The British stratigraphy of the Ordovician System has long been a standard reference, as summarised by Fortey et al. (Reference Fortey, Bassett, Harper, Hughes, Ingham, Molyneux, Owen, Owens, Rushton, Sheldon, Barnes and Williams1991), who list six series: in ascending order, Tremadoc, Arenig, Llanvirn, Llandeilo, Caradoc and Ashgill. Responding to the need for more precise definitions of series and stages, especially for international use and correlation, Fortey et al. (Reference Fortey, Harper, Ingham, Owen and Rushton1995) revised the chronostratigraphy of the Ordovician of the historical type area in England and Wales. The main novelty in their 1995 paper was the extension of the Llanvirn to include the lower part of the Llandeilo, as a stage called the Llandeilian. The upper part of the former Llandeilo was added to the Caradoc, bringing the base of the Caradoc down to the base of the Nemagraptus gracilis Biozone, an important and widely recognisable horizon. In conclusion, they obtained a five-fold division of the Ordovician System, into the Tremadoc, Arenig, Llanvirn, Caradoc and Ashgill series. The Ashgill Series is divided into four stages (Ingham & Wright Reference Ingham and Wright1970), i.e., the Pusgillian, Cautleyan (Ingham’s 1966 shelly fossil Zones 1–4), Rawtheyan (Ingham’s 1966 Zones 5–7) and the Hirnantian (Ingham’s 1966 Zone 8) stages, with the Cautley district in northern England serving as its type area (given its more instructive section than that of Ash Gill in the Lake District, where the origin of the name to the series lies). Since then, the Subcommission on Ordovician Stratigraphy (ISOS) of the International Commission on Stratigraphy has adopted a tripartite division of the system into a Lower, Middle and Upper Series and has installed a sub-series division into seven stages. The international Upper Ordovician Series is divided into the Sandbian, Katian and Hirnantian stages (Bergström et al. Reference Bergström, Finney, Xu, Pålsson, Wang and Grahn2000, Reference Bergström, Finney, Xu, Goldman and Leslie2006, Reference Bergström, Chen, Gutiérrez-Marco and Dronov2009). The base of the Sandbian equates to the base of the British Caradoc Series (as defined by Fortey et al. Reference Fortey, Harper, Ingham, Owen and Rushton1995), while the base of the Katian correlates into the mid-Caradoc. Comprehensive cross-calibrations of the British and international chronostratigraphies are detailed in Cocks et al. (Reference Cocks, Fortey and Rushton2010, fig. 1), but also see Molyneux et al. (Reference Molyneux, Harper, Cooper, Hollis, Raine, Rushton, Smith, Stone, Williams, Woodcock and Zalasiewicz2023) for a recent review.
Since the mid 19th Century, fossils have been collected from the Girvan area, consisting of rich Ordovician–Silurian brachiopod, echinoderm, trilobite and graptolite faunas (Ingham Reference Ingham, Lawson and Weedon1992; see Molyneux et al. Reference Molyneux, Harper, Cooper, Hollis, Raine, Rushton, Smith, Stone, Williams, Woodcock and Zalasiewicz2023 for recent overview). The Ordovician part of the succession has been divided into six graptolite biozones, and graptolites of the Pleurograptus linearis Biozone have been used historically to identify the boundary interval between the British Caradoc and Ashgill series (Ingham Reference Ingham1966; Ingham & Wright Reference Ingham and Wright1970; Fortey et al. Reference Fortey, Harper, Ingham, Owen and Rushton1995, Reference Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000; Molyneux et al. Reference Molyneux, Harper, Cooper, Hollis, Raine, Rushton, Smith, Stone, Williams, Woodcock and Zalasiewicz2023). This is largely based on observations of these graptolites in the Whitehouse Subgroup in Girvan, where they straddle the C/A boundary as defined by shelly fauna (Ingham & Wright Reference Ingham and Wright1970). However, in the type Cautleyan-Rawtheyan area of northern England, graptolites of the same biozone have been observed to range into the Rawtheyan (upper Ashgill) by Rickards (Reference Rickards2002), and would thus be significantly younger. Rickards (Reference Rickards2002) expressed uncertainty about the presence of the biozone in the Cautleyan or lower levels. A comparable upward extension of the P. linearis Biozone has not been documented elsewhere, however, and this has resulted in the zone having a different stratigraphic extent in different areas. Zalasiewicz et al. (Reference Zalasiewicz, Taylor, Rushton, Loydell, Rickards and Williams2009, fig. 1), for example, showed the P. linearis Biozone extending from below the Pusgillian Stage to the Rawtheyan Stage in northern England, but having a more restricted stratigraphical extent in Scotland, from below the Pusgillian Stage to a level in the Cautleyan Stage.
We here document chitinozoan faunas from across the P. linearis Biozone in Girvan to add to the record of shelly, graptolitic and conodont faunas, and investigate correlation between key sections at Girvan, on Bornholm and in northern England using chitinozoans as an independent biostratigraphic tool. Chitinozoan microfossils have proven to be an increasingly important and powerful biostratigraphical tool for long-distance correlations in the Ordovician, due to their short ranges and generally wide distributions. This functionality has been applied to create a robust biostratigraphy in key localities in the Anglo-Welsh basin (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005, Reference Vandenbroucke, Williams, Zalasiewicz, Davies and Waters2008, Reference Vandenbroucke, Ancilletta, Fortey and Verniers2009; Vandenbroucke Reference Vandenbroucke2008a, Reference Vandenbroucke2008b; Challands et al. Reference Challands, Vandenbroucke, Armstrong and Davies2014; Amberg et al. Reference Amberg, Molyneux, Zalasiewicz and Vandenbroucke2023), Baltica (Nõlvak & Grahn Reference Nõlvak and Grahn1993; Vandenbroucke Reference Vandenbroucke2004; Vandenbroucke et al. Reference Vandenbroucke, Recourt, Nõlvak and Nielsen2013) and Laurentia (Achab Reference Achab1989; McLaughlin et al. Reference McLaughlin, Emsbo, Desrochers, Bancroft, Brett, Riva, Premo, Neymark, Achab, Asselin and Emmons2016; Esteves et al. Reference Esteves, McLaughlin, Velleman, De Baker, De Weirdt, Miller and Vandenbroucke2025).
