North Berwick Fossil Hunting

1897 – The volcanic cliffs illustrated
Archibald Geikie illustrated the eastern side of Oxroad Bay in his account of Britain’s ancient volcanoes. The illustration provides an early record of the cliff geology, rather than a dated discovery of the fossil flora.

1930 – Gordon discovers the classic cliff flora
W. T. Gordon found fossil plants in the green ash of the cliff east-south-east of Tantallon Castle. His block came from above an angular unconformity within an over-steepened channel, in the exposure later labelled A. Removed in pieces, it became the basis of decades of anatomical research.

1938 – Tetrastichia bupatides named
Gordon formally described Tetrastichia bupatides from the classic A material. Its stems and associated organs subsequently became important to attempts to reconstruct an early seed plant; 1938 is the description date, rather than the date of the original find.

1941 – Salpingostoma dasu described
Gordon named the ovule Salpingostoma dasu from the same classic assemblage. Later work revised aspects of its interpretation, illustrating how repeated study of historic material can change ideas about early seed-plant reproduction.

1956–1957 – Anatomical research and collecting revived
The development of acetate-peel preparation helped stimulate Barnard’s renewed study of Gordon’s block from 1956. A. G. Long visited the locality in 1957. This revival led to further work on stems and reproductive organs, without making every later publication a new field discovery.

1959 – Eosperma oxroadense named
Barnard described Eosperma oxroadense from 60 ovules in Gordon’s A material. These small reproductive structures extended knowledge of the seed plants represented in the anatomically preserved flora.

1960 – Cupules of Calathospermum fimbriatum described
Barnard named Calathospermum fimbriatum from cupules in the classic A assemblage. Their organisation became central to later attempts to connect isolated reproductive organs with the stems and fronds of whole plants, although several early proposed connections remained uncertain.

1962 – Root anatomy investigated
Barnard’s revision of Amyelon included the roots named Amyelon bovius, initially attributed to Oxroad. A later reassessment suggested that the type block’s matrix indicated Castleton instead. Similar roots do occur at Oxroad, so the local record and the uncertain type provenance must be distinguished.

1965 – Oxroadia gracilis named
Alvin described Oxroadia gracilis and its associated megaspore, originally named Triletes subpalaeocristatus, from a loose shore block attributed to the Oxroad cliffs. Later in-situ specimens independently confirmed the plant’s local occurrence. Subsequent research revised the original interpretation of its growth form.

1966 – A second distinctive megaspore described
Alvin named Triletes pannosus, subsequently known as Setispora pannosa, from material recorded at exposure B. Later specimens linked this spore to the reproductive structures of Oxroadia conferta; the spore name does not represent an additional whole plant.

1969 – Megaspore development examined
Pettitt studied megaspores and ovules from B supplied by Alvin. The recognition of three abortive megaspores attached to a functional one preserved evidence of a developmental tetrad, adding detail beyond the outward shape of the fossil spores.

Early 1970s–1976 – A larger block collection assembled
Long collected and prepared many loose foreshore blocks; 276 had been cut and peeled by 1976. The programme greatly increased the material available for study, but later investigations showed why a loose block’s origin must be tested rather than assuming that every shore specimen came from the local cliffs.

1973 – Buteoxylon and Tantallosperma described
Barnard and Long named the stem Buteoxylon gordonianum and seeds of Tantallosperma setigera from in-situ A material. Their association in the same rock was significant, but did not by itself establish that they belonged to one whole plant.

1975 – Triradioxylon primaevum established
Barnard and Long described Triradioxylon primaevum using Gordon’s A material and loose blocks probably derived from A. Work on its stems and petioles helped refine the interpretation of early seed plants; the wider study also included material from elsewhere in south-east Scotland.

1976 – Calathopteris heterophylla named
Long described a stem bearing two kinds of petiole as Calathopteris heterophylla. Its loose greenish-grey tuff block was attributed to A, rather than collected in place. Proposed connections with reproductive organs remained hypotheses, not demonstrated whole plants.

1977–1979 – Cupule organisation investigated
Long’s studies of ovulate cupules, then discussed under Hydrasperma, developed interpretations of their branching and reproductive organisation. Later revision separated the cupules from the ovules taxonomically. A rare bisexual specimen was interpreted as a developmental anomaly, not evidence of flowering plants.

1984 – A new foreshore assemblage discovered
Richard Bateman discovered the foreshore plant deposit later labelled C. A programme involving Bateman, Andrew Scott and Gar Rothwell mapped and sampled multiple exposures, replacing the idea of Oxroad as essentially one cliff assemblage with a much more varied picture of its flora and environments.

1984 – Oxroadopteris parvus described
Long named Oxroadopteris parvus from a stem with petioles in a loose block attributed to A. This anatomical discovery is separate from the discovery of C in the same year, and its inferred block provenance remains part of the record.

1984 – A late Tournaisian age model developed
Scott and colleagues used miospores to support a CM-zone, late Tournaisian interpretation. That became an influential age framework for the flora. Later direct sampling at Oxroad supplied evidence for a Viséan interpretation, so the earlier result is an important historical model rather than the final word on every plant bed.

1988 – Field and palaeobotanical synthesis in a doctoral thesis
Bateman’s doctoral work brought together detailed field investigation and palaeobotanical research, providing a foundation for subsequent descriptions and reconstructions. Taxa formally named in later papers retain those later naming dates.

1989 – Pullaritheca longii separates cupules from ovules
Rothwell and Wight established Pullaritheca longii for Oxroad cupules previously included in interpretations of Hydrasperma. Distinguishing the surrounding cupule from the ovule improved the taxonomy and the evidence used in whole-plant reconstruction.

