow shelf settings, and is proposed to define in future the upper Frasnian substage base. On a global scale, the Martenberg section is currently the best bed-by-bed documented section for facies changes, conodont and goniatite biostratigraphy at the middle/upper Frasnian transition. Therefore, it is a prime candidate for a future GSSP selection. A global literature survey identified more than 20 other pelagic conodont successions that have the potential for precise correlation and a better understanding of the environmental changes associated with the semichatovae Event.


Introduction
The chronostratigraphic subdivision of the Devonian into series and stages was accomplished more than 20 years ago, although, some revisions are under way (see recent review by Becker et al. 2020a). For most of the stages, substage definitions have been proposed, but none of these have been ratified by the International Commission on Stratigraphy (ICS). Therefore, substages are widely used but are still informal. In the Frasnian, Ziegler and Sandberg (1997)  3. Development of a refined conodont zonation in pelagic facies, with subzones, and a focus on the disputed meaning and justification of a jamieae Zone (see Klapper and Becker 1999;Ovnatanova and Kononova 2020) 4. Proposal for a future, conodont-based upper Frasnian substage definition, including aspects of ammonoid bio-and sequence stratigraphy.

Locality
The Martenberg section lies northeast of Diemelsee-Adorf in the Waldeck region at the eastern margin of the Rhenish Massif, Germany (GPS: 51°22'30.4"N, 8°48'46.1"E). It can be reached by following, from the eastern end of Adorf, the road towards Giershagen. After ca. one kilometre, the Martenberg "Klippe" (Figs. 1 and 2) is situated west of the road in a small depression.
It is marked at a small parking area as a protected natural monument of the "Geopark Grenzwelten" (Mertmann 2017). The outcrop was originally part of the Christiane Mine, which started from an open pit (e.g. Bernauer 1890;Masling 1911;Teeke 1953;Emde 1965). From the 13 th century until 1963, it exploited a thick exhalative hematite ore body formed at the top of a basaltic volcanoe (Schlüter 1927;Paeckelmann 1928aPaeckelmann , 1936Bottke 1962Bottke , 1965, overlain by fossiliferous, condensed seamount limestones. The latter form the upper part of the Martenberg "Klippe" or "Rosenschlößchen" (Fig. 2), a monolith about 15 m in diameter and 10 m high (Sandberg et al. 1989a). The Martenberg section represents the axial dome of a second order anticline with smaller-scale special folding and a thrust zone (e.g. Schlüter 1927;Paeckelmann 1928a;Ree 1953;Becker 1984). It formed during the main Variscan orogeny as a subunit of the northeastwards branching, first order East Sauerland Anticline, due to the rheological resistance of the metabasalts in its core. The famous eastern face of the cliff (e.g. House and Ziegler 1977: fig. 1), which housed Ziegler´s lateral sections I, II, IV, and V, collapsed several years ago, which destroyed the formerly prominent, dense bed numbering. Fortunately, this did not affect the important section on the northern side (Figs. 2 and 3), which was the main section for Ziegler´s conodont sampling from 1958 to 1971, by Becker (1998, 1999), and for this study.
Despite its small outcrop size and thickness, the Martenberg section became a key locality for Rhenish biostratigraphy when Wedekind (1913) used it to develop his lower Upper Devonian Fig. 1 Geographic location of the Martenberg north of Diemelsee-Adorf on a simplified geological map for the differentiated eastern part of the Ostsauerland Anticline (extracted from Becker 1984). The star marks the location of the section goniatite zonation of the "Mantiococeras-Stufe", which later became the "Adorfstufe". Wedekind (1918) added some goniatites and Matern (1929) suggested some revisions. When the stratigraphic significance of conodonts was realised in Germany, Devonian research concentrated on condensed limestones with goniatites, such as the Martenberg (Müller 1956). Bischoff and Ziegler (1957) revised the Middle/Upper Devonian transition with the help of conodonts and in a footnote (p. 35) they pointed out that the Martenberg section was systematically examined for conodonts by W. Ziegler. His results were published one year later (Ziegler 1958). He emphasised that the eastern cliff side was not very suitable for detailed investigations due to the rapid wedging out of layers. Therefore, he restricted himself to the north-western side, followed later (Ziegler 1971) by a numbered section at the northern side, which is also the section re-studied by us. Kullmann and Ziegler (1970) aimed to complete gaps in goniatite stratigraphy and its correlation with the developing conodont stratigraphy. However, as it later became clear House and Kirchgasser 1993), the studies faced problems caused by the extreme condensation and local incompleteness of upper Givetian to lower middle Frasnian strata. House and Ziegler (1977) re-illustrated important goniatite type-material and were able to further refine the correlation between the goniatite and conodont succession. They re-established Wedekind´s (1913) distinction of do Iß and do Iγ faunas, which characterise the Frasnian zones UD I-I and UD I-J sensu . However, it is important to note that the new faunas with Stilleoceras retorquatum, Costamanticoceras nodulosum and Playfordites tripartitus (generic assignments updated) came from a higher level than Wedekind´s very thin Iß interval, Denckmann´s "Webel Limestone". Based on the presence of Beloceras, the original Iß correlates with UD I-H.  and Becker (2002) documented some important Martenberg goniatite types and new records (e.g. the Australian Trimanticoceras cinctum), followed most recently by tornoceratid descriptions in Korn (2021aKorn ( , 2021b. Sandberg et al. (1989a) revised the conodont zonation at the Middle/Upper Devonian transition and established new semichatovae Event interval between bed R-Q and Bed Q, characterised by thin-bedding (interval of "Sheet 1-3"). The ruler is folded out to a length of 30 cm. Bed P begins just above its top. (photos by T. Söte) Fig. 3 Proposed Frasnian substage position, ammonoid zonal keys after  and Wedekind (1913) (in brackets), conodont zone assignments of various authors, bed numbering, lithological log (scale 1 : 10), re-transgressive trends, position of depophases of Johnson et al. (1985), positions of new conodont samples, and microfacies data for Martenberg section NE. Between beds R-Q and Q lies the regressive interval, here exaggerated (scale 1 : 1) in order to show details, with previously neglected sheet cracks, unconformities, a microbial layer, and volcaniclastics (tuffites) "standard zones", which partly used Martenberg as a reference section. This was followed by the major revision of the Frasnian and lower Famennian zonation by Ziegler and Sandberg (1990). Unfortunately, it was partly incompatible with the more detailed Montagne Noire Frasnian zonation established one year before by Klapper (1989). The schemes use several different zonal index species but also different taxonomic concepts of the same species, which prevented a simple correlation. Therefore, Becker (1998, 1999) resampled the northern Martenberg section (which is "section q" of Ziegler and Sandberg 2000) in order to correlate the "Late Devonian Standard Conodont Zonation" and Montagne Noire Zonation based on data from the same section. This resulted in the recognition of problems at the middle/upper Frasnian transition, especially concerning the precise correlation between the Frasnian Zone 11 and jamieae/ Lower rhenana zones, leading to discussions by Sandberg and Ziegler (1998) and Ziegler and Sandberg (2000). Ovnatanova and Kononova (2020) re-examined the Martenberg conodont collection of Willi Ziegler in 1994 and denied the justification of the jamieae Zone, a topic that is re-evaluated here based on new data.
Frasnian "Standard" and Montagne Noire zonations Ziegler (1958Ziegler ( , 1962 studied the biostratigraphic distribution of Upper Devonian conodonts and established a succession of zones based on the genera Ancyrodella, Ancyrognathus, Palmatolepis, Polygnathus, Bispathodus, and Scaphignathus. His work built on earlier results of Sannemann (1955), Müller (1956), Bischoff (1956), and Bischoff andZiegler (1956, 1957). Following a revision of the middle/upper Famennian zonation (Ziegler and Sandberg 1984), the so-called "Standard Conodont Zonation" of the Frasnian and lower Famennian was refined and revised by Sandberg et al. (1989a) and Ziegler and Sandberg (1990). In the Frasnian, it is based on the supposed autochronologic phylogeny of Palmatolepis and its precursor Mesotaxis. Most Frasnian palmatolepids fall in Manticolepis Müller, 1956, a (sub)genus which rarely has been used by later authors (e.g. in Dzik 2002, who added the Frasnian genera Kielcelepis and Lagovilepis). We apply the traditional, wide generic concept but admit that a future dissolution of the "megagenus" Palmatolepis into well-defined, monophyletic (sub)genera is desirable.
Based on the studies of several conodont-rich sections in the Montagne Noire, Klapper (1989) showed that the regional Frasnian deposits can be divided into thirteen conodont zones instead of only seven zones of the "Standard Zonation" of Ziegler (1958Ziegler ( , 1962. Klapper´s Montagne Noire (MN) Zonation was based on iterative speciations in the genera Ancyrodella, Ozarkodina (s.l.), Ancyrognathus, and Palmatolepis. It considered the ranges of all species present in pelagic carbonate assemblages, if they were sufficiently short-ranging, as many show a consistent stratigraphic position in relation to Palmatolepis. He thus did not follow the "phylogenetic" (= autochronological) approach of Ziegler and Sandberg (1984), which was, anyway, not stringent since exceptions (e.g. "Uppermost marginifera Zone" defined by Scaphignathus velifer velifer) were allowed. Klapper (1989) intended the MN succession to be a regional zonation, which at that time had no claim to global application. This was underlined by the parallel introduction of a different regional zonation for the Canadian Rocky Mountains, published in the same Calgary Symposium volume by Klapper and Lane (1989). Klapper also referred to the preliminary results of graphic correlation, which could offer a finer resolution than conventional zoning and would thus probably replace it in the long term. From the data collected since the 1980s, Klapper et al. (1995) were able to present such an alternative to conventional zoning. Their "Frasnian Composite Standard" was based on the graphical correlation of 27 Frasnian sections from the Montagne Noire (France), western New York, and Western Australia. Values in CSUs (Composite Standard Units) provide either global or local (total) bases (= FADs/-FODs) and tops (= LADs/LODs) of species ranges, including marker taxa that define the bases of MN zones, some of which were used to fix the lines of correlation. The composite was later refined by data from the Timan-Pechora Basin in Russia . Important data for ranges of Frasnian conodonts were published subsequently from the Cantabrian Mountains (García-López and Sanz-López 2002), NW Australia (Klapper 2007), the Pyrenees (e.g. Liao and Valenzuela-Ríos 2012), the Anti-Atlas of southern Morocco (e.g. Hartenfels et al. 2013;Aboussalam and Becker 2016;Becker et al., 2018), Algerian Sahara (Mahboubi et al. 2015), eastern North America (Klapper and Kirchgasser 2016), the Rhenish Massif (e.g. Becker et al. 2016aBecker et al. , 2016c, the Iowa Basin (Day and Witzke 2017), South China (Zhang et al. 2019), and the Moroccan Meseta (Becker et al. 2020b;Aboussalam et al. 2020). Ranges of marker forms were combined in updated Frasnian composits (Klapper 1997;Klapper and Kirchgasser 2016). The latter study replaced the term "MN Zone" by "Frasnian Zone", but we prefer to name zones after their index species, because the names of fossils are easier to remember than simple numbers. Therefore, Frasnian Zone 10 is called plana Zone, Frasnian Zone 11 feisti Zone, and Frasnian Zone 12 winchelli Zone (see Becker et al. 2020a), but both are given in order to ease the stratigraphic understanding. Shallow-water Frasnian successions mostly lack the pelagic index species and require alternative regional zonations based mostly on polygnathids or icriodids (e.g. Matyja 1993;Kirilishina and Kononova 2004;Ovnatanova and Kononova 2008).