This is the last paper in a series of three interlinked studies to explore this strategy. A first paper (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005) documented the chitinozoans in abundance from the very graptolite slabs of the type Cautleyan-Rawtheyan area that were the basis of Rickard’s (Reference Rickards2002) study; a detailed chitinozoan biostratigraphy was established in these sections. The Ashgill chitinozoans from England and Scotland (Vandenbroucke Reference Vandenbroucke2008a) proved to have a Baltoscandian biogeographic signature, suggesting we would also be able to correlate with the chitinozoan biostratigraphy of the island of Bornholm in Denmark, which became the subject of the second paper in this series (Vandenbroucke et al. Reference Vandenbroucke, Recourt, Nõlvak and Nielsen2013). In Bornholm, the chitinozoan biozones were tightly calibrated against the – then newly revised – graptolite biostratigraphy, which included a clear documentation of the P. linearis Biozone. This third paper of the trilogy now focuses on yet another key section that covers the Caradoc–Ashgill boundary interval and includes the P. linearis Biozone. So far, extensive fossil collections from the Girvan area consist of rich Upper Ordovician–lower Silurian shelly and graptolite faunas. Additionally, a pilot study based on a limited selection of Upper Ordovician and lower Silurian samples from Girvan demonstrated the potential of the chitinozoan assemblages as tools for cross-continental correlation as they contained key chitinozoan index species that displayed a mixed Laurentian, Avalonian and Baltoscandic signature (Vandenbroucke et al. Reference Vandenbroucke, Verniers and Clarkson2003).
The objectives of the current paper therefore are: (i) to document the chitinozoans from the Whitehouse Subgroup in Girvan, with a particular focus on the levels that also contain P. linearis graptolites; (ii) to confirm their utility as a link between Baltic, Avalonian and Laurentian biozonal schemes; (iii) to correlate strata attributed to the P. linearis Biozone in Girvan into Cautley, using chitinozoans, in order to help alleviate some of the stratigraphic unclarities.
1. Geological setting
Girvan is located in the Midland Valley of south-central Scotland. During the Ordovician, Girvan was positioned on the southeastern margin of the palaeocontinent Laurentia (Cocks & Fortey Reference Cocks and Fortey1982; Scotese & McKerrow Reference Scotese and McKerrow1990; Ingham Reference Ingham, Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000). In the Midland Valley, the lower Palaeozoic rocks are present as a series of inliers, of which the Girvan area is the most southern one and the only one that contains Ordovician deposits as well as Silurian rocks. The other inliers, toward the north and east, exclusively contain Silurian successions (Fortey et al. Reference Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000). The Girvan inlier lies north of the Iapetus Suture, which was formed by the closure of the Iapetus Ocean when the Laurentian palaeocontinent collided with the microcontinent Avalonia (Armstrong & Owen Reference Armstrong and Owen2001). The Ordovician rocks of the Girvan area are exposed both north and south of the Water of Girvan (Fig. 1). To the north lies the Craighead Inlier as an asymmetrical anticline with Ordovician rocks in its centre. To the south lies the ‘Main Outcrop’ of Silurian rocks in a narrow strip from Saugh Hill to Straiton (Cocks & Toghill Reference Cocks and Toghill1973) with the Ordovician rocks exposed south of the ‘Main Outcrop’. Ordovician samples were collected from the exposures south of the Water of Girvan at three different locations, i.e., Penwhapple Burn (supplementary Fig. 1, available at https://doi.org/10.1017/S1755691026101054) and two coastal exposures (supplementary Figs 2, 3). This study focuses entirely on the Upper Ordovician Ardwell Farm Formation and Whitehouse Subgroup, which can be accessed in the aforementioned three locations. Of particular interest is the Penwhapple Formation as it is the only level that yielded P. linearis Biozone graptolites, but this unit was not investigated in the pilot study of Vandenbroucke et al. (Reference Vandenbroucke, Verniers and Clarkson2003). The Penwhapple Formation has now been sampled at the inland location of Penwhapple Burn, a gorge that cuts through the ‘Main Outcrop’ deposits, and from a coastal exposure at Myoch Foreshore, just south of Girvan. The Penwhapple Formation does not outcrop on the Whitehouse Shore because of structural complications (Ingham Reference Ingham, Lawson and Weedon1992).
Simplified geological map of the Palaeozoic deposits in the Girvan area (modified after Cocks & Toghill Reference Cocks and Toghill1973). Samples for this study were collected from Penwhapple Burn and Whitehouse Shore. See supplementary Figs 1–3 for detailed geological maps.
2. Material and methodology
Samples for palynology have been collected with a focus on the Whitehouse Subgroup from the Girvan area (containing, in ascending stratigraphical order, the South Shore, Three Mile, Penwhapple, Myoch, Mill and Shalloch formations). Additionally, the underlying Ardwell Farm Formation was sampled, and the results are also included here as well.
The chitinozoan data of a total of 84 samples, collected during three different sampling campaigns, are integrated to combine a chitinozoan dataset through the Ardwell Farm Formation and the Whitehouse Subgroup. An initial exploratory campaign was undertaken in 1999 by Euan Clarkson, Jacques Verniers and the senior author of this paper (Vandenbroucke et al. Reference Vandenbroucke, Verniers and Clarkson2003), testing the potential of the Ordovician–Silurian succession of the area for chitinozoan studies; this included 14 samples from the Whitehouse Foreshore and Myoch Foreshore sections that are integrated in the current study. Vandenbroucke et al.’s (Reference Vandenbroucke, Verniers and Clarkson2003) study did not contain samples from the Penwhapple Formation, which is poorly exposed on the coast. These samples are labelled TVDB 0xx. Vandenbroucke briefly revisited Girvan in 2004 and collected one sample from the P. linearis-containing Penwhapple Formation on the Myoch Foreshore (TVDB 04-016). The results were briefly discussed in Vandenbroucke’s PhD dissertation (unpub. data, Reference Vandenbroucke, Rickards and Verniers2005). Then, finally, the Caradoc–Ashgill of the Whitehouse Subgroup and the underlying Ardwell Farm Formation were extensively and exhaustively sampled in three different locations, during a third sampling campaign in 2007. The locations were sampled by two of the current authors (T.R.A.V. and J.K.I.) in Penwhapple Burn, on Whitehouse Foreshore and Myoch Foreshore (Fig. 2). The 2007 collection included a total of 78 samples, 69 of which have been analysed and used for this report; these samples are labelled TVDB 07-xxx. Their positions are indicated on detailed geological maps of J.K.I. (supplementary Figs 1–3).
The protocol for the first series of samples (TVDB 0xx) is discussed by Vandenbroucke et al. (Reference Vandenbroucke, Verniers and Clarkson2003). The samples collected in 2004 and 2007 were dissolved for palynology in several batches at the Universities of Ghent and Lille. In Ghent, the samples were processed using hydrochloric and hydrofluoric acids according to a standard palynological protocol: they were broken into 0.5–1 cm pieces and decarbonated using 20 % hydrochloric acid (HCl). They were then heated to 65°C with 150 ml of 40 % hydrofluoric acid (HF) over 48 h. Any newly formed fluorosilicates were removed using a second 17–20 % HCl treatment, at 60–65°C. Samples were neutralised and filtered at 53 μm (see De Backer et al. Reference De Backer, Day, Emsbo, McLaughlin and Vandenbroucke2024 for details). The Lille method differs slightly from the method used in Ghent (see Vandenbroucke et al. Reference Vandenbroucke, Hints, Williams, Wallis, Velleman, Kurihara, Tanaka, Komatsu, Mannik, Siveter and De Backer2019) in that 34–38 % HCl is used in the first step, 150–200 ml of 48 % HF in the second step, where samples were stirred for c. 24 h rather than heated, and that sieve mesh of 51 µm instead of 53 µm is used during filtering. The organic fraction larger than 53/51 µm of each sample was hand-picked for chitinozoans using a stereomicroscope. The picked specimens were mounted on a metal stub, gold-coated and studied with scanning electron microscopy (SEM). The specimens were photographed using a FEI Quanta 200 (University of Lille), JEOL JSM-5310LV and a Tescan TIMA3-X GMU (University of Ghent). Specimens, samples and subsamples for further research are stored at the Department of Geology at Ghent University.
3. Results
The samples are moderately productive in chitinozoans and the recovered assemblages are reasonably well-preserved and diverse. The chitinozoan specimens are generally flattened. The 84 analysed samples yield a total of 8,088 chitinozoans for which 47 different species are recognised. Twenty-three of these taxa remain in open nomenclature. Occurrences of the taxa in absolute numbers are provided as supplementary files, for the three studied sections, i.e., Penwhapple Burn (supplementary Table 1), Whitehouse Shore (supplementary Table 2) and Myoch Foreshore (supplementary Table 3). Graptolites, scolecodonts, melanosclerites and acritarchs were also observed in many samples, but are not added into the count. The dominant long-ranging species that are present throughout the Ardwell Farm Formation and Whitehouse Subgroup in all three sections include Belonechitina robusta, Spinachitina bulmani, Belonechitina america, Belonechitina micracantha, Cyathochitina campanulaeformis and Desmochitina minor. Additionally, several samples yielded biostratigraphically significant species. Lagenochitina baltica occurs from the Ardwell Farm Formation into the Mill Formation. The species is joined by Lagenochitina prussica. The next distinctive occurrence is that of Acanthochitina latebrosa in the South Shore Formation. This species carries some resemblance with Acanthochitina barbata but is less elongated and has a dense mesh-like network where individual spines can no longer be identified. One sample within the P. linearis-yielding Penwhapple Formation from the Lower Whitehouse Subgroup contained the distinctive Armoricochitina reticulifera, while two samples collected higher in the Penwhapple Formation yielded Tanuchitina bergstroemi, easily recognised by its long total vesicle length and carina, but very often found fragmented. Sample TVDB04-016 yielded the age-diagnostic Hercochitina cristata, which also occurs higher up the section but there it was only identified questionably. Finally, Acanthochitina barbata specimens occur sparsely in all three sections, within the Upper Whitehouse Subgroup. Figure 2 displays the ranges of selected chitinozoan species against the stratigraphy for the three different sections, while supplementary Figure 4 illustrates their correlation. A systematic description of key species is provided as part of the supplementary material. Figures 3–6 display key species from the Girvan succession.
Illustrations of chitinozoans from Girvan, with indications of length of the scale bar. (a) Lagenochitina baltica, sample TVDB 07-126, 100 µm; (b) L. baltica, sample TVDB 07-061, 100 µm; (c) L. baltica, sample TVDB 07-098, 100 µm; (d) detail of (c) L. baltica, 10 µm; (e) Lagenochitina prussica, sample TVDB 07-077, 100 µm; (f) L. prussica, sample TVDB 07-054, 100 µm; (g) L. prussica, sample TVDB 07-074, 100 µm; (h) L. prussica, sample TVDB 07-126, 100 µm; (i) Armoricochitina reticulifera, sample TVDB 07-088, 100 µm; (j) A. reticulifera, sample TVDB 07-088, 100 µm; (k) A. reticulifera, sample TVDB 07-088, 100 µm; (l) A. reticulifera, sample TVDB 07-088, 100 µm; (m) Fungochitina ?spinifera, sample TVDB 07-096, 100 µm; (n) F. ?spinifera, sample TVDB 07-096, 100 µm; (o) Saharochitina ?fungiformis, sample TVDB 07-089, 100 µm; (p) S. ?fungiformis, sample TVDB 07-089, 100 µm.