Due to the underlying disparate taxonomies, it was not possible to correlate the "Late Devonian Standard Conodont Zonation" and the Montagne Noire Zonation precisely until Becker (1998, 1999) resampled the Martenberg reference section and successfully applied the taxonomic concepts of the Montagne Noire zonation. As mentioned above (see Fig. 3), this raised problematical alignments of some of the "standard zones" by Ziegler and Sandberg (1990). This involved their original use of a generalised lithological log, which was corrected by Ziegler and Sandberg, 2000, fig. 2). More important was the uncertain correlation of the Frasnian Zone 11 found in Bed Q (= q, numbering of Ziegler 1971) either with the jamieae Zone, as suggested by the log of Ziegler and Sandberg, 1990, Fig. 3), or with the lower part of the "Early" rhenana Zone (see comments by Sandberg and Ziegler 1998). Ziegler and Sandberg (2000) restudied their samples (beds P, Q, R / samples VI'/6 to 12) from Martenberg and corrected their zonal identifications by re-assigning Bed Q (Sample VI/11b) to the "Early" rhenana Zone. In addition, they re-assigned the underlying upper part of Bed R (Sample VI'/10) to the jamieae Zone because of a supposed rare occurrence of Pa. jamieae. However, this did not clarify the situation and correlation problem, because Klapper and Becker (1999, p. 343) recognised that Pa. jamieae sensu Ziegler and Sandberg (1990) included different species (e.g. Pa. jamieae s.str. and Palmatolepis sp. B of Foster Jr. 1986, now Pa. feisti Klapper, 2007). Since the holotype of Pa. jamieae came from higher strata at Schmidt Quarry in the Kellerwald region, and in the absence of an illustration of the oldest specimens of Ziegler and Sandberg (2000), the entry level of Pa. jamieae s. str. at Martenberg remained unclear. As a consequence, the justification for a jamieae Zone in its type locality was completely open. Recently, Ovnatanova and Kononova (2020) correlated the regional Timan-Pechora "conodont associations III-XI" of Ovnatanova et al. (1999) with the "Late Devonian Standard Conodont Zonation" and Frasnian zonation. This involved the re-examination of the Frasnian conodont collection of the Rhenish Massif in the laboratory of Willi Ziegler in 1994. They did not recognise any true Pa. jamieae in the Lower rhenana Zone at Martenberg, or below, and concluded that the jamieae Zone should not be considered as a separate biostratigraphic unit. Since the alleged jamieae specimens from Ziegler and Sandberg (2000) have not been revised, their true identity remained unsolved, a circumstance that we decided to overcome by new samples, which were partly exceedingly rich in conodonts.
Global semichatovae Transgression/Event Sandberg et al. (1989bSandberg et al. ( , 1992Sandberg et al. ( , 2002 and Sandberg and Ziegler (1998) established the semichatovae Transgression as a shortterm, major eustatic deepening event. The expanded term semichatovae Event refers to the sudden sea-level rise combined with its effects on lithofacies and biota. It is recognisable as a lithological perturbation and bioevent, especially by migration pulses and radiations in several fossil groups, such as conodonts and ammonoids. The transgression and litho-/bioevent are named after Pa. semichatovae, a morphologically very distinctive species, which appeared as a cryptogenic taxon and spread fast pantropically, especially in shallower areas of carbonate platforms/ramps that were uninhabitable for most other palmatolepids. Sandberg et al. (1989b) pointed out that this opportunistic species may make up three-quarters of all Palmatolepis in neritic facies, while in deeper settings it only reached an abundance of 10% -or it is even rarer, as at Martenberg (Ziegler and Sandberg 2000) or other German seamount sections, such as Schmidt Quarry Ovnatanova and Kononova 2020). The semichatovae Transgression can be recognised pantropically (Fig. 4), but there is no published review of global distribution. Therefore, a comprehensive compilation is supplied here, roughly from west to east, giving the relevant regional lithological units, evidence for deepening/facies changes, some conodont information, and references for further reading (unclear localities are preceded by a question mark  (Streel et al. 1974;Sandberg et al. 1992;Boulvain et al. 1999;Da Silva et al. 2010 Meseta: Re-onset of condensed, pelagic limestones after a hiatus high in the plana Zone (Frasnian Zone 10), followed by continuing deepening marked by argillaceous, nodular limestones with Pa. semichatovae and a major change of water mass circulation evidenced by Nd isotopes (Lazreq 1999;Dopieralska et al. 2015;Becker et al. 2020b) (Fig. 4: Point 16). & Eastern Anti-Atlas, southern Morocco: Transgressive onset of Kellwasser-type black goniatite limestones above unconformities on the Tafilalt and Maïder platforms, again associated with a major change in Nd isotopes (e.g. Wendt and Belka 1991;Becker et al. , 2018Dopieralska et al. 2015;Wendt 2021) (Fig. 4: Point 17).
& Dra Valley, western Anti-Atlas, southern Morocco: Sudden change from well-oxygenated, red nodular limestones with some corals and FZ 10 conodonts to hypoxic goniatite shales with rich UD I-I faunas (Becker et al. 2004) (Fig. 4: Point 17). & Algerian Sahara: No distinctive semichatovae Transgression but a more continuous maximum of transgression within pelagic outer shelf facies (Mahboubi et al. 2019) (Fig. 4: Point 18). & Holy Cross Mountains, Poland: Change from conodontpoor proximal reef debris limestones to condensed pelagic limestones with rich deep-water conodont faunas (Matyja and Narkiewicz 1995); the "Lower gigas Regressive Pulse" of Narkiewicz (1989) probably represents a regional tectonic pattern ( Fig. 4 (Ma et al. 2004(Ma et al. , 2009Ma and Zong 2010). In a later review of South China Devonian sea level changes (Ma et al. 2017), the semichatovae Transgression is regionally named as "Qilijiang high stand" ( It is likely that the transgression is recognisable in further regions (e.g. Iran, Afghanistan, other parts of Central Asia), where biostratigraphic information is still too limited for precise international correlation.

Material and methods
Our reinvestigation at Martenberg concentrated in the northern section ("section q" of Ziegler and Sandberg 2000) on the critical interval that was established by the previous studies. It includes five samples for microfacies and eleven new samples for conodonts (R top = upper 7 cm of the bed, R-Q base, R-Q 8-14 cm below top, R-Q 4.5-8 cm below top, R-Q bulk = complete bed, Q base = 5-10 cm above base, Q 18-30 cm above base, "Sheet 1, 2, 3", P base = 0-8 cm above base), three of which (the "sheet samples") yielded no conodonts. Correlation of our samples ( Fig. 3) with that in Klapper and Becker (1999) and Ovnatanova and Kononova (2020) is straight forward since they derived from the same section. The correlation with samples from sections VI and VI´in Ziegler and Sandberg (1990) is based on the lateral tracing of beds in House and Ziegler (1977) and in Sandberg (1990, 2000). Since thicknesses change along strike, the position of their samples within beds has to be approximated by interpolation, which did not result in any contradictions (e.g. different zonal assignments).
A total of 8575 Pa (= P1) elements were counted and identified; ramiform elements were picked but not identified, as many Upper Devonian multielement reconstructions are incomplete or doubtful. We give the number of Nothognathella elements, which are thought to represent Pb (= P2) elements of Mesotaxis/Palmatolepis, showing their strong underrepresentation although they are massive elements that should not have undergone different hydraulic sorting than the Pa (= P1) elements.
For the recovery of conodonts, each sample was dissolved in 10% formic acid. Washed residues were separated into >0.1, >0.315, and >0.63 mm fractions. As it is widespread standard practice, only the voluminous smallest fractions were treated before picking by heavy liquid separation, using diluted sodium polytungstate (3Na 2 WO 4 × 9WO 3 × H 2 0). The light fraction was visually checked whether any significant amount of small-sized/juvenile conodonts finished accidently in that section, which was not the case. For each sample, the local alpha diversity was calcaluated, counting species and subspecies, not morphotypes.
The applied taxonomic concept distinguishes clearly between intraspecific morphotypes, subspecies, and species. Morphotypes are variants with distinctive morphological features that are part of populations but which intergrade. They are also used for forms with minor morphological differences when there is no clear evidence of different distributions in time, space, and biofacies in relation to the types of taxa. We recommend that the holotype/lectotype of a taxon should define Morphotype 1, although there have been divertions from this practice in the past. In subspecies, a few intermediate specimens may occur in relation to the nominate subspecies but both show different distributions in space, time, or biofacies, which suggests at least partial genetic separation. This definition is an essential part of present-day biological taxonomy and conservation biology. Subspecies are populations characterised by dominant factual genetical isolation and a key element towards speciation. In full species, there should be no intermediates between populations after speciation was complete; intermediates are restricted to the mostly short time intervals of the gradually evolving lineage, which is not preserved in cryptogenic taxa. We are aware that we can apply this concept currently only to the Pa elements. There could be species with common or very similar Pa but more different other elements. It is notable that subspecies have not been used in Frasnian palmatolepids while they are long established and very successfully applied in biostratigraphy in Famennian lineages of the genus. This difference has no logical base and led us to recognise Frasnian subspecies.
Our fauna agrees roughly with records of Klapper and Becker (1999), who took samples from the upper 7 cm and 10 cm, respectively, of Bed R. They found in addition Ag. leonis (= tsiensi auct.), Ag. amplicavus, Ag. barbus, Pa. proversa, and Pa. housei (= aff. proversa). The correlation with the data of Ziegler and Sandberg (1990) is not straight forward since their Table 1 combines records from their samples Ma VI'/11 and VI/ 12a; the latter corresponds to our Beds R plus R-Q. Their record of Ag. triangularis is doubtful due to their wide taxonomic concept of the species and since a representative was not illustrated. Ovnatanova and Kononova (2020) suggested that the record refers to Ag. iowaensis. They confirmed the presence of Ag. leonis (= tsiensi auct.) as well as Pa. proversa, and additionally recorded Pa. domanicensis and Pa. transitans. We doubt the latter identification, as for higher beds, since Pa. transitans normally does not range above the lower half of the middle Frasnian (Klapper 1997). In the absence of illustrations, we do not know which palmatolepid was locally identified as Pa. transitans but Pa. punctata martenbergensis is an option. In Pa. transitans, there may be an anterior platform sinus but no carina offset and no posterior platform sini. An important note in Ovnatanova and Kononova (2020) is the observation that, in Ziegler´s collection, Pa. plana occurs only in Sample MA VI´/10 (Bed R-Q), not in the two samples that represent Bed R. The combined alpha diversity (not counting morphotypes) for the upper part of Bed R is 18.
Bed R-Q was sampled with very high stratigraphic resolution due to the existing uncertainties concerning its precise age. It was divided into three segments: R-Q base, R-Q 8-14 cm below top, and R-Q 4.5-8 cm below top. The results of Sample R-Q bulk, taken initially, are presented as well and can be easily correlated with the samples of previous authors. The rich Sample R-Q base yielded a total of 28 taxa (2295 specimens), with a record gap for Pa. domanicensis. Compared to our preceding Sample R top, the following conodonts were additionally found: Ag. amplicavus  Figs. 8i,j). It is a matter of subjective judgement to recognise Pa. bohemica as full species or subspecies of Pa. punctata. We think that the differences of Pa. sepkoskii Bardashev and Bardasheva (2012), the slight angle of the fine posterior carina and slightly more sinuous posterior platform margin in the holotype, are not significant enough to warrant distinction beyond the morphotype level, the taxonomic concept applied here.
Sample R-Q 8-14 cm below top yielded only 19 taxa (599 specimens), with a record gap of nine further taxa, probably due to the reduced specimen number. Compared to Bed R-Q base, only Ag. guanxiensis, represented by a weakly ribbed variant (Fig. 9k), is a new addition. The rich Sample R-Q 4.5-8 cm below top yielded a total of 19 taxa (2905 specimens), with a local record gap for further seven taxa. New forms, compared to Sample R-Q 8-14 cm below top, are Pa. hassi   Branson and Mehl, 1934 The moderately rich Sample R-Q bulk yielded, combined with previous records, 24 taxa (574 specimens), with a record lack for further two species known from the subunits. Compared to the other samples from the bed, Pa. domanicensis (Fig. 11j), Pa. cf. domanicensis, with a rather narrow, lappet-like side lobe (Fig. 12f), and Icriodus sp. 1 with an offset denticle on the cusp (Fig. 12e) are additions. Unusually, the bulk sample did not contain Pa. feisti, which may mean that it did not include much material from the upper part of the bed. Ziegler and Sandberg (2000) recorded Pa. jamieae from the "upper part of Bed r", which probably is equivalent to our Bed R-Q. Since no specimen was illustrated, and with respect to the wide taxonomic concept applied in Ziegler and Sandberg (1990), which included, e.g. Pa. feisti specimens (see Klapper 2007), this record remains doubtful; such records are quoted in the following as "Pa. jamieae auct.".
As noted above, the thin volcaniclastics and tuffite layers between beds R-Q and Q yielded no conodonts at all. Sample sizes are normally smaller than in the other cases, but the complete lack of conodonts, even in micritic layers with subordinate volcanic clasts, is surprising. Bed Q was previously divided into several segments Klapper and Becker 1999). Aiming at high stratigraphic resolution, we separated a sample from the base (Q base, correlating with samples VI´/9 and VI/11b of Ziegler and  and from the middle part (Q 18-30 above base), which includes roughly Sample VI´/8 of Ziegler and Sandberg (1990: 15-24 cm above base at a laterally reduced thickness of only 33 cm) and Sample Q 20-25 above base of Klapper and Becker (1999); the first and third are combined here (Table 1). For the upper part of the bed, there are Sample Q 3-13 cm below top of Klapper and Becker (1999) and the samples of Ziegler and Sandberg (1990: VI´/6, 7 = upper 13 cm and IV/1b, which was shown at ca. the level of VI/7).