Illustrations of chitinozoans from Girvan, with indications of length of the scale bar. (a) Tanuchitina bergstroemi, sample TVDB 07-089, 500 µm; (b) T. bergstroemi, sample TVDB 07-089, 500 µm; (c) T. bergstroemi, sample TVDB 07-089, 500 µm; (d) Conochitina ?incerta, sample TVDB 07-063, 500 µm; (e) C. ?minnesotensis, sample TVDB 07-076, 200 µm; (f) Rhabdochitina gracilis, sample TVDB 07-065, 200 µm; (g) Conochitina ?elegans, sample TVDB 07-073, 200 µm; (h) Spinachitina cf. coronata, sample TVDB 07-057, 100 µm; (i) Spinachitina coronata, sample TVDB 07-109A, 100 µm; (j) Acanthochitina barbata, sample TVDB 07-101, 100 µm; (k) A. barbata, sample TVDB 07-101, 100 µm; (l) A. barbata, sample TVDB 07-101, 100 µm; (m) Acanthochitina latebrosa, sample TVDB 07-073, 100 µm; (n) A. latebrosa, sample TVDB 07-074, 100 µm.
Illustrations of chitinozoans from Girvan, with indications of length of the scale bar. (a) Spinachitina bulmani, sample TVDB 07-081, 100 µm; (b) Belonechitina micracantha, sample TVDB 07-089, 100 µm; (c) Belonechitina americana, sample TVDB 07-101, 100 µm; (d) B. americana, sample TVDB 07-085, 100 µm; (e) Belonechitina robusta, sample TVDB 07-102, 100 µm; (f) B. robusta, sample TVDB 07-102, 100 µm; (g) Belonechitina cf. robusta, sample TVDB 07-113, 100 µm; (h) Spinachitina cf. cervicornis, sample TVDB 07-057, 100 µm; (i) S. cf. cervicornis, sample TVDB 07-073, 100 µm; (j) Spinachitina spp., sample TVDB 07-059, 100 µm; (k) Spinachitina spp., sample TVDB 07-055, 100 µm; (l) Spinachitina spp., sample TVDB 07-073, 100 µm; (m) Belonechitina wesenbergensis, sample TVDB 07-057, 100 µm; (n) B. wesenbergensis, sample TVDB 07-085, 100 µm; (o) Hercochitina sp. A, sample TVDB 07-110, 100 µm; (p) Hercochitina sp. B, sample TVDB 07-113, 100 µm; (q) Hercochitina sp. C, sample TVDB 07-074, 100 µm; (r) Hercochitina cristata, sample TVDB 04-016, 100 µm; (s) H. cristata, sample TVDB 04-016, 100 µm; (t) Hercochitina cf. cristata, sample TVDB 07-112, 100 µm.
Illustrations of chitinozoans from Girvan, with indications of length of the scale bar. (a) Ancyrochitina cf. merga, sample TVDB 07-101, 50 µm; (b) Ancyrochitina cf. merga, sample TVDB 07-101, 50 µm; (c) Spinachitina spp., sample TVDB 07-057, 100 µm; (d) Spinachitina spp., sample TVDB 07-101, 100 µm; (e) Angochitina communis, sample TVDB 07-118, 50 µm; (f) A. communis, sample TVDB 07-118, 50 µm; (g) Spinachitina spp., sample TVDB 07-089, 100 µm; (h) Belonechitina hirsuta, sample TVDB 07-102, 50 µm; (i) B. hirsuta, sample TVDB 07-102, 50 µm; (j) Kalochitina multispinata, sample TVDB 07-088, 50 µm; (k) K. multispinata, sample TVDB 07-088, 50 µm; (l) Belonechitina ?hirsuta, sample TVDB 07-085, 50 µm; (m) B. ?hirsuta, sample TVDB 07-085, 50 µm; (n) Eisenackitina sp. A, sample TVDB 07-118, 50 µm; (o) Eisenackitina sp. B, sample TVDB 07-118, 50 µm; (p) Eisenackitina sp. C, sample TVDB 07-118, 50 µm; (q) Eisenackitina sp. C, sample TVDB 07-118, 50 µm; (r) Eisenackitina sp. D, sample TVDB 07-101, 50 µm; (s) Angochitina cf. capillata, sample TVDB 07-089, 100 µm.
4. Biostratigraphy
4.1. Chitinozoan biozonation and correlation
Chitinozoan biozonations have historically been developed for separate palaeocontinental regions and provinces (Webby et al. Reference Webby, Cooper, Bergström, Paris, Webby, Droser, Paris and Percival2004), including one originally defined by Nõlvak & Grahn (Reference Nõlvak and Grahn1993) in Baltoscandia. This biozonation was the basis for, and has been integrated into, an Upper Ordovician Avalonian biozonation, with some additions (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005, Reference Vandenbroucke, Recourt, Nõlvak and Nielsen2013; Vandenbroucke Reference Vandenbroucke2008a), de facto creating a hybrid Avalonian–Baltoscandian Upper Ordovician biozonation. This integrated biozonal scheme of Vandenbroucke (Reference Vandenbroucke2008a) can be recognised in the assemblages from Girvan, in surprising detail given its peri-Laurentian position (Fig. 7). The occurrences of selected chitinozoan taxa are compared to their occurrences in several Laurentian, Avalonian and Baltoscandian locations (Figs 8, 9).
Composite log of the Girvan area sections, with identified biozones. For litholog colour code see supplementary Figs 1–3. Grey horizontal lines are sampled horizons. Numbered chitinozoan ranges: (a) Hercochitina spinetum, (b) Cyathochitina campanulaeformis, (c) Desmochitina minor, (d) Belonechitina hirsuta, (e) Belonechitina robusta, (f) Spinachitina bulmani, (g) Belonechitina americana, (h) Belonechitina micracantha, (i) Belonechitina wesenbergensis, (j) Conochitina ?elegans, (k) Conochitina ?incerta, (l) Conochitina ?minnesotensis, (m) Cyathochitina kuckersiana, (n) Rhabdochitina gracilis, (o) Spinachitina cf. cervicornis, (p) Spinachitina cf. coronata, Desmochitina erinacea, (q) Rhabdochitina magna, (r) Kalochitina multispinata, (s) Angochitina communis, (t) Spinachitina coronata, (u) Desmochitina lata.
Chronostratigraphic correlation framework between Girvan, Bornholm and the Cautley district. This diagram highlights graptolite correlations using the original position of the Pleurograptus linearis Biozone by Ingham (Reference Ingham1966) for Cautley as well as its extension by Rickards (Reference Rickards2002) alongside correlations using the Avalonian–Baltoscandian Fungochitina spinifera and Tanuchitina bergstroemi chitinozoan biozones.
Total ranges of selected chitinozoan taxa, in a series of key sections, plotted against the British graptolite biozonation (Zalasiewicz et al. Reference Zalasiewicz, Taylor, Rushton, Loydell, Rickards and Williams2009) and the Baltoscandian/Avalonian Biozonation (Vandenbroucke Reference Vandenbroucke2008a).
In Girvan, the Fungochitina spinifera Biozone, originally defined by Nõlvak & Grahn (Reference Nõlvak and Grahn1993) as corresponding to the entire range of their index species Fungochitina spinifera, is recognised here based on the first occurrences of L. baltica and L. prussica, as suggested by Vandenbroucke (Reference Vandenbroucke2008a). Their occurrences are a proxy for a level slightly above the base of the biozone, in the absence of the index fossil (only rare and very doubtful occurrences of Fungochitina spinifera and the smooth equivalent Saharachitina fungiformis were documented in Girvan; Fig. 7). The F. spinifera Biozone incorporates the Caradoc–Ashgill boundary at Pus Gill in northern England, where the biostratigraphy is directly tied to the chronostratigraphy (Vandenbroucke et al Reference Vandenbroucke, Rickards and Verniers2005). The recognition of the local A. latebrosa–A. onniensis Bio(sub)zone (Vandenbroucke Reference Vandenbroucke2008a), based on the occurrence of A. latebrosa, allows correlation with the Onnian of the Onny Valley (Vandenbroucke et al. Reference Vandenbroucke, Ancilletta, Fortey and Verniers2009) and the Whitland area of southwest Wales (Vandenbroucke et al. Reference Vandenbroucke, Williams, Zalasiewicz, Davies and Waters2008). Vandenbroucke (Reference Vandenbroucke2008a) provided two alternatives, i.e., that the A. latebrosa–A. onniensis Bio(sub)zone sits below the F. spinifera Biozone or is an equivalent to the lower part of the F. spinifera Biozone. Based on our Girvan data, the A. latebrosa–A. onniensis Bio(sub)zone may be best considered a lateral equivalent of (part of) the F. spinifera Biozone, and we choose to refer to it as a bio(sub)zone to emphasise our doubt as to its (sub?)biozonal status. The A. reticulifera Subzone was defined by Nõlvak & Grahn (Reference Nõlvak and Grahn1993) as corresponding to the entire range of its eponymous index species. This definition of the biosubzone was also applied in the Avalonian biozonation (Vandenbroucke Reference Vandenbroucke2008a). The precise position of the subzone within the F. spinifera Biozone remains uncertain and can unfortunately not be untangled in the Girvan area, as the species was only present in one sample of the Penwhapple Formation. The next sampled horizon already contains T. bergstroemi, the index species of the biozone that succeeds the F. spinifera Biozone. This would position the A. reticulifera Subzone near the top of the F. spinifera Biozone in Girvan, which is slightly different from its definition as a subzone halfway through the F. spinifera Biozone by Nõlvak & Grahn (Reference Nõlvak and Grahn1993) and in several observations elsewhere in the UK (Vandenbroucke Reference Vandenbroucke2008a). The A. reticulifera Subzone of the F. spinifera Biozone has been observed to co-occur with graptolites of the Dicranograptus clingani–Dicellograptus morrisi Biozone, both in Dob’s Linn, Scotland (Verniers & Vandenbroucke Reference Verniers and Vandenbroucke2006) and in the Cardigan area of southwest Wales (Vandenbroucke et al. Reference Vandenbroucke, Williams, Zalasiewicz, Davies and Waters2008). However, in the Whitland area of southwest Wales they are associated with graptolites of the Normalograptus proliferation interval (Vandenbroucke et al. Reference Vandenbroucke, Williams, Zalasiewicz, Davies and Waters2008). The initial proliferation of the Normalograptus graptolites had been previously tentatively correlated with the base of the P. linearis Biozone by Zalasiewicz et al. (Reference Zalasiewicz, Rushton and Owen1995). The presence of A. reticulifera has also been observed in the Greenscoe section of the southern Lake District (Van Nieuwenhove et al. Reference Van Nieuwenhove, Vandenbroucke and Verniers2006).
The T. bergstroemi Biozone is identified in Girvan based on the occurrence of its index species, complemented by the recoveries of several A. barbata specimens that allowed the recognition of the Baltoscandian A. barbata Subzone. Accessory chitinozoans of the T. bergstroemi Biozone included Belonechitina americana, Belonechitina wesenbergensis, Conochitina incerta, L. baltica and L. prussica. So far, the A. barbata Subzone has not been recognised in Avalonia. But in Baltoscandia, A. barbata is an excellent guide fossil as the species is short-ranging and occurs in several Baltoscandian sections, in Sweden and Estonia (Laufeld Reference Laufeld1967). The species has been identified on the Laurentian Anticosti Island of Quebec as well, where it is included as an accessory characteristic chitinozoan species, in combination with B. americana and L. baltica, in the Laurentian Conochitina vaurealensis Biozone. In Girvan, the chitinozoans of the T. bergstroemi Biozone and its subzone co-occur with graptolites of the upper part of the P. linearis Biozone and at least part of the D. complanatus Biozone.