For the upper part of Bed Q, we summarise data from the literature, which include Samples VI´/6, 7 (upper 9 of 33 cm), parts of Sample VI/11a, Sample IV/1b of Ziegler and 17, with additional record gaps for two species, Pa. kireevae and Po. lodinensis. Notable are the youngest local records of Ad. curvata (Early Form) and Pa. ederi.
Bed P was divided by Klapper and Becker (1999) into two segments (0-8 cm and 8-18 cm). We re-sampled the lower part (Sample P base) and confirmed in comparison to the combined evidence of all previous papers the FODs of Pa. winchelli (Fig.  12n), Pa. brevis (Fig. 12p), Pa. rhenana (Fig. 12m), Pa. jamieae jamieae (Fig. 12o), Ad. ioides (M1, Fig. 12t), and Ad. ioides (s.str.). There is a single Pa. jamieae rosa n. ssp., three Pa. jamieae savagei n. ssp. (M2, Figs. 12q-s), and Pa. jamieae jamieae is represented by only one specimen. As in older beds, Pa. hassi is dominant. In Samples VI´/5 and VI/10e of Ziegler and Sandberg (1990) from the lower part of Bed P, the locally youngest Pa. proversa were recorded, which meant an upper range extension in relation to the composite range of Klapper et al. (1996). We have reservations concerning Bed P records in Ziegler and Sandberg (1990) of Pa. punctata and Pa. plana.

Stratigraphic interpretation and diversity trends of Martenberg succession
No zonal marker was found in Sample R top by our re-sampling, but Ziegler and Sandberg (1990), confirmed by Ovnatanova and Kononova (2020), found Pa. proversa, the index fossil of Frasnian Zone 9 (proversa Zone) in Bed R and much further below. Ziegler and Sandberg (1990) recorded in addition Pa. plana, index fossil of Frasnian Zone 10 (plana Zone), but this was rejected for the detailed succession of Section VI´by the revision of samples by Ovnatanova and Kononova (2020, p. 116). Typical Pa. amplificata enter, after Klapper et al. (1996), in the higher part of Frasnian Zone 10 while Bardashev (2009) gave a lower range for the closely related Pa. manzuri. We collected specimens that are intermediate between the types of the two species, which are, therefore, in accord with the published ages. Ovnatanova and Kononova (2020) showed, for the Timan, a joint entry of Pa. amplificata with Pa. proversa. Ancyrognathus amplicavus enters already at the base of Frasnian Zone 7 , Ag. iowaensis at the base of Frasnian Zone 9 (Klapper and Kirchgasser 2016), Pa. ljaschenkoae near the top of Frasnian Zone 8 (Klapper 1997), and Pa. mucronata in the upper part of Frasnian Zone 9. The cooccurrence of Ag. barbus and Pa. mucronata found by Klapper and Becker (1999) meant an upper range extension for the first taxon; previously (Klapper 1997), Ag. barbus was thought to range only into the middle of Frasnian Zone 8. In  summary, it seems that the top of Bed R is not younger than the upper Frasnian Zone 9 (proversa Zone).
Due to the first occurrence of Pa. plana (Figs. 6m, n) in our re-sampling, Sample R-Q base can be assigned to Frasnian Zone 10 (plana Zone). Previously, Bed R-Q has not been studied with our stratigraphic precision. Palmatolepis plana is found in all three segments (R-Q base, R-Q 8-14 cm below top, R-Q 4.5-8 cm below top) as well as in Sample R-Q bulk. Therefore, the boundary between Frasnian zones 9/10 is placed at the base of Bed R-Q. In addition, Sample R-Q base includes Pa. jamieae savagei n. ssp. (M1) (Figs. 8d, e), Pa. adorfensis n. sp. (Figs. 7b, r, s, 9i, j), and Pa. jamieae rosa n.ssp. (Fig. 8a), which become accessory Frasnian Zone 10 indicators. "Palmatolepis jamieae" specimens of Ziegler and Sandberg (2000) were probably based on one or some of these taxa. This provides a correlation of the base of the Frasnian Zone 10 (plana Zone) with the jamieae Zone sensu Ziegler and Sandberg (2000), but all early forms resembling, to some extent, Pa. jamieae are too rare to be used as zonal index taxa. The strong increase of alpha diversity in Frasnian Zone 10 (plana Zone) is probably an artefact of the strong increase of available specimens; most newcomers occur in small numbers.
Unexpected was the early entry of Ag. triangularis (Figs. 6c, e, f), which has not been verified in Frasnian Zone 10 (plana Zone) until now (see composite ranges in Klapper 1997 andKirchgasser 2016). The diagnostic free blade of these early representatives is partly preserved (Fig.  6c). The basal pit becomes relatively smaller with growth. As a consequence, Ag. triangularis loses its status as auxiliary Frasnian Zone 11 (feisti Zone) indicator, but becomes important to recognise levels from Frasnian Zone 10 (plana Zone) upwards. This has implications for the precise dating of reef extinctions and the separation of the two deepening pulses in the middle/upper Frasnian transition.
The single Ag. guangxiensis from Sample R-Q 8-14 below top is of some stratigraphic value. The species was originally described from the middle of the Lower rhenana Zone of Sihongshan, Guangxi (Wang and Ziegler 1983), but a slightly older, weakly ribbed representative was found above the entries of Pa. plana and Ag. triangularis at Longmen (Wang 1994). This gives a perfect match with our weakly ribbed Martenberg specimen from the higher Frasnian 10 Zone (plana Zone), which at the same time means a new record for Europe. The Montagne Noire record of Ag. guangxiensis (Klapper 1989) belongs to the related Ag. barbus (see Sandberg et al. 1992, p. 52, andKlapper 1997, p. 124). In comparison to the base of the zone, the alpha diversity stagnated when the sample-size related record gaps are taken into consideration.
Because of Pa. nasuta and rare Pa. semichatovae found by Ziegler and Sandberg (2000), the lower part of Bed Q can be assigned to a distinctive subdivision of the Frasnian Zone 11, the new Frasnian Subzone 11b or nasuta Subzone. The base correlates straight with the base of the Lower rhenana Zone sensu Ziegler and Sandberg (1990) and with the base of the semichatovae Subzone sensu Morrow and Sandberg (2008) that is typical for more shallow-water successions. The entry of Po. lodinensis adds to the subzone distinction; the species enters in the composite range of Klapper et al. (1996) slightly above the base of Frasnian Zone 11. The record of Po. uchtensis at this level is somewhat unusual since the species does not overlap with Pa. semichatovae in its Timan type region (Ovnatanova and Kononova 2020, Fig. 3). Palmatolepis semichatovae is very rare in the Rhenish Massif. Apart from the occurrence at Martenberg and from four specimens from the Schmidt Quarry (Ziegler and Sandberg 1990, Bed 21), the only other record is a single specimen from the Hölloch Valley of the Lahn Syncline (Gereke 2007, p. 58).
The local decline, not necessarily their final extinction, of some previously characteristic species seems typical for the transition from Bed R-Q to Bed Q: Ag. coeni, Ag. leonis, Pa. amplificata, Pa. domanicensis, and Pa. kireevae. Frasnian Subzones 11a and 11b are physically separated at Martenberg by an interval with unconformities and current-induced sedimentation immediately preceding the semichatovae Transgression (see microfacies analyses below). This sedimentary break did not result in a reduction of total alpha diversity. The  disappearance of taxa was balanced by newcomers; the semichatovae Event is locally characterised by faunal turnover, with continuing high levels of episodic record gaps.
The middle and upper parts of Bed Q represent the middle and upper Frasnian Subzone 11b (nasuta Subzone). The local entry of Ad. curvata Late Form (new records) is interesting but the morphotype has a lower range, starting at the base of the Frasnian Zone 10 (plana Zone) in the Frasnian composite (Klapper and Kirchgasser 2016). The entry of Palmatolepis uyenoi recorded by Klapper and Becker (1999) seems suitable to characterise a higher level of the Frasnian 11b Subzone (see composite range in Klapper 2007, p. 529: CSU 113.1-124.4). The restriction of Pa. ederi to Frasnian Subzone 11b (nasuta Subzone) at Martenberg agrees with the upper composite range of Klapper et al. (1996) but Ziegler and Sandberg (1990) reported younger specimens from other sections. Following the original illustrations of Ziegler and Sandberg (1990) (compare synonymy of Klapper 1997), revisions of Ovnatanova and Kononova (2020), and our re-sampling, there are no Pa. jamieae s.str. (= jamieae jamieae) in the Martenberg Frasnian Subzone 11b (nasuta Zone). At Schmidt Quarry in the Kellerwald region, the type-level of Pa. jamieae jamieae coincides with the local FOD of Pa. nasuta   Table 2, Sample 84-GER-1 = Bed 23), but associated Ad. ioides indicate a level above Frasnian Zone 11, which is supported by the position of the Pa. semichatovae level, indicating Frasnian Subzone 11b (nasuta Subzone) well below (Bed 21). However, there may be Frasnian Zone 11 records of Pa. jamieae jamieae from other regions (see stratigraphic range in taxonomic chapter).
Because of the first occurrences of Pa. winchelli and Ad. ioides, the lower part of Bed P falls in the basal Frasnian Zone 12 (winchelli Zone). Based on the revised Martenberg record of Ovnatanova and Kononova (2020, p. 117), on our two new specimens, and because of the association of the holotype with Ad. ioides at Schmidt Quarry, this is the type and main level of Pa. jamieae jamieae. We confirm the upper range extension of Ovnatanova and Kononova (2020) for Pa. brevis, which Ziegler and Sandberg (1990) regarded as typical for the Lower rhenana Zone. The local downwards range extension of Pa. rhenana (= rhenana rhenana) by Ziegler and Sandberg (2000) and Ovnatanova and Kononova (2020, p. 117) is also confirmed by our new sample from the base of Bed P. It means that the index species of the Upper rhenana Zone enters near the base of Frasnian 12 Zone (winchelli Zone). In the Frasnian composite of Klapper (1997), its FAD is at CSU 124.3, higher in the Frasnian Zone 12, and only slightly below the Lower Kellwasser level (compare FOD at Steinbruch Schmidt, Ziegler and . We prefer Pa. winchelli (= subrecta) as zonal index species but are pleased about the improved precision in the correlation of Frasnian Zonation and "Standard Zones".
The literature data (see Table 1) suggest a decline of alpha diversity in Frasnian Subzone 11b (nasuta Subzone) and Frasnian Zone 12 (winchelli Zone), down to 19-23 taxa (from 26-28, local Lazarus Taxa included). The proven alpha diversity of only 16 recorded taxa in the lower part of Bed P probably reflects the much smaller number of recovered specimens (365, Revised diagnosis: Narrow platform, which can be straight (M1) to slightly curved (M2), with 5-9 transverse rows of denticles from median stages on; posterior half of the median row distinctly higher than lateral rows; it may extend posteriorly by two to four denticles beyond the lateral rows; denticles of the median row somewhat laterally compressed, connected by a thin, longitudinal ridge; denticles of lateral rows pointed, subcircular and isolated; narrow part of the basal cavity extents posteriorly to a subcircular outline. Discussion: Two main morphotypes are distinguished based on the curvature of the middle and posterior part of the platform, excluding the anterior apex. Morphotype 1 is more or less straight, with a curvature < 10°, Morphotype 2 is markedly curved (> 10°). Some variance can be observed in the two morphotypes.
There are median to adult stages with five to six transverse rows and posteriorly two or more denticles beyond the lateral rows and median to adult stages with seven to nine transverse rows and posteriorly normally one or two (Fig. 5h), rarely up to four small denticles (Figs. 5i, o) beyond the lateral rows in Morphotype 1.
In Morphotype 2, median to adult stages with five to six transverse rows and posteriorly three to four denticles beyond the lateral rows and median to adult stages with seven to nine transverse rows and posteriorly one to three denticles beyond the lateral rows (Figs. 5p, 8b, c) can be observed.
Morphotype 1 corresponds to the lectotype of I. symmetricus and Morphotype 2 to the holotype of I. curvatus. Morphotypes are assigned in order to provide a base to establish in future possible different distributions in time, space, and facies (possible ecomorphotype patterns). Ziegler (1975) mentioned that originally the asymmetrical outline of the basal cavity was thought to be the characteristic of I. curvatus, but that it is encompassed within the variability range of I. symmetricus. In his opinion, if I. curvatus was to be treated separately, this should be done on the basis of its long middle row that extends posteriorly three to four denticles beyond the lateral rows. However, he also mentioned that this feature has not been found in later described curved specimens. Stratigraphic and geographic range: The species ranges throughout the Frasnian and can be found pantropically. Narkiewicz and Bultynck (2010, p. 609 Description: The platform is curved (ca. 20°) with six transverse rows. The posterior half of the median row is distinctly higher than the lateral rows. The denticles of the median row are laterally compressed. They continue posteriorly as a ridge with three denticles, the second of which is developed as a short transverse ridge. The denticles of the lateral rows are pointed. Prominent is an offset, large denticle on the right side of the widened cusp at the level of the most posterior transverse row, where the normal right side denticle is missing. Discussion: Our single specimen is kept in open nomenclature since we cannot exclude that the irregular denticulation is pathological. Dzik (2006) published similar but more straight specimens from the upper Famennian styriacus Zone of the Holy Cross Mountains as "Latericriodus (or Antognathus) rarus". We do not think that our specimen is conspecific. Two different icriodids with marginal ridges and single small nodes of the upper cusp were illustrated by Lys and Serre (1957, pl. IX) from the Frasnian of the Montagne Noire as I. cf. nodosus and Icriodus sp. Wang (1994) illustrated as I. symmetricus from the lower Frasnian of the Sihongshan section, Guangxi, a straight specimen with only five rows of denticles, with the median row merged to a complete longitudinal ridge, and with an additional prominent denticle on the right side of the upper cusp.   Material: One specimen from Sample P base (Fig. 12o).