It is remarkable that in the inland section of Penwhapple Burn only questionable identifications of L. baltica and L. prussica are made in the Penwhapple Formation and the underlying formations, while the two species appear in large numbers in the Whitehouse Shore coastal exposures, both below and above the Penwhapple Formation, which does not occur in the coastal outcrop for structural reasons (Ingham Reference Ingham, Lawson and Weedon1992). However, in the Penwhapple Formation of Penwhapple Burn, T. bergstroemi has been observed in two of the samples. This species is very often poorly preserved and frequently only recovered as fragmented vesicles (Vandenbroucke et al. Reference Vandenbroucke, Williams, Zalasiewicz, Davies and Waters2008). The majority of the T. bergstroemi specimens from the Penwhapple Formation, however, were obtained in their entirety (except for the carina). This excludes a preservation issue as the accountable factor for the lack of (the more robust) L. baltica and L. prussica in the Three Mile and Penwhapple formations of Penwhapple Burn. Palaeo-ecological reasons linked to the deep-water nature of the Penwhapple Formation also seem insufficient to explain the absence of L. baltica, given that, in the Fjäcka Shales of Sweden, T. bergstroemi and L. baltica have been recorded together (within the same sample) by Laufeld (Reference Laufeld1967). Both the Fjäcka Shales and the Penwhapple Formation can be described as relatively deep-water setting deposits. In addition, the overlying Myoch Formation has been described as one of the deepest water deposits of the entire Ordovician in Girvan (Ingham Reference Ingham, Lawson and Weedon1992; Ingham Reference Ingham, Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000). Within the Myoch Formation, L. baltica reoccurs, but T. bergstroemi was not recovered above the Penwhapple Formation. The presence of the A. barbata index fossil in the Myoch and Mill formations indicates that the top of the T. bergstroemi Biozone has not been reached yet. A more precise position of top of the T. bergstroemi Biozone is likely to emerge from a more thorough investigation of the Shalloch Formation (which was not the main target of this study) and the overlying Drummuck Subgroup.
Given Girvan’s palaeogeographic location on the southern margin of Laurentia, integration of the assemblages into the Laurentian biozonation scheme, which was developed by Achab (Reference Achab1989) in Quebec and western Newfoundland, seemed logical. However, correlation with Laurentian biozones was only possible at two levels, and one of these even only provisionally. In between these two biozones, several other biozones were originally defined in Canada, but indicators of these intervening biozones were not observed at all in Girvan. The Laurentian biozonation scheme is placed alongside the composite log of the Ordovician Whitehouse Subgroup and Ardwell Farm Formation in Figure 7. The Conochitina pygmaea–Hercochitina cristata Biozone was defined as a concurrent-range zone by Achab (Reference Achab1989) based on the presence of both index fossils in the St Lawrence Lowlands and secondarily on Anticosti Island, Quebec, in the Macasty Formation. In Girvan, H. cristata was recovered from the Myoch Foreshore in a sample (TVDB 04-016) collected in the Penwhapple Formation and questionable identifications were made in the Mill Formation. C. pygmaea, however, was not identified. H. cristata itself has a restricted range within the upper part of the Climacograptus pygmaeus graptolite Biozone in the Quebec and Anticosti Island area (Achab Reference Achab1987). The pygmaeus graptolite Biozone has a local Cincinnatian–Maysvillian age. The base of the C. pygmaeus graptolite Biozone has been tentatively correlated with the base of the P. linearis Biozone (Achab Reference Achab1987; Webby et al. Reference Webby, Cooper, Bergström, Paris, Webby, Droser, Paris and Percival2004). In Girvan, the recovery of H. cristata also coincides with a part of the P. linearis Biozone, confirming the suggested correlation between the C. pygmaeus and P. linearis graptolite biozones.
Equivalents of the total range chitinozoan biozones that overlie the C. pygmaea–H. cristata Biozone in Laurentia (the Conochitina sp. 2 Biozone, the Cyathochitina hyalophrys Biozone and the Conochitina senta Biozone) were not recognised in the Girvan sections due to the absence of the biozonal index fossils. Some relatively long-ranging species that occur in these biozones, such as S. bulmani, A. merga and K. multispinata, have been observed in Girvan, but this is insufficient to identify correlative levels of the biozones. The Conochitina vaurealensis partial-range Biozone was defined in the Vaureal Formation of Anticosti Island, Quebec by Achab (Reference Achab1989), characterised by the appearance of the index fossil Conochitina vaurealensis up to the lowest occurrence of T. anticostiensis, the index fossil of the overlying biozone (also see McLaughlin et al. Reference McLaughlin, Emsbo, Desrochers, Bancroft, Brett, Riva, Premo, Neymark, Achab, Asselin and Emmons2016). C. vaurealensis does not occur in Girvan. Other characteristic chitinozoans are B. americana, A. barbata and L. baltica. These three chitinozoan species occur together in Girvan in the Upper Whitehouse Subgroup (the Myoch to Shalloch Formation) and allow a tentative identification of the C. vaurealensis Biozone. As B. americana and L. baltica have their first appearance datums well below that of A. barbata, the biozonal equivalent at Girvan is mostly suggested by the occurrence of A. barbata. The ?C. vaurealensis Biozone would correspond to the upper P. linearis Biozone and the entire D. complanatus Biozone at Girvan, while in Quebec the biozone forms a major part of the D. complanatus Biozone and has a local Cincinnatian–Richmondian age. Using A. barbata as a characteristic chitinozoan species, the C. vaurealensis Biozone can be correlated with the A. barbata Subzone of the T. bergstroemi Biozone.
If our questionable identifications of H. ?cristata in the Mill and Shalloch formations (samples TVDB 07-113, -114, -115, -112, -109A, -106, -103, -104) can be used to extend correlatives of the C. pygmaea–H. cristata Biozone upwards, the biozone would also correspond to strata bearing D. complanatus Biozone graptolites at Girvan and overlap equivalents of the overlying ?C. vaurealensis chitinozoan Biozone (and the three intercalated Conochitina sp. 2, C. hyalophrys, C. senta Biozones).
4.2. Integrated graptolite-chitinozoan biostratigraphic correlation between Girvan, Bornholm and Cautley
In Girvan, chitinozoans from the T. bergstroemi Biozone co-occur with graptolites of the upper part of the P. linearis Biozone and part of the D. complanatus Biozone. The co-occurrence of P. linearis graptolites with T. bergstroemi has been recorded elsewhere, for example, both in the Fjäcka Shales in Sweden (Laufeld Reference Laufeld1967) and in Bornholm, Denmark (Vandenbroucke et al. Reference Vandenbroucke, Recourt, Nõlvak and Nielsen2013). In the type Cautleyan-Rawtheyan, the T. bergstroemi Biozone (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005) is restricted to the Cautley stage, below the P. linearis Biozone which is in the Rawtheyan stage and ‘the only reliable graptolite dating of the type Ashgill’ (Rickards (Reference Rickards2002), 1), although Rickards (Reference Rickards2002) recognises that ‘it is possible that the underlying Cautley Stage is also of linearis Biozone age’, i.e., that it belongs to the P. linearis chronozone. This package includes all strata that are equivalent in age to the biozone regardless of the presence of its diagnostic fossils (Salvador Reference Salvador2013, 83), and is here used as shorthand to refer to the rock succession of the Onnian to Cautleyan stage at Cautley that do not have the typical P. linearis Biozone graptolite fauna, but that are of the same age as the P. linearis Biozone rocks in Girvan. A correlation between the chitinozoan and graptolite biozonations of the Girvan area in Scotland, the Cautley district in northern England and Bornholm, Denmark around the P. linearis Biozone is shown in Figure 8. The following paragraphs describe the correlation between these areas in more detail.
The British stages of the Ashgill series are defined in the Cautley district, so in this area the British chronostratigraphic units of the Ashgill are considered to be a given fact. Originally, graptolites of the P. linearis Biozone were considered to straddle the C/A boundary in Girvan (Ingham Reference Ingham1966; Ingham & Wright Reference Ingham and Wright1970), given that these strata contained the same shelly fauna as the C/A boundary strata in Cautley. This resulted in an Onnian to Pusgillian chronostratigraphic position for the P. linearis and the Penwhapple Formation. The boundaries of the P. linearis Biozone are however unclear. The P. linearis may extend downwards to include the South Shore Formation as this stratigraphic level was described by Laufeld (Reference Laufeld1971) as ‘limestone member of the Lower Whitehouse Formation of “P. linearis’ age”’, but it is unclear if this formation actually yielded any graptolites. Likewise, this biostratigraphic unit ranges to the top of the Myoch Formation, although this formation does not contain graptolitic fauna. The lower part of the Mill Formation contains a mixed assemblage that is indicative of the P. linearis Biozone, but already includes the index fossil D. complanatus, the biozone overlying the P. linearis Biozone (Ingham Reference Ingham, Lawson and Weedon1992).
The chitinozoan-based correlations between Bornholm and Girvan are fairly straightforward, using the F. spinifera and the T. bergstroemi biozones that were identified in both areas (Fig. 8). In Bornholm and in Girvan, the F. spinifera Biozone was identified based on the first occurrences of L. baltica (and of L. prussica), instead of the index fossil, and these are expected to closely equate to the base of the biozone. The biozone in Girvan ranges up to the lower half of the P. linearis Biozone. This is the case in Bornholm as well. Both in Bornholm and in Girvan, the top of the T. bergstroemi has not been identified with confidence. In Bornholm, index fossil T. bergstroemi was present in the highest sampled interval, so it probably ranges above this level, while in Girvan the chitinozoan assemblages above the A. barbata Subzone did not contain elements of either the T. bergstroemi Biozone or of the overlying C. rugata Biozone. The correlation between these regions is therefore most solid around the base of the T. bergstroemi Biozone, but less tight at its top.
The Pusgillian and Cautleyan stages in the Cautley district yielded the upper part of the F. spinifera Biozone (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005). The biozone therefore includes the C/A boundary and has a chronostratigraphic position that is Onnian to lower Cautleyan, additionally based on correlations with Pus Gill in northern England (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005). The F. spinifera Biozone is calibrated with the top of the D. clingani Biozone and the lower part of the P. linearis Biozone in South Wales (Vandenbroucke et al. Reference Vandenbroucke, Williams, Zalasiewicz, Davies and Waters2008) and Bornholm (Vandenbroucke et al. Reference Vandenbroucke, Recourt, Nõlvak and Nielsen2013). Based on observations of the index species in the Cautley district, the T. bergstroemi Biozone occupies a Cautleyan chronostratigraphic position, although the index fossil itself ranges into the Rawtheyan, where it continues in the overlying C. rugata, S. fossensis and B. umbilicata chitinozoan Biozones (Vandenbroucke et al. Reference Vandenbroucke, Rickards and Verniers2005). The T. bergstroemi Biozone in Girvan and Bornholm equates with the upper part of the P. linearis Biozone and at least part of the D. complanatus Biozone.
In summary, taking our new chitinozoan correlations at face value, the Fungochitina spinifera Biozone suggests correlation of the upper Onnian, Pusgillian and lower Cautleyan stages in the Cautley district with sections from the upper D. clingani Biozone into the lower P. linearis Biozone on Bornholm and at Girvan. The base of the Tanuchitina bergstroemi Biozone in the type Cautleyan Stage correlates with a level within the P. linearis Biozone on Bornholm and at Girvan (i.e., within the Penwhapple Formation in Girvan). The top of the T. bergstroemi Biozone is in the upper part of the Cautleyan Stage at Cautley, but is poorly constrained on Bornholm and at Girvan. The overlying chitinozoan biozones at Cautley have not been identified in the other two areas but are possibly represented above the sampled sections; further work is needed to establish if this might be the case. The chitinozoan correlations reinforce the argument that the P. linearis Biozone extends higher at Cautley than its original restriction there to the upper Onnian–lower Pusgillian stages (e.g., Ingham Reference Ingham1966). The occurrence of P. linearis Biozone graptolites recorded by Rickards (Reference Rickards2002) at Cautley with chitinozoans of the rugata, fossensis and umbilicate biozones may indicate correlation with a level above the P. linearis Biozone at Girvan and on Bornholm. This suggests in turn that the Rawtheyan part of the P. linearis Biozone at Cautley may be laterally equivalent to the D. complanatus and anceps graptolite biozones in the other two areas, reinforcing the correlations depicted by Zalasiewicz et al. (Reference Zalasiewicz, Taylor, Rushton, Loydell, Rickards and Williams2009, fig. 1).