Revised diagnosis of Klapper, 2007: Outline of platform roughly triangular to pyriform; outer lobe relatively narrow, midline of lobe directed laterally; outer posterior sinus commonly distinct; blade-carina sigmoidal but straight anteriorly, curving towards lobe just anterior of central node; posterior carina short, reaching or almost reaching posterior tip, curved inwardly; rim of marginal nodes on inner side of platform; blade extends anterior of platform a short distance. Discussion: All three specimens illustrated by Ziegler and Sandberg (1990) from the same bed at Schmidt Quarry are typical Pa. jamieae jamieae, which suggests a restricted variability in the type population, which is stratigraphically younger than specimens from the supposed "jamieae Zone" at Martenberg. At Martenberg, we found one specimen with a rather narrow platform in strata corresponding to the type level but the typical subspecies is more common in equivalent beds at Beringhauser Tunnel (Saupe and Becker, in prep.). A typical feature of Pa. jamieae jamieae is the concave to near-straight anterior inner platform marked by a rim of raised nodes. Following the refined diagnosis for Pa. jamieae of Klapper (2007) and based on the survey of all specimens illustrated in the literature and our re-sampling at Martenberg, three subspecies (Pa. jamieae jamieae, Pa. jamieae savagei n. ssp., Pa. jamieae rosa n. ssp.) and two related new species (Pa. adorfensis n. sp., Pa. descendens n. sp.) are named. Their distinction is useful to establish refined stratigraphical ranges. Additionally, two other Fig. 13 Microfacies of sampled beds a Top of Bed R: flaser-bedded, bioturbated bioclastic wackestone with micritic to microsparitic matrix, isolated styliolines, stylioline nests (arrows), ostracods (O), and redbrownish iron enrichments. a1 Detail of the top of Bed R, showing a stylioline nest (arrow). a2 Detail of the top of Bed R, showing a lump of small, Frutexites-type microstromatolites surrounding small stylioline shells (arrows). b Bed R-Q base: flaser-bedded, bioturbated bioclastic wackestone with micritic to microsparitic, partly nodular matrix and two levels of fine-grained, angular to subangular, partly ironmineralized extraclast enrichments (crinoid debris/volcaniclasts) at the base and top. Note that a very thin sparite seam follows the sharp base of the upper level on the right side but not in the middle, where the underlying fine micrite is truncated by extraclast packstone. b1 Detail of the base of Bed R-Q, showing fine extraclast enrichments above the iron-stained basal surface (arrow). b2 Detail of the base of Bed R-Q, showing the sharp base of the upper extraclast interval. c "Sheet 1": flaser-bedded, chloritized and carbonatic tuff (volcaniclastic packstone) with diagenetic iron and clay mineralizations. c1 Detail of "Sheet 1", showing the chloritization of rather well-sorted angular clasts, as typical for recrystallized tuff. d "Sheet 2": very fine-grained, recrystallized (chloritic), calcareous tuffite at the base (1 Layer 1), followed above an undulating unconformity and a very thin, sparitic sheet crack (2 Layer 2) by a grey band of microbialite with clotted micrite fabric (3 Layer 3), in the upper part with embedded, partly coated, subrounded to angular clasts derived from the basal layer plus irregular and chloritized microlapilli, then, above a rippled unconformity by a thicker sheet crack with floating, cavernous, chloritic volcaniclasts in the lower part of Layer 4 (4), becoming increasingly abundant towards the top (5 Layer 5). d1 Details of "Sheet 2", showing the fine, clastic nature of Layer 1 at the base (1), the thin sparite sheet of Layer 2 (2), diffuse micrite clotting and tube-like meandering internal structures (T) in the microbial Layer 3, with chloritized volcaniclasts (V) at the rippled (R) top, and partly ironcoated, size-sorted extraclasts at the base of Layer 4 (4). e "Sheet 3": flaser-bedded, unfossiliferous, reddish, micritic to microsparitic mudstone with some isolated, chloritized, lumped microlapilli (M) within the middle part.   fig. 6.20). The position of the central node projects laterally to the end of the platform lobe and the nodose anterior inner platform margin is slightly concave to straight, as in the typical subspecies. Another form, provisionally called Palmatolepis sp. ε, has a small, nodose platform, lacks an outer posterior sinus, and displays a well-developed, coarse posterior carina, which turns straight inwards without bending (e.g. Ovnatanova et al. 2017: p. 1084. A weak development of the posterior outer sinus may also occur in some Pa. jamieae jamieae (one of the syntypes, Ziegler and Sandberg 1990, pl. 6, Fig. 9) and in Pa. jamieae savagei n. ssp. (e.g. Ovnatanova and Kononova 2008, pl. 11, Fig. 7, pl. 14, Fig. 10). Stratigraphic and geographic range: Pa. jamieae jamieae is a rare (< 1 % of palmatolepids) to moderately common (1-5 % of palmatolepids) form in populations from its type locality, possibly from Frasnian Subzone 11b (Bed 21 with Pa. semichatovae, and Bed 22, Ziegler and , no specimen figured), from Frasnian Zone 12 (Bed 23 with Ad. ioides, type level, to Bed 26/5 with the FOD of Ag. asymmetricus, Lower Kellwasser Limestone), and possibly from the lower part of Frasnian Zone 13a (Beds 27/6 and 1/7, no specimen figured). According to our partly very rich samples, it is absent from the beds originally  or later (Ziegler and Sandberg 2000) assigned to the jamieae Zone at Martenberg (Frasnian Zone 10 and Frasnian Subzone 11a), which confirms the revision of Ovnatanova and Kononova (2020). Records backed by illustrations from other regions are restricted to Frasnian Zone 12 (winchelli Zone) of the Ardennes (Tiènne du Lion backmound section, Belgium, Sample 85-BEL-120, Sandberg et al. 1992), the subsurface of Pomerania (Matyja 1993), the subsurface of the Holy Cross Mountains (Malec 2015), the Timan of northern  fig. 1  marginal nodes or short ridges on inner side of platform usually present; blade extents anterior of platform a short distance. Discussion: The new subspecies differs from Pa. jamieae jamieae in its rather narrow, more elongate platform, lacking a bulbous posterior outer platform around the central node, and, in typical forms, by the more convex, arched inner platform margin. The posterior sinus is moderately (in the holotype, Fig. 8e) to weakly developed (e.g. in paratype GMM B9A.13-103, Fig. 10g), and missing in extreme variants from northern Russia (Ovnatanova and Kononova 2008, pl. 11, Fig. 7, pl. 14, Fig. 10) and Inner Mongolia (Lang and Wang 2010, pl. 1, Fig. 12). Palmatolepis jamieae ssp. δ differs in its smooth inner platform (e.g. Ovnatanova and Kononova 2008, pl. 10, figs. 16-18, pl. 11, Fig. 3) and a straight to slightly concave anterior inner platform margin, as in Pa. jamieae jamieae. We distinguish two morphotypes of Pa. jamieae savagei n. ssp. based on the position of the central node relative to the lobe. In Morphotype 1, which includes the holotype, the central node is located at the level of the apex of the lobe (Figs. 8d, e, 10g, i), while in Morphotype 2 (Figs. 12q, s; e.g. Ji and Ziegler 1993, pl. 27, Fig. 1; Wang 1994, pl. 6, Fig. 14;Izokh et al. 2004, pl . 1, Fig. 12), its position projects towards the posterior end of the lobe, as in typical Pa. jamieae jamieae. In median-sized Morphotype 1 specimens (Figs. 10g-i), the posterior carina is short to almost absent. One atypical Martenberg specimen, assigned questionably to Morphotype 2 (Fig. 10j), combines a moderately wide platform and rimmed, concave inner platform margin resembling Pa. jamieae jamieae but differs from the typical subspecies in its completely reduced posterior carina, while another one (Fig.  12r) is intermediate between Morphotype 1 and Morphotype 2 due to the position of the central node in relation to the platform.
Savage (2013,2019) illustrated from much younger levels (upper part of Frasnian Zone 13a) of Mae Sariang, NW Thailand, an elongated specimen as Pa. aff. jamieae that resembles Morphotype 1. However, its lobe and central node sit in a very posterior position, as in other specimens assigned by him to Pa. khaensis. A typical feature of the latter is its long, straight carina anterior of the central node. In addition, we have the impression that the platform shapes of some Mae Sariang specimens are affected by distortion. A supposed jamieae specimen illustrated by Komatsu et al. (2018) from northern Vietnam is also significantly younger than typical Pa. jamieae savagei n. ssp. and differs by its smooth platform and a long, fine, posterior carina. The latter feature excludes it from Pa. jamieae ssp. δ. Stratigraphic and geographic range: Palmatolepis jamieae savagei n. ssp. has a different, lower range than typical Pa. jamieae but its upper range overlaps with the typical subspecies. In the Rhenish Massif, Morphotype 1 was found from Frasnian Zone 10 (plana Zone) to Frasnian Subzone 11a Fig. 14 Correlation of the conodont succession (sequence of important marker taxa), CSU units (Klapper et al. 1995;Klapper et al. 1996;Klapper and Kirchgasser 2016), Frasnian zones (FZ), and "standard zones" sensu Ziegler and Sandberg (1990) around the semichatovae Event and middle/upper Frasnian transition at Martenberg (feisti Subzone), Morphotype 2 in Frasnian Subzone 11a (records in Stichling et al., this vol.) and in the basal part of Frasnian Zone 12 (winchelli Zone). Elsewhere, Morphotype 1 occurs in the Frasnian Zone 10/ Frasnian Subzone 11a interval of the Moroccan Meseta (Lazreq 1992), Frasnian Subzone 11b (nasuta Subzone) to Frasnian Zone 12 (winchelli Zone) in South China (Ji 1993), at a similar level (regional Mendym Stage) in the Lemva River Basin of the Polar Urals , in the Volgograd region of the Russian Platform (Galushin and Kononova, 2004), and in the Rudnyi Altai (Izokh et al. 2004). The age of Morphotype 1 in Inner Mongolia (Heilongjiang Province, Niqiuhe Formation and Daminshan Formation, Lang and Wang 2010) cannot be clearly determined in a faunal association that represents the undivided Frasnian Zones 11-12 (feisti to winchelli Zone) interval. Morphotype 2 was found in South China in Frasnian Subzone 11b (nasuta Subzone, Sihongshan section, Sample CDC 372, Wang 1994Wang , 2016  Derivation of name: After the type locality, which in local folklore is referred to as "Rosenschlösschen", meaning small rose castle; rosa is Latin for rose. Material: Holotype GMM B9A.13-89 (Fig. 9l), paratype GMM B9A.13-55 (Fig. 8a), and a third specimen from Sample P base. Six further specimens have previously been illustrated from three distant regions. Type locality and level: Martenberg, Sample R-Q 8-14 cm below top, Frasnian Zone 10 (plana Zone), upper Frasnian. Diagnosis: Outline of platform asymmetrically rhombic to slightly pyriform; broad, indistinctively delimited lobe with apex lying close to the position of the central node, bordered by near-straight anterior and posterior margins with weak to absent sini; carina sigmoidal, bends anterior of central node towards lobe side; posterior carina often thin, rather straight, ends before tip of platform; at maturity with rim of marginal nodes or short ridges on strongly convex inner side of platform; blade extents anterior of platform a short distance. Discussion: The outline of the platform of the new subspecies lies between the pyriform shape of Pa. jamieae jamieae and the more broadly rhombic shape of Pa. feisti. The shape of the lobe is similar to Pa. feisti but is located in a more posterior position and the carina is more sigmoidal than in typical Pa. feisti. The arched inner platform margin differs from that in Pa. jamiae jamieae and agrees with that in the much narrower Pa. jamieae savagei n. ssp., which, again, differs clearly in its long, markedly concave anterior outer platform margin (compare Fig. 8e). In Chinese specimens, the platform is rather smooth, partly because of their small size, which indicates a transition towards Pa. jamieae ssp. δ that, however, also displays a long and strong anterior outer platform sinus. Derivation of name: After the region of the type locality at Diemelsee-Adorf. Material: Holotype GMM B9A.13-53 (Fig. 7r), paratypes GMM B9A.13-54 (Fig. 7s), GMM B9A.13-86 (Fig. 9i), GMM B9A.13-87 (Fig. 9j), and the slightly less typical GMM B9A.13-38 (Fig. 7b). Additional specimens occur in the Winsenberg section near Adorf (Becker et al. 2016c) and in the Hönne Valley region (Stichling et al., this vol.). Five more specimens have previously been illustrated from four distant regions. Type locality and level: Martenberg, Sample R-Q 8-14 cm below top, Frasnian Zone 10 (plana Zone), upper Frasnian. Diagnosis: Outline of platform roughly subtriangular, wide in the centre and just anterior of central node; broad, well-rounded, lappet-like lobe ending in a position lateral to the central node; anterior and posterior sinus of lobe distinct; posterior outer platform margin straight or convexly arched; carina weakly sigmoidal, bends anterior of central node towards lobe; posterior carina fine or partly reduced, turns straight inwards with little to no bending; rim of marginal nodes or short ridges on inner side of platform; blade extents anterior of platform a short distance.