5. Palaeobiogeography
While Girvan was situated on the southeastern margin of Laurentia, the chitinozoan faunal assemblages are characterised by a mixed Baltoscandian/Avalonian signature instead of the expected Laurentian aspect. However, the expression and recognition of this palaeobiogeographic signature seems to be related to the presence of index species of the Baltoscandian/Avalonian biozonal scheme, the absence of Laurentian index species and the absence of Laurentian endemics, rather than the presence of Baltoscandian/Avalonian endemics. Most species recognised here as having a ‘Baltic’ signature, e.g., L. baltica and A. barbata, were also recovered in Laurentian Canada. They even occur in northern Gondwana, though, intriguingly, higher in the chronostratigraphy (Paris et al. Reference Paris, Verniers, Achab, Albani, Ancilletta, Asselin, Xiaohong, Fatka, Grahn, Molyneux, Nõlvak, Samuelsson, Sennikov, Soufiane, Xiaofeng and Winchester-Seeto1999).
The chitinozoan assemblage from Girvan includes two chitinozoan species (H. cristata and H. spinetum) that occur solely in the Laurentian palaeocontinent, three species (including A. latebrosa and K. multispinata) are limited to both Laurentia and Avalonia, while three other species (including A. barbata and A. capillata) have not been recorded in Avalonia, but are present in Laurentia and Baltica. Twenty-six of the species, however, occur on all three land masses. Several of these species (including A. barbata, L. baltica and S. fungiformis) were also documented in northern Gondwana. Such considerations underline that notions of what we consider ‘typical’ Baltic or Laurentian species are mainly rooted in where such species were originally described (e.g., L. baltica), or where they have been included, with a certain degree of formality, in a biozonal scheme (e.g., A. barbata), rather than their currently known spatial distribution patterns.
The similarity in faunal assemblages between the eastern Laurentian palaeocontinent and both Baltica and Avalonia can be explained by their proximity alone. Additionally, the simulations of the past ocean circulation in the Ordovician carried out by Pohl et al. (Reference Pohl, Nardin, Vandenbroucke and Donnadieu2016) using the general circulation model FOAM show the development of an anticlockwise circulation between these three land masses in the Iapetus and Tornquist oceans during the Middle Ordovician. This ensures an efficient exchange of these water masses. Following the northward migration of Avalonia and Baltica from the Middle to the Late Ordovician, these anticlockwise circulations ceased to exist. The distance between eastern Laurentia, north Avalonia and western Baltica decreased and in between these palaeocontinents a meridional current flowed from the equator towards the Southern Hemisphere Iapetus Ocean, as the Iapetus Current (Pohl et al. Reference Pohl, Nardin, Vandenbroucke and Donnadieu2016). The increasingly cosmopolitan composition of faunal assemblages, upwards through the Girvan succession, was also noted by Ingham (Reference Ingham, Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000). While the shelly fauna exhibits an American (Laurentian) aspect until the Llanvirn and Caradoc, rather than a British Avalonian signature, by the late Caradoc and Ashgill times Baltic and Anglo-Welsh elements make their appearance in the deep-water faunal assemblages. The cyclopygid trilobites recovered from the Myoch Formation even display a Gondwanan signature (Ingham Reference Ingham, Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000).
6. Conclusions
The chitinozoan study of 84 samples collected in the Girvan Whitehouse Subgroup and the underlying Ardwell Farm Formation yielded 8,088 chitinozoans and led to the development of a solid chitinozoan biostratigraphic framework. The characteristic chitinozoans from Girvan have a predominantly Baltoscandian/Avalonian signature, based on the presence of the index species of the Baltoscandian/Avalonian biozonation. However, data also suggest that an objective approach as to which species are considered ‘typically’ Baltoscandian/Avalonian/Laurentian is required. In Girvan, the Fungochitina spinifera and Tanuchitina bergstroemi biozones and the Acanthochitina latebrosa–Ancyrochitina onniensis, Armoricochitina reticulifera and Acanthochitina barbata subzones were recognised.
Using the F. spinifera and T. bergstroemi biozones, a chitinozoan-based correlation was made between the three different locations that were separated by the Iapetus and Tornquist oceans at the time. Taken at face value, the ‘British’ chronostratigraphic position (derived from the type area at Cautley) of the P. linearis Biozone in Girvan would be lower Pusgillian to mid-Cautleyan (lower Ashgill); this is somewhat higher than its original position suggested by Ingham (Reference Ingham1966) but would not reach into the higher Rawtheyan positions of Rickards (Reference Rickards2002). As such, our data confirm that the Cautleyan in its type area at Cautley, and the Pusgillian, are part of the P. linearis chronozone as suggested by Rickards (i.e., of similar age to the P. linearis Biozone elsewhere, but without P. linearis Biozone fossils in Cautley).
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1755691026101054
Acknowledgements
This contribution is part of series of papers celebrating the impressive career of Prof. Euan Clarkson. Euan first introduced the senior author (T.R.A.V.) to Scottish geology, and Girvan in particular, as the field area for my master degree research project; this launched me into the field of Palaeozoic stratigraphy and I am still grateful for Euan’s guidance and mentorship. As my own students (C.M.W.V.L .and I.V.) have now taken over the baton, using the area for their own projects and rounding the circle, this paper seemed like an appropriate way to honour Euan and his contributions to our science. We also thank the Palaeontological Association for financial support through an Undergraduate Research Bursary (PA-UB201602) to C.M.W.V.L. I.V. and T.R.A.V. were funded by the Research Foundation, Flanders (FWO) through research grant G038722N (Monsters of the Apocalypse). We thank Sabine Van Cauwenberghe (Ghent University) and Laurence Debeauvais (Lille University) for laboratory preparation, and Philippe Recourt (Lille University) for scanning electron microscopy. SEM images were partly generated using research infrastructure funded through FWO grant I013118N. Neil Clark is acknowledged for his help with J.K.I.’s geological maps. We thank Alan Owen for the invitation to write this paper and we are highly indebted to the very insightful reviews of two anonymous colleagues.
Competing interests
The authors declare there are no competing interests.