Discussion: There is no morphological intergradation with relatives of Pa. jamieae, which justifies the introduction of a new species. It differs from all subspecies of Pa. jamieae in its distinctive, well-defined, subsymmetric, broadly rounded and more anteriorly situated lobe bordered by marked anterior and posterior sini. In Pa. feisti, there are two weak posterior sini, instead of one pronounced one while Pa. plana is characterised by a strongly asymmetric side lobe with very pronounced to almost rectangular anterior sinus. Palmatolepis amplificata and the holotype of Pa. manzuri share a lappet-like side lobe but it is even wider and the posterior platform is distinctively longer, pointed and constricted at the end. In Pa. unicornis, the side lobe has a similar shape, but the platform bears very strong nodes and the free blade consists of a single large denticle. The specimen of Lazreq (1992)  Derivation of name: After the posterior direction of the lobe, from Latin descendens = to step down. Material: Holotype GMM B9A.13-107 (Fig. 10k), paratypes GMM B9A.13-147 (Fig. 12g), GMM B9A.13-148 (Fig. 12h).
A fourth specimen has been illustrated from the far distant region of Inner Mongolia (Lang and Wang 2010). Type locality and level: Martenberg, Sample Q 18-30 cm above base, Frasnian Subzone 11b (nasuta Subzone), upper Frasnian. Diagnosis: Outline of platform roughly subtriangular to pyriform; well-rounded, lappet-like, asymmetric, short lobe oriented posteriorly and ending slightly after the central node; anterior outer margin descending with a shallow sinus and with pronounced, subrectangular posterior sinus followed by an arched posterior margin; carina weakly to moderately sigmoidal, bends anterior of central node towards lobe; posterior carina pronounced or fine, turns straight inwards with little to no bending; rim of marginal nodes or short ridges on inner side of platform. Discussion: Lang and Wang (2010)  Palmatolepis ljaschenkoae Ovnatanova, 1976 (M1: Figs. 6d, g, 9r, 11m;M2: Figs. 5c, 8t, u, 11n;M3: Figs. 6j, 9d, 10a, b, 11o) Revised diagnosis: Platform elongate, with asymmetric, moderately wide, anteriorly clearly demarcated lobe directed to the side, positioned well anterior of central node, margin posterior to platform lobe without any sinus, with a constricted, pointed tip, or with an additional weak to distinctive sinus defining the posterior lobe; short free blade and carina first straight, then slightly sigmoidal, fine (with three to seven nodes) after the central node, directed straight or kinked inwards, shortened or reaching the platform tip; platform weakly to moderately ornamented, small nodes arranged along platform margins may form a denticulated margin especially in anterior part. Discussion: Specimens identified in the literature as Pa. ljaschenkoae display a large variability of platform shape and ornamentation. This supposed large variability contrasts with narrower taxonomic concepts in other Frasnian palmatolepids. As a step towards refinement, three morphotypes are distinguished, mostly based on the different shape of the posterior outer platform. Since they co-occur in several Martenberg beds, we do not suggest a subspecies differentiation. Morphotype 1, which includes the holotype, is characterised by an asymmetrically arched platform margin, without any sinus, from the tip of the lobe to the posterior end. The lateral lobe and the round posterior part of the platform are, therefore, very weakly defined. The surface of the platform is rather smooth apart from nodes at the interior anterior margin in the holotype. Morphotype 2 lacks a clear sinus delimiting the posterior end of the lobe but has a constricted and pointed posterior platform tip. Marginal nodes and fewer nodes inside the platform are typical. Morphotype 3 has a pronounced sinus directly after the lobe and also mostly a constricted and pointed tip. The ornament resembles M2 but can be even stronger in some specimens. Palmatolepis ljaschenkoae resembles Pa. proversa because of the partially (slightly) anteriorly oriented lobe. However, the angle of the anterior sinus is significantly more acute in Pa. proversa and its platform is stronger ornamented. Matveeva and Zhuravlev (2014) showed that juveniles of Pa. proversa may resemble morphotypes 1 and 2 of Pa. ljaschenkoae. Morphotype 1 of Pa. ljaschenkoae approaches Pa. simpla, which, however, differs by an extended, wider and anteriorly longer platform. Stratigraphic and geographic range: The species ranges from higher parts of Frasnian Zone 8 (housei Zone) to within Frasnian Subzone 11b (nasuta Subzone; compare composite range in Klapper 1997) and occurs in Western Canada, the Illinois Basin, the Rhenish Massif (Germany), Montagne Noire (France), the Timan-Pechora Region (northern Russia), western Urals, and Canning Basin (Western Australia). Tagarieva (2013) illustrated a much younger (linguiformis Zone) supposed Pa. ljaschenkoae from the southern Urals, which possesses a more posteriorly situated side lobe, at the level of the central node. We regard it as a Pa. winchelli with weak posterior platform sini. Revised diagnosis: Subtriangular platform, moderately to strongly ornamented, anterior part usually with rostrum, posterior part asymmetrically constricted, pointed and bent down; free blade and carina anterior to central node straight to slightly curved; carina posterior to central node straight, fine, approaching posterior tip; narrow side lobe strongly anteriorly oriented, with acute anterior sinus and distinct notch, located anterior of central node; underside of lobe with keel (Morphotype 1) or without (Morphotype 2). Discussion: The initial German description by Ziegler (1958) contained details omitted in the revised diagnosis by Ziegler and Sandberg (1990). These points are considered in the present revised diagnosis and contribute to the assignment of two morphotypes, which are established in order to provide a base to establish possible different distributions in time, space, and facies. Morphotype 1 has a keel on the aboral side of the lobe, which runs towards the centre of growth at an angle of ca. 45°i n relation to the longitudinal keel, Morphotype 2 lacks the keel. In general, we follow the synonymy of Klapper and Foster Jr. (1993), which is not repeated, but add recent records and supply older references that showed lower views, where the morphotype affiliation is clear. This species resembles Pa. ljaschenkoae in which the angle of the anterior sinus is significantly wider (> 90°), with the lobe oriented to the side, and its platform is usually weaker ornamented. Juveniles of both species may be similar (Matveeva and Zhuravlev 2014). Klapper and Foster Jr. (1993) synonymized Pa. barba Pa. redana Irwin andOrchard, 1991. The latter is an extreme variant with a spine-like side lobe oriented at < 45°t o the carina and with a reduced posterior carina. Such forms are currently unknown from Europe. Therefore, it qualifies as a palaeogeographically restricted form, either as a regional full species or subspecies. Stratigraphic and geographic range: The species ranges from Frasnian Zone 9 (proversa Zone) to Frasnian Subzone 11b (nasuta Subzone) and can be found in the Canadian Rocky Mountains, Iowa (USA), Belgium, Rhenish Massif (Germany), Montagne Noire (southern France), Morocco, Holy Cross Mountains, Pomerania (both Poland), Timan-Pechora Region (northern Russia), South China, and Canning Basin (Western Australia). Ovnatanova, 1996 (Figs. 5f, 7l, 11d-e, r;cf.: Figs. 7l, 9m) Discussion: Specimens with a markedly curved carina, with fine denticles after the central node, and raised, strongly ornamented inner margin, but without a rostrum on the platform (Figs. 5f, 7l, 11d), are intermediate towards Pa. manzuri Bardashev, 2009 (see Fig. 6k), in the sense of its holotype, but were part of the type series (Klapper et al. 1996, fig. 7.4). Specimens with only a few coarser nodes at the margin supports intergradation between the two species but they partly lack the posterior carina (Figs. 11e, r). Because of the close similarity, Klapper and Kirchgasser (2016) placed Pa. manzuri in synonymy with Pa. amplificata. Since rostra or raised anterior platform margins were not documented in the Asian populations (Bardashev 2009), we decided to report both forms until variabilities are better understood based on larger collections. Specimens identified as Pa. cf. amplificata (Figs. 7l, 9m) differ in an only weakly sinuous carina consisting after the central node of one or a few coarser, partly irregular, merged nodes. The platform is not flat but has raised margins both on the outer and inner side, including the side lobe, which bears a short side carina. Klapper and Kirchgasser (2016, fig. 12.15) included a somewhat different variant with weakly bent, coarse and shortened carina in Pa. amplificata. Further material is required to define subspecies or morphotypes within Pa. amplificata and to establish their ranges in space and time. Stratigraphic and geographic range: Frasnian Zone 10 (plana Zone) to Frasnian Zone 12 (winchelli Zone) of the Canadian Rocky Mountains, Iowa, New York State, Texas, SW England (Devon), eastern Rhenish Massif, Holy Cross Mountains, and the Timan (Russia). The cf. specimens from Martenberg extend the range to Frasnian Zone 9 (proversa Zone).

Description:
Outline of platform moderately wide, triangular. Relatively broad, laterally directed lobe with shallow to moderately developed posterior sinus. Carina nearly straight, with slight curvature anterior of central node and posteriorly either fine, short, or missing. Discussion: Specimens identified as Pa. aff. feisti differ from typical Pa. feisti in their comparatively narrower platform and their posterior carina, which is finer, short, or completely missing. In contrast to Pa. jamieae rosa n. ssp., their outer anterior platform is strictly concave, as in Pa. feisti. Lazreq (1999, pl. 8, fig. 4) identified a somewhat similar specimen as Pa. jamieae with a relatively pronounced sinus of the posterior platform margin. Stratigraphic and geographic range: Frasnian Zone 11a (feisti Zone) in the eastern Rhenish Massif. The Moroccan specimen is probably from the base of Frasnian Zone 12 (winchelli Zone).

Microfacies
On the basis of the nomenclature of Dunham (1962), Hartenfels (2011) defined 19 modified microfacies types for outer shelf settings, which can be assigned to the standard facies zones (1B to 4) of Flügel (2004) and can be divided into two facies series, MF-A: facies below storm wave base, MF-B: facies supposedly within the influence of storm waves or bottom currents. According to Hartenfels (2011), this modification is necessary because hemipelagic or neritic carbonates cannot be classified with sufficient differentiation in the classic microfacies zone model of Wilson (1975) and its critical review by Flügel (2004). However, it should be noted that current-induced sedimentation in the subphotic, pelagic realm may originate from contourites rather than from storms. In the Rhenish Massif, the modified classification sensu Hartenfels (2011) was successfully applied by Lüddecke et al. (2017) on the Upper Ballberg Quarry, a middle Famennian hemipelagic seamount section. Since the Martenberg section was deposited on top of a drowned volcanic seamount, an older but comparable depositional setting is assumed to interprete the microfacies (Fig. 13).
In the micrite-saturated, partly micro-sparitized matrix of the reddish to reddish-brown, flaser-bedded wackestones (MF-A4 / MF-B1) of the top of Bed R (Fig. 13a) and the base of Bed R-Q (Fig. 13b), the proportion of skeletal hard parts varies between 10 and 30 % (compare Hartenfels 2011). The matrix is heavily churned due to bioturbation of unknown origin, which causes a partly nodular texture (Fig. 13a). Poorly preserved dacryoconarids occur isolated or as small packstone nests (Fig. 13a1) together with ostracods of Thuringian and "Entomozoacean" ecotypes. There are iron encrustations at the pressure solution seams and some dispersed aggregates with small, Frutexites-type microstromatolites (Fig. 13a2). The lower part of Bed R-Q differs from the top of Bed R by two minor erosion horizons (Figs. 13b1, 13b2), which subdivide the interval. Above undulating truncation levels, there are thin accumulations of angular to subangular, small-sized white ironmineralized extraclasts, probably partly volcaniclasts, within a micrite-saturated, partly micro-sparitized matrix.
"Sheet 1" (Fig. 13c) is a well-sorted, fine-grained, partly chloritized (Fig. 13c1), calcareous, volcaniclastic packstone without any fossils. It resembles hyaloclastite matrix described from the region by Sunkel (1990). Dark iron minerals and clay minerals form the flaser seams, which represent diagenetic dissolution fronts. "Sheet 2" (Fig. 13d) consists of several layers, layers 1-5. Layer 1 consists of a very fine, recrystallized tuffite with some chlorite in its greenish to greenish-brown micro-sparitic matrix. It is followed by an undulating discontinuity surface that is partly enhanced by a thin, sparitic sheet crack, Layer 2 ( Fig.  13d1). Layer 3 is a thin band of grey, clotted micrite, partly with tube-like, meandering internal structures (Fig. 13d1). It contains some fine, angular clasts, more frequent subrounded intraclasts derived from the lowermost layer, partly with dark rims and minute hematite crystals, and greenish, angular to irregular, partly slightly larger, altered (chloritized) volcaniclastics. The upper boundary of the grey band is sharp, rippled, and followed by an interval of coarse, light-grey to white orthosparite, Layer 4 (Figs. 13d, 13d1). Angular, moderately sorted, often cavernous volcaniclasts, partly associated with iron minerals, and a small amount of fine white clasts, float isolated in the sparite, mostly at the base (Fig. 13d). Layer 5 is the upper part of the thick sparite seam and characterised by a high number of floating, mostly chloritic volcaniclasts. Some of the white clasts represent crinoidal detritus since a few small crinoid ossicles were found in the barren conodont residue.
The flaser-bedded "Sheet 3" interval ( Fig. 13e) consists mostly of a reddish, micritic mudstone with a small amount of small angular, white extraclasts. At the base (Fig. 13e3) and the top (Fig. 13e1), there are, again, thin volcaniclastic layers, combined with iron-mineralized dissolution seams. The volcanic particles differ from those in "Sheet 1 and 2" and contain packed, small, spherical, chloritized aggregates (Fig. 13e2). In the middle of the unit, there are a few larger lumps of these microlapilli (Fig. 13e4). Distinctive is the almost lack of identifiable skeletal remains and the lack of microfauna. Bed Q begins with a last greenish volcaniclast layer (Fig. 13f3), overlain by reddish-brown, partly micro-sparitized, flaser-bedded, micritic cephalopod limestone (Fig. 13f) with fragmentary or complete goniatites that include Ponticeras and Manticoceras (see House and Ziegler 1977), ostracods, fragmented, partly ribbed bivalves, and rare last microlapilli (Fig. 13f2). Some goniatites show horizontal, geopetal filling (Fig. 13f1) with very fine micrite and orthosparite.

Facies development across the semichatovae Event
The middle-upper Frasnian palaeogeographic setting of the volcanic Martenberg seamount can be characterised as constantly shallow pelagic and below the euphotic zone. This is in accord with the rich conodont assemblages and macrofaunas (e.g. Holzapfel 1882;Paeckelmann 1936;House and Ziegler 1977) that are dominated by cephalopods associated with specialised, subphotic benthic faunal elements, such as buchiolid and other bivalves, rare gastropods, very rare proetid trilobites, and smallsized, deeper-water solitary rugose corals. The fauna and microfacies of Bed R represent MF A4 sensu Hartenfels (2011), which suggests calm and slow deposition below the photic zone and storm wave base. Based on the strong bioturbation, the sea-floor was fully oxic. Microstromatolites (Fig.  13a2) are typical for such condensed pelagic sedimentation (compare Préat et al. 2008;Hartenfels 2011;. Dacryoconarid nests may stem from the burrowing (Fig. 13a1).
The minor erosion horizons of Bed R-Q indicate brief episodes of discontinuous sedimentation with iron mineralizations (Fig. 13b1), episodic winnowing/scouring (Fig. 13b2), and fine extraclast (volcaniclast) influx caused by bottom currents, either by distal storms or contourites. A characteristic feature of such MF-B1 wackestones (sensu Hartenfels 2011) is the combined occurrence of components of different provenance. The episodically increased bottom turbulence suggests a minor regressive trend in the terminal middle Frasnian. It is recorded in the conodont biofacies by a dominance of Polygnathus and partially Icriodus (Table 2), which persists from Sample R top to Sample R-Q 4.5-8 cm below top (Frasnian Zone 9 = proversa Zone to Frasnian Subzone 11a = feisti Subzone). This trend is briefly interrupted in Sample R-Q 8-14 cm below top by a small resurge of Palmatolepis ( Table 2).
The top of the middle Frasnian, recorded in the thin "Sheets 1-3" interval, is marked by a further increase of bottom turbulence, early lithification, minor reworking, several phases of nonsedimentation, a complete perturbation of the fossil succession, at a time with resumed local volcanism. Previously, the youngest pyroclastic eruptions of the region were known until the lower/ middle Frasnian transition (e.g. Bottke 1965; Sunkel 1990). It is remarkable that the significant changes of deposition were not recognised during the initial bed-by-bed surveys of the section (e.g. Ziegler 1958Ziegler , 1971, perhaps because of the lateral unconformities increasing on the "classic" eastern face of the cliff. Klapper and Becker (1999) noted the layered interval with sheet cracks at the base of Bed Q but, due to macroscopic similarity, did not separate crinoidal bioclasts from the volcaniclasts.
The partly chloritic clasts of "Sheet 1" (Fig. 13c1) resemble the fine crystalline matrix of tuffites that are poor in degassing structures as described from the region by Sunkel (1990). There is some sorting but no grading or cross-bedding, as typical for epiclastites that are commonly redeposited on volcanic slopes. Since there are no coarse lapilli or volcanic breccias, it is not likely that the Martenberg was re-activated as an eruption centre. Due to the partial good sorting in "Sheets 1 and 2", the volcaniclasts may represent distal fallout from one of the many other volcanoes of the region. Layer 1 of "Sheet 2" consists of even finer (Fig. 13d1), probably more distal fallout mixed with carbonate mud. The overlying fine sheet crack (Layer 2) indicates a depositional interruption due to increased bottom water agitation. The overlying clotted micrite (Layer 3, Fig. 13d1), especially in combination with Stromatactis-type sheets (Fig.  13d), is typical for a microbial origin (e.g. Zhou and Pratt 2019). Therefore, we relate the short, diffuse tubular structures also to microbial biomineralization. Minor bottom currents led to the re-deposition of reworked and coated tuffite clasts within Layer 3 and an influx of microlapilli that were later chloritized (Fig. 13d1). The sharp, rippled surface of Layer 3 provides further evidence for weak bottom currents. This culminated in a longer phase of non-deposition marked by the thick second sheet crack (Figs. 13d,13d1). Since chloritized microlapilli float in its late diagenetic, blocky orthosparite, there must have been an original microbial mat in the space of layers 4-5. Fine volcaniclastic and biogenic detritus was washed in by episodic currents and could not sink in the growing organic layer. This "volcaniclastic bindstone" is a rather unique type of sediment that lies outside the normal microfacies classification scheme. In the overlying "Sheet 3", pelagic micritic limestone deposition resumed but the fossil content remained very sparse (Fig. 13e) and the influx of fine volcaniclastics (Figs. 13e1-4) continued episodically.
The biotic perturbation at the end of the middle Frasnian includes a strong local decline of planktonic dacryoconarids, which proliferate in general under eutrophic conditions. The absence of conodonts in all three sheet samples, especially in the micritic "Sheet 3", suggests a phase of poor living conditions. As noted in the chapter on conodont stratigraphy, several taxa disappeared or declined locally at the top of Bed R-Q (e.g. Ag. amplicavus, Ag. coeni, Pa. amplificata, Pa. domanicensis, Pa. jamieae savagei n. ssp.). Microbial mats develop as typical opportunists in crisis times, especially when there are no grazers and during strong oligotrophy (e.g. Prieto-Barajas et al. 2018). In other Rhenish successions, zebra and Stromatactis limestones yield very rich pelagic conodont faunas (Becker et al. 2016a, c). It is not likely that the regionally youngest volcanic episode caused directly the local ecosystem crisis. The Givetian lapilli tuffites at the base of the Martenberg cliff are characterised by a highly diverse fauna of benthos and nekton.
The onset of the transgressive semichatovae Event is placed above the last thin, sorted, and current-induced volcaniclastic layer right at the base of Bed Q (Fig. 13f3). This is supported by the initial appearance of Pa. semichatovae and Pa. nasuta (Frasnian Subzone 11b / nasuta Subzone) and by a conodont biofacies shift to dominant Palmatolepis (Table 2). Although our Sample Bed Q base is not diagnostic due to the very small number of conodonts, this trend can be observed well in the higher samples of Bed Q (18-30 cm above base). It is noteworthy, that no new specimens of Pa. semichatovae were found during resampling. As shown by the onset of micritic cephalopod floatstones (Fig. 13f), the basal upper Frasnian environmental conditions at Martenberg were too pelagic; the species prefers more neritic facies, as in the Ardennes or North America. The deepening trend is supported by a sudden bloom of goniatites (e.g. Wedekind 1913;House and Ziegler 1977) and pelagic bivalves, but the dacryoconarid population did not recover. The spread of Stilleoceras (= Maternoceras) and entry of Playfordites within Wedekind´s do Iß characterise UD I-I, as part of a global, rapid, basal upper Frasnian goniatite radiation (e.g. House 1993, 1997). It should be noted that the position of Iß is incorrectly marked in House and Ziegler, 1977, Fig. 2); the true level of rich Stilleoceras faunas is given in the bed-by-bed goniatite record. After the volcaniclastic influx had faded very gradually in the main part of Bed Q (Fig. 13f2), the hemipelagic, subphotic, and calm deposition continued in Bed P. Upsection, its cephalopod floatstones with Manticoceras were increasingly altered by late diagenetic dolomitisation, resulting in yellow to pink weathering colours and poor preservation of macrofaunas (compare Wedekind 1913;House and Ziegler 1977).

Conodont zonation and the middle/upper Frasnian boundary
The problem of the precise correlation of the jamieae Zone in its type locality with the Frasnian zonation has been briefly outlined above. It has to be emphasised that the Martenberg was designated as the reference section for the zone and that it is, therefore, critical to reach a refined understanding and to clarify if it is a useful biostratigraphic unit at all. As shown below in detail, the zone has been uncritically used by many authors on a global scale. Therefore, we are aware that it will not be easy to delete it from the future literature. Ziegler and Sandberg (1990) assigned beds R and R-Q and laterally equivalent samples (samples VI´/12-10, VI/12a) to the Upper hassi Zone, Bed Q (samples VI´/9-6, VI/11b-11a) to the jamieae Zone, and Bed P (samples VI´/5, VI/10e) to the Lower rhenana Zone (Figs. 3 and 14). However, no Pa. jamieae was figured from the jamieae Zone and due to their wide concept of the species, including Pa. feisti and forms placed by us in separate species/subspecies, the Martenberg level of Pa. jamieae sensu its holotype remained undocumented. Klapper and Becker (1999) correlated Bed R-Q with Frasnian Zone 10 (plana Zone), Bed Q, the supposed jamieae Zone, with Frasnian Zone 11 (feisti Zone), and Bed P, the assumed basal Lower rhenana Zone, with the lower part of the Frasnian Zone 12 (winchelli Zone). This created a contrast with the Lion Quarry section of Sandberg et al. (1992), where the Lower rhenana Zone falls in Frasnian Zone 11,not in Frasnian Zone 12. This stimulated faunal revisions and corrections by Ziegler and Sandberg (2000), who re-assigned the "upper part of Bed R" (VI'/10), which refers to Bed R-Q, to the jamieae Zone, because of a rare occurrence of supposed Pa. jamieae, Bed Q (Sample VI/11b) to the Lower rhenana Zone, and Bed P to the Upper rhenana Zone. The jamieae Zone shifted completely its position and became very thin. This resulted in a new correlation of the Frasnian Zone 10 with the assumed jamieae Zone, of Frasnian Zone 11 with the lower part of the Lower rhenana Zone, as at Lion, and of Frasnian Zone 12 with the basal Upper rhenana Zone. As in 1990, it remained unknown, which form included in Pa. jamieae (here "Pa. jamieae auct.") does occur in the new jamieae Zone level. Clarification was partly provided by the revision of original collections by Ovnatanova and Kononova (2000), who did not find any true Pa. jamieae in beds R-Q and Q, only in Bed P, which equals the type level of the species (and subspecies) at Schmidt Quarry. Consequently, the justification for the jamieae Zone as a separate biostratigraphic unit was denied.
Our new samples enable a further clarification of the problem. We can confirm that no typical Pa. jamieae, Pa. jamieae jamieae of our revised taxonomy, occur in beds R to Q. We also did not find any among the 1763 new Pa elements from the middle part of Bed Q. This suggests that it is locally a very rare form with random occurrences that is NOT suitable as a zonal index taxon. The recognition of Pa. jamieae savagei n. ssp., Pa. jamieae rosa n. ssp., and Pa. adorfensis n. sp. in the lower/middle part of Bed R-Q explains the Ziegler and Sandberg (2000) jamieae record from this bed and supports the implied revised correlation with Frasnian Zone 10 (plana Zone). However, our new taxa are rare and Pa. plana, represented by 50 specimens in our two samples, is a distinctive, much better zonal index form. We follow Ovnatanova and Kononova (2000) and postulate to abandon the jamieae Zone completely. If authors do not wish to apply the Montagne Noire zonation, they should replace the jamieae Zone by the plana Zone, which has the big advantage to enable a straight forward correlation of both zonation schemes. Based on data in Sandberg et al. (1992), the entry of I. praealternatus praealternatus predates in Belgium the entry of Pa. plana, but not at Martenberg. Ancyrognathus triangularis is an additional Frasnian Zone 10 marker but becomes much more common in the overlying zones.
As explained above, the Frasnian Zone 11 (feisti Zone) is subdivided into Frasnian Subzone 11a (feisti Subzone) and Frasnian Subzone 11b (nasuta Subzone). The first begins just before the probably eustatically controlled regression preceding the semichatovae Transgression and may have been a short interval. However, this view is based on the reduced thickness at Martenberg, which is perhaps distorted by times of non-deposition. Based on the successions of the Timan from Member 1 (TP-VIIb Assemblage of Ovnatanova and Kononova 2020) and basal part of Member 2 of the Lyaiol Formation (basal TP-VIII Assemblage of Ovnatanova et al. 1999), and the Canning Basin (Klapper 2007), Pa. elegantula, Pa. ederi, Pa. timanensis, and Pa. brevis also occur in Frasnian Subzone 11a (feisti Subzone) but may partly enter slightly earlier than Pa. feisti (compare composite range for Pa. ederi in Klapper et al. 1996 and range charts in Ovnatanova and Kononova 2008Kononova , 2020. Therefore, there is not a precise correlation with the regional, northern Russian elegantula-semichatovae Zone of Ovnatanova and Kononova (2008), which ranges from the top of Frasnian Zone 10 (plana Zone) to the top of Frasnian Subzone 11b (nasuta Subzone).
The semichaetovae Transgression or semichatovae Event s.str. characterises the base of Frasnian Subzone 11b, with Pa. nasuta as index species (as for the Lower rhenana Zone). It is important to distinguish Pa. nasuta correctly from the morphologically somewhat similar, partly older, partly overlapping Pa. mucronata and from the younger Pa. boogardi (see Klapper et al. 1996, p. 147: focus on the carina curvature and orientation). Auxiliary index forms are Pa. semichatovae (index of the shallow-water semichatovae Subzone of Morrow andSandberg 2008 andof Alberta Zone 5 of Klapper and and Po. lodinensis (compare Klapper 2007 andNarkiewicz andBultynck 2011; but see an alleged lower range predating the entry of Pa. plana in the Polar Urals, . In northern Russia and Australia, Pa. anzhelae and Pa. playfordi are further guide species (see Klapper 2007;Ovnatanova and Kononova 2020); in the Timan probably also Po. siratchoicus and Pa. lyaiolensis (Ovnatanova and Kononova 2008;Ovnatanova et al. 2017;. In New York State and Iowa, Po. unicornis and Tortodus deformis enter with Pa. semichatovae (Klapper and Kirchgasser 2016;Day and Witzke 2017). Palmatolepis eureka appears in Belgium at the same level as Pa. nasuta Sandberg et al. 1992). In the shallow-water regions of the Russian Platform, Pa. semichatovae provides a correlation with the Polygnathus subincompletus Zone of Ovnatanova and Kononova (2008).
The combined FADs of palmatolepids, polygnathids, and species of other genera give an optimal correlation potential. Therefore, we propose to place the future base of a formal upper Frasnian substage at the base of Frasnian Subzone 11b (base nasuta Subzone = base Lower rhenana Zone = base semichatovae Subzone), which conforms with the original proposal of Ziegler and Sandberg (1997). The so defined substage would be easily recognisable in ammonoid biostratigraphy by the onset of UD I-I faunas and in sequence stratigraphy by the TST base. Detailed chemostratigraphic work is required. The final ending of basaltic phreatomagmatic volcanism at Martenberg may be a coincidence; elsewhere in the northern Rhenish Massif, last "Diabas" lava flows were found up to the highest Frasnian (e.g, at Wuppertal-Barmen, Paeckelmann 1928b).
Previous regional records of supposed Pa. jamieae and the jamieae Zone In the light of the new knowledge, previous records of the jamieae Zone and supposed regional Pa. jamieae occurrences are reviewed, giving the relevant literature for further reading. This is an essential precondition to revise former age identifications, international correlations, and an important step towards a correct substage recognition in many regions.
In the Great Basin of the Western United States, at Devils Gate, Nevada, Ziegler and Sandberg (1990) showed the supposed position of the jamieae Zone but without any record of its index form above the FOD of Pa. plana and below the FOD of Pa. nasuta. A supposed specimen from the top of the "Early rhenana Zone" was re-assigned by Klapper (2007) to Pa. uyenoi. There is also no faunal record of the jamieae Zone at Tempiute Mountain (Morrow 2000), where not-figured "jamieae auct." were first recorded with Pa. nasuta, which is Frasnian Subzone 11b (nasuta Subzone). In the chert facies at Whiterock Canyon (Morrow 2000), there is no record of "Pa. jamieae auct." in the noted jamieae Zone, as at Devils Gate; only Pa. proversa and Pa. simpla were found, which suggests Frasnian zones 9/10 (proversa/plana zones). Levman and Bitter (2002) illustrated as Pa. jamieae a specimen from the Long Rapids Formation of northern Ontario (Canada), which combines the platform shape of Pa. feisti with an ornament as in Pa. winchelli. This homoeomorphic new form comes from the top of Frasnian Zone 13a (top bogartensis Zone) and has nothing to do with Pa. jamieae jamieae, nor with the older Pa. feisti. This is probably also true for three further not figured specimens and for another one from the subsequent basal Frasnian Zone 13b (linguiformis Zone). The region has no jamieae Zone record but Frasnian Zone 10 (plana Zone) was distinguished by Klapper et al. (2004). Helsen and Bultynck (1992;compare Bultynck et al. 1998) identified the jamieae Zone in the Nismes section of the Ardennes and showed it questionably in the Mariembourg section in the southern flank of the Dinant Syncline. However, both Nismes specimens illustrated are not Pa. jamieae jamieae, but Pa. jamieae rosa n. ssp. (their pl. 3, fig. 7) and Pa. wildungensis (their pl. 3, fig. 6). The first comes from high in Member 2 of the Neuville Formation (Bed C10B) and from well above the level of the semichatovae Event, suggesting Frasnian Subzone 11b, the second from the higher part of Member 3 (Bed C13B), above the FOD of Ad. ioides (in Bed C12) and above the FOD of Pa. winchelli (in Bed C12c; identified by Bultynck et al. 1998, pl. 2, fig. 7 as Pa. gigas), therefore, from Frasnian Zone 12 (winchelli Zone). The listed (not figured) oldest "Pa. jamieae auct." enter at Nismes with Ag. triangularis and above the first Pa. plana in Bed C1. Therefore, they come from Frasnian zones 10 to 11 (plana to feisti zones); which gives a good match with the Martenberg. Younger "Pa. jamieae auct." from the Upper rhenana Zone (Frasnian Zone 12) may have included typical specimens. The specimen from the Frasnes-W section illustrated by Bultynck et al. (1998) does not represent Pa. jamieae jamieae but Pa. adorfensis n. sp. (pl. 1, Fig. 13). As the not figured "jamieae auct." specimens from below (Sample CFrW 2), it comes from Frasnian Subzone 11b (nasuta Subzone); the jamieae Zone was not recognised in that section, nor at Neuville and Lessive, where not figured "Pa. jamieae auct." occur in the Upper rhenana Zone. Sandberg et al. (1992) reported Pa. jamieae from the Lion Quarry access road section and the north side of the Tiènne du Lion section and used it to determine the local lower limits of the jamieae Zone. One of the two specimens of Pa. jamieae shown is Pa. feisti (their pl. 3, fig. 5). It was found at a level (Sample 85-BEL-118) above the entry of Pa. nasuta (85-BEL-115), Pa. winchelli (= subrecta) and Ad. ioides (both in 85-BEL-116), therefore, in Frasnian Zone 12 (winchelli Zone). The other figured specimen is Pa. jamieae jamieae. It is even younger (85-BEL-120, their pl. 3, fig. 3) and was associated with Pa. rhenana, as in the Kellerwald type level. The not figured "Pa. jamieae auct." enter with (Tiènne du Lion) or above (access road section, base of Boussou-en-Fagne Member) Pa. plana in Frasnian Zone 10, which gives another good fit with Martenberg. In the railroad cut west of the Lion Quarry and in the section at the south end, the jamieae Zone was not recognised but Frasnian Zone 10 (plana Zone) is recognisable by its index species, near the base of the Boussou-en-Fagne Member. Therefore, the Lion successions underline the usefulness of the plana Zone. In summary, no early member of the Pa. jamieae Group has ever been figured in the Ardennes from time equivalents of the supposed jamieae Zone at Martenberg. The typical subspecies occurs regionally at its type-level while the precise affinities of even younger specimens, from the lower part of the Matagne Formation, are currently unknown.
In the eastern Rhenish Massif, the jamieae Zone was recognised by Ziegler and Sandberg (1990) in the Heimberg section, which still exists but the middle-upper Frasnian transition lies at a very steep slope above an old quarry and is dangerous to access. Only some beds have been sampled so far. The jamieae Zone was recognised above the entry of Pa. plana and Ag. triangularis (Frasnian Zone 10 / plana Zone) and below the entry of Pa. nasuta (Frasnian Subzone 11b / nasuta Subzone). None of the "jamieae auct." specimens were illustrated and Pa. feisti or early Pa. jamieae subspecies would be expected between the two recognised levels. The section has potential but resampling is required. Pas et al. (2013) showed the jamieae Zone in the Burgberg section situated on a drowned volcano south of the Brilon Reef. They referred to conodont data of Stritzke (1990), who did not report any Pa. jamieae, and to their own samples, which results were not given.
At Benner near Bicken in the southern Rhenish Massif (Lahn Syncline), the middle-upper Frasnian transition shown in the section log by Ziegler and Sandberg (1990) is condensed. The more than 3 m thick interval from the supposed base of the Lower rhenana Zone to the base of the Lower Kellwasser level does not correlate well with the more detailed log of Schindler (1990), who noted that older beds became inaccessible after flooding by a fish pond, including strata up to the middle Frasnian (Schülcke 1995). Ziegler and Sandberg´s not figured "Pa. jamieae auct." enter jointly with Pa. plana, as at Martenberg, and range into faunas with Pa. rhenana and Pa. winchelli (= subrecta) from just below the Lower Kellwasser level (higher Frasnian Zone 12 = winchelli Zone). The latter specimens may have included typical Pa. jamieae. The local supposed jamieae Zone (beds 59-60), here re-assigned to the plana Zone/feisti Subzone interval, was thin (ca. 25 cm); it needs to be restudied.
In the Carnic Alps, Spalletta and Perri (1998) recognised the jamieae Zone based on the entry of Pa. foliacea, which, however, enters later (high in Frasnian Zone 11, Klapper and Kirchgasser 2016) than the supposed jamieae Zone levels at Martenberg. In France and Spain, the jamieae Zone has not been identified although the term was used unspecifically by Sanz-López (2002) in a discussion of the Pyrenees Frasnian.
In the more southern part of the Mrirt region, at Tougguiou-Allal (see map in Becker et al. 2020b), the jamieae Zone was again recognised by Lazreq (1999) in beds with Ag. ioides (s.l.), above the FOD of Ag. triangularis and below the FOD of Pa. semichatovae, which suggests the Frasnian Zone 10/ Frasnian Subzone 11a interval, as at Martenberg. However, no specimen was figured. From Mrirt far to the northwest, in the northern El Hamam Zone, the jamieae Zone was recognised by Lazreq (1999) at Bou Alzaz North and Aïn Azza in the same position; again, the affinities of not figured "Pa. jamieae auct." are unclear. Re-sampling is required, also to clarify whether the repeated early Ag. ioides records of the Meseta refer to Ag. ioides M1 sensu Klapper (2021) that enters in the Rhenish Massif in Frasnian Subzone 11a (feisti Subzone, Stichling et al., this vol.). Matyja and Narkiewicz (1995) described the jamieae Zone from the Janczyce I borehole section (at 373.3 m depth) in the eastern Holy Cross Mountains (Poland) and pictured in a slightly oblique view one of their two oldest specimens, which is probably Pa. jamieae n. ssp. δ. Associated are mostly polygnathids, I. praealternatus praealternatus, and Pa. simpla, which suggests Frasnian Zone 10 (plana Zone). A locally youngest, not-figured "Pa. jamieae auct." was found in the same level as Pa. nasuta (at 347 m depth), indicating Frasnian Subzone 11b. In the Pagow IG 1 borehole of the Malopolska Massiv, Malec (2015) illustrated specimens close to Pa. kireevae as supposed Pa. jamieae from an interval with Pa. proversa and Pa. ederi, which was assigned to the jamieae-Lower rhenana Zone interval. A Pa. jamieae jamieae from a level with numerous Pa. subrecta (= winchelli; Frasnian Zone 12) was identified as "Pa. cf. rotunda".
In the southern Timan of the northern Russian Platform, Pa. jamieae jamieae is restricted to the regional TP-IX assemblage of Ovnatanova et al. (1999), which occurs in the upper part of Member 3 and Member 4 of the Lyaiol Formation (Ovnatanova and Kononova 2008) that falls in Frasnian Zone 12 (e.g. Klapper et al. 1996;House et al. 2000). Palmatolepis jamieae ssp. δ and an atypical variant of Pa. jamieae savagei n. ssp. M2 without posterior platform sinus are associated; no jamieae Zone was regionally distinguished. In the northern Timan/Subpolar Urals region, Savage and Yudina (2001) illustrated a specimen as Pa. jamieae, which is a typical Pa. feisti. Therefore, their jamieae Zone, recognised in Sample 87-2/83 in the Syv`yu River section, falls in Frasnian Subzone 11a (feisti Subzone). Their younger (Sample 87-2/98) Pa. cf. jamieae includes a probable Pa. nasuta with short side lobe and a Pa. ?hassi, suggesting a Frasnian Subzone 11b age (nasuta Subzone). Our revisions agree with Ovnatanova et al. (2017, tab. 4), who did not recognise any Pa. jamieae (s.str.) in their extensive samples from the same section and why they questioned the regional meaning of the jamieae Zone, which previously, however, was used by Tsyganko (2011). In the adjacent Shar´yu River section, not illustrated "Pa. jamieae auct." of Ovnatanova et al. (2017) occur within the Vorota Formation with Pa. semichatovae and Pa. nasuta in the Frasnian Subzone 11b (nasuta Subzone). Slightly younger specimens (Ovnatanova et al. 2017, pl. 35, Figs. 3 and 4, Sample 217) are re-assigned to Pa. jamieae rosa n. ssp. and Palmatolepis sp. ɛ; they co-occurred with Pa. gyrata, indicating Frasnian Zone 12 (see Klapper et al. 1996). From the Bolś haya Sa´uyga River, there are only cf. records of "Pa. jamieae auct." associated with Pa. cf. semichatovae , Table 2). In the Kozhim River section of the same region, Matveeva (2013) recognised the jamieae Zone above the FOD of Pa. plana and below the FODs of Pa. nasuta, correlating with the higher FZ 10 and FZ 11a interval, but the reported palmatolepid fauna lacks any Pa. jamieae records. For the Lemna River Basin of the Polar Urals, Ovnatanova et al., 2017, pl. 41, Fig. 1) documented a Pa. jamieae that represents ssp. δ. Details of its assemblage and precise level at Nadota River are unknown. In summary, the available date don´t justify a jamieae Zone below equivalents of Frasnian Subzone 11b in all of northern Russia.
In the Tatarstan region west of the Urals, Pa. jamieae ssp. δ was documented by Ovnatanova and Kononova, 2008, pl. 11, Fig. 3) from probable beds of Frasnian Zone 13a (bogartensis Zone). Not figured "Pa. jamieae auct." range lower, probably into Frasnian Zone 12 levels; no jamieae Zone was recognised. In the western South Urals, not figured "Pa. jamieae auct." were reported with Pa. plana and the endemic Pa. triquetra from within the regional "Domanik Horizon" of the Gabdyukovo section (Artyushkova et al. 2011), therefore, from within Frasnian Zone 10 (plana Zone). In the much more densely sampled Ryauzyak section, the jamieae Zone was based on the entry of not figured "Pa. jamieae auct." just before the FOD of Pa. plana, well below the very closely spaced successive entries of Pa. semichatovae, Pa. nasuta, and Pa. bogartensis (= rotunda) in the overlying Mendym Horizon (Artyushkova et al. 2011). Alleged jamieae specimens range very high in the Upper rhenana Zone (Frasnian Zone 13a). From the subsurface of the Volgograd region of the Russian Platform, Pa. jamieae savagei n. ssp. M1 was illustrated from an interval in borehole 96-Kamyshiiskoj that also contained Pa. lyaiolensis, Pa. foliacea, and Pa. subrecta (= winchelli), indicating Frasnian Zone 12. Not figured "Pa. jamieae auct." enter already below, at a level with Po. lodinensis and Pa. lyaiolensis, indicating Frasnian Subzone 11b (nasuta Subzone), and at an unspecified level of Frasnian zones 11-13 in a second borehole .
In the Rudny Altai of southern Siberia, the jamieae Zone was recognised by Izokh et al. (2004) based on occurrences of the typical subspecies. However, these come from well above the entry of UD I-I ammonoid faunas (unpublished material collected by RTB in 2005), indicative of Frasnian Subzone 11b, and are obviously associated with Pa. rhenana, Pa. muelleri, and Pa. jamieae savagei n. ssp. M2, suggesting a Frasnian Zone 12 (winchelli Zone) age for figured specimens. The supposed occurrence of Pa. jamieae in the Tien Shan of Uzbekistan is based on an atypical specimen closer to Pa. foliacea than to Pa. jamieae (Erina in Kim et al. 2007). A jamieae specimen from Tadzhikistan (Bardashev and Bardasheva 2014;Bardashev 2018) is very close to our Pa. adorfensis n. sp. Ji and Ziegler (1993) assigned in the Lali section of Guangxi, South China, their Bed 33 to the jamieae Zone but did not illustrate any specimen from this level, which contained Pa. plana, suggesting Frasnian Zone 10 (plana Zone). A not figured cf. jamieae specimen was reported from the level of Pa. proversa (Frasnian Zone 9 = proversa Zone) in the bed below. The overlying Frasnian Subzone 11b with Pa. nasuta yielded not figured "Pa. jamieae auct." (beds 34 and 36) and Pa. jamieae ssp. δ (pl. 27, fig. 2, Bed 37). Supposed jamieae specimens, re-assigned to Pa. jamieae savagei n. ssp. M2 (pl. 27, Fig. 1) and Pa. winchelli (pl. 27, Fig. 3), come from the overlying beds 38 and 39 with Pa. subrecta (= winchelli), proving Frasnian Zone 12. Zhang et al. (2019) did not recognise the jamieae Zone in their new Lali section, which is not the same as the now covered section of Ji and Ziegler (1993).
In the Shetianqiao section of Hunan, Ma et al. (2004) and Ma and Zong (2010) assigned a conodont-free interval below Pa. cf. semichatovae to the jamieae Zone. In summary, from all of South China, no jamieae specimen has ever been figured from time equivalents of the supposed jamieae Zone at Martenberg. From higher beds, there is no proven record of Pa. jamieae jamieae but of the three other subspecies. Additional forms seem to characterise Frasnian Zone 13a and require further work. The jamieae specimens from Inner Mongolia shown by Lang and Wang (2010, pl. 1, Figs. 12 and 13) are elongate variants of Pa. jamieae savagei n. ssp. M1 and come from a unit with mixed faunas.
Savage (2013,2019) described Pa. jamieae jamieae jointly with Ad. ioides, close relatives of Pa. subrecta (= winchelli; named as Pa. khaensis in 2019) and "Pa. aff. rotunda" from the upper part of Frasnian Subzone 12 (winchelli Zone) of Northwestern Thailand. A specimen named as Pa. aff. jamieae resembles Pa. jamieae savagei n. ssp. Morphotype 1 (Savage 2013: figs. 5, 13 and14 = 2019: figs. 9, 3, 4 and5) but may have been elongated by tectonic deformation. It is much younger than the Martenberg specimens and comes from the upper part of Frasnian Zone 13a (bogartensis Zone). Ta et al. (2021) recognised in Central Vietnam a jamieae Zone between an Ag. triangularis Zone below, beginning well above the FOD of Pa. plana, and a Pa. rhenana nasuta Zone above, therefore from the upper Frasnian Zone 10 to Frasnian Subzone 11a interval, as at Martenberg. However, none of the supposed Pa. jamieae from four beds have been illustrated, which prevents a re-interpretation. The same applies to "Pa. jamieae auct." noted in the range chart of Matsuo et al. (2020) for their Frasnian-Famennian boundary section (stratotype of the Xom Nha Formation) in the same region, which were noted between the FODs of Pa. rhenana and Pa. boogardi, therefore in Frasnian Zone 12 (winchelli Zone).
This summary underlines the very heterogeneous nature of previous literature records of Pa. jamieae and local versions of the jamieae Zone. Various oldest records fall in the globally recognisable Frasnian Zone 10 (plana Zone), as at Martenberg, but there is no evidence that a precisely correlated time interval was meant in the many different regions. Palmatolepis feisti has not been recorded widely but it may be hidden among "Pa. jamieae auct." specimens that were not illustrated. We expect that the feisti Zone/Subzone will be recognised in pelagic successions of more regions in the future.

Potential basal upper Frasnian GSSP sections
The northern slope of the protected Martenberg section currently has potential as a future GSSP for the base of a formal upper Frasnian substage. Geochemical work would have to be added to the data on conodont and ammonoid biostratigraphy, sedimentology/microfacies, and sequence stratigraphy. The extreme rarity of zircons in Rhenish basaltic volcanites strongly delimits the options for an absolute dating of the thin volcaniclastic layers. The closest zircon-bearing distal fallout layer belong to the Pegasus Group of the tephrostratigraphic correlation of Winter (2015). They lie a third up in the Lower rhenana Zone at Benner (Lahn Syncline), in the Lion Quarry of Belgium, and in the upper Oos Formation (in the supposed jamieae Zone) of the Prüm area, Eifel Mountains (Grimm and Rothausen 1992).
On a global scale, very few other sections with a sufficient bed-by-bed record of conodont faunas, especially of the top Frasnian Zone 9 (proversa Zone) to Frasnian Subzone 11b (nasuta Subzone) interval, and more specifically around the regressive-transgressive phases around the semichatovae Event, have been published. One exception is Devils Gate in Nevada , which includes a small unconformity between the entries of Pa. nasuta (Sample 10) and of rare Pa. semichatovae (Sample 9B), and which lacks other biostratigraphic markers, such as goniatites. There are other sections with potential for detailed correlation that should be re-investigated at finer detail before any GSSP decision. We omit from a short compilation sections that are very difficult to access (e.g. in the Russian Far East or in the military zone at the Moroccan-Algerian border), temporary (boreholes, filled trenches, e.g. at La Serre, southern France, overgrown, e.g. the original Lali section of Ji and Ziegler 1993; active quarries, such as Kowala in the Holy Cross Mountains), characterised by long siliciclastic intervals without conodonts (e.g. Appalachian Foreland, Tafilalt Basin, pelagic shelf basin facies of Germany), with extreme condensation or significant unconformities (e.g. central Tafilalt Platform), outcrop gaps (e.g. southern Timan river sections), or deposited in shallowwater facies (e.g. Pomerania, Russian Platform, Iran). This leaves the following sections: 7. Despite its condensed nature, the highly fossiliferous Martenberg section would be suitable as a future upper Frasnian GSSP section. Trace element and isotope stratigraphy data and magnetic susceptibility studies are still lacking. A global literature survey identified 22 further conodont sections that expose the middle/upper Frasnian transition but in most of them research is still at an early stage, with required revisions, re-sampling, closure of sampling gaps, full documentation of faunas, microfacies and chemostratigraphy studies. also grateful to guest editors Christoph Hartkopf-Fröder (University of Cologne) and Sven Hartenfels (Geologischer Dienst NRW, Krefeld) for their comments on the manuscript. Traudel Fährenkemper assisted with some of the graphics, Davina Mathijssen and Lukas Afhüppe processed conodont samples, Rike Zimmermann and her team helped with the preparation of thin sections, and Till Söte took new photos of the outcrop (all University of Münster). Hanna Cieszynski (University of Cologne) thankfully photographed some additional conodonts in the scanning electron microscope.
Funding Open Access funding enabled and organised by Projekt DEAL.
Data availability statement All data generated or analysed during this study are included in this published article.

Declarations
Conflict of interests The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.