1990 – The flora and volcanic environments reassessed
Bateman and Rothwell’s floristic study and Bateman and Scott’s environmental study documented eight exposures at four probable correlated levels. The floristic inventory distinguished 43 anatomical organ categories and 19 adpressed organ-species, representing at least 11 whole plants. Together the studies linked differences in vegetation and preservation to a changing volcanic, river and lake landscape.

1991 – A heterosporous sphenopsid reconstructed
Bateman described Protocalamites longii and the reproductive structures Protocalamostachys farringtonii from C. Their reconstruction provided early evidence for heterospory in the sphenopsid line, using rooted vegetative material and reproductive remains rather than treating each isolated organ as a separate plant.

1992 – Oxroadia reconstructed and a second species named
Bateman revised Oxroadia gracilis and named Oxroadia conferta, clarifying the architecture of these small rhizomorphic lycopsids. The work replaced an overly simple herbaceous interpretation and distinguished the C plant from O. gracilis, particularly well represented in D.

1992 – Stamnostoma oliveri named
Rothwell and Scott described Stamnostoma oliveri and its ovulate cupule systems from Oxroad. The revision clarified earlier records discussed under other seed names; those older identifications should not be counted again as additional species.

1995 – The locality’s conservation significance synthesised
Cleal and Thomas’s Geological Conservation Review assessment drew together Oxroad’s international palaeobotanical importance, the differences between its assemblages and the need to exclude misleading transported records. The assessment is a scientific conservation milestone, not the date of the first fossil discovery or evidence of collecting permission.

2007 – The rare lycopsid Hestia eremosa named
Bateman, Kenrick and Rothwell described Hestia eremosa from rare axes represented in A, C and D. The material included historic collections and specimens recovered during the later research programme. These were not three new field finds made in 2007.

2011 – SSSI notification reviewed
The reviewed notification for the Firth of Forth SSSI, dated 29 March 2011, explicitly recognised Oxroad Bay’s fossil-flora interest within the protected coast. This is a documented notification-review date, not a claim that protection began for the first time in 2011.

2012 – Tetrastichia anatomy reassessed
Dunn and Rothwell studied 59 stems together with fronds and roots. Variation in the vascular system, leaf arrangement and secondary wood showed that the plant was anatomically more variable than older narrow diagnoses suggested, improving the basis for its reconstruction.

2019 – Spore walls examined at ultrastructural scale
Wilson A. Taylor investigated Oxroadia gracilis microspores and megaspores using transmission electron microscopy. Material from D revealed fine wall layering, extending the locality’s importance from visible anatomy to the microscopic construction of reproductive structures.

2021–2022 – A fuller Tetrastichia reconstruction
Rothwell, Dunn and Scott combined evidence from Oxroad A and Ballyheigue in Ireland to reconstruct Tetrastichia bupatides and establish Tetrastichiaceae. The study appeared online in 2021 and in its 2022 issue. It associated Calathospermum-like cupules with the plant, while the identity of its ovules remained uncertain; not every specimen in the reconstruction came from Oxroad.

2025 – New direct palynology revises the age interpretation
Marshall and colleagues’ paper, published online on 20 October 2025 in a nominal 2024 volume, reported spores from B, D and near C. Rare Lycospora supports a Viséan interpretation. The more precise early Viséan inference was applied to a correlated horizon at Gin Head, not presented as an exact date measured on every Oxroad plant bed.

2025–2026 – Rootlet patterning reconstructed
Wyman and colleagues used two Oxroad specimens, of Oxroadia gracilis and Oxroadia conferta, alongside separate Whiteadder material to investigate how rootlets were added and arranged. Published online on 31 October 2025 and in the June 2026 issue, the study revealed intercalation and an equilateral-triangular pattern through three-dimensional reconstruction. It was anatomical research, rather than a new independent dating of the beds.

The locality covered here. The fossil locality in this North Berwick guide is Oxroad Bay, east to south-east of Tantallon Castle. Its cliffs and foreshore preserve plant-bearing volcanic and sedimentary rocks, including reworked ash, calcareous bands and small lenses. The castle stands on the adjacent Tantallon Vent, towards the western to north-western side of the bay. The town shore, Castleton Bay and Gin Head have their own exposures and fossil records; their plants must not automatically be added to the Oxroad assemblage.

STRATHCLYDE GROUP

The Strathclyde Group is the current formal parent group of the volcanic formation exposed here. It supplies the regional stratigraphic framework, rather than a thickness measurable in the Oxroad cliff alone.

Garleton Hills Volcanic Formation

Regional volcanic succession. The formation includes basaltic and trachytic lavas, tuffs and subordinate sedimentary rocks derived from volcanic material. Regionally it lies above the Ballagan Formation and below the Gullane Formation. This does not mean that all of these rocks, both formation boundaries or the formation’s complete lava succession are exposed at Oxroad. BGS gives formation thicknesses of 380 m in the Garleton Hills, 240 m in the North Berwick area and 257 m in the Spilmersford Borehole. The igneous Geological Conservation Review account instead quotes 520 m for the Garleton Hills. These are separate regional estimates, not measurements of the fossil-bearing bay or values to combine into a local log.

North Berwick Pyroclastic Member

Plant-bearing volcaniclastics. The member contains tuffs, lapilli-tuffs and local agglomerates, together with water-deposited and reworked volcanic sediment. Regional interbeds include dolostone, siliciclastic mudstone, ripple-marked sandstone and lacustrine limestone. At Oxroad, tuffs and tuffites, meaning rocks made from volcanic material that has also undergone sedimentary reworking, contain the important plant assemblages; calcareous lenses and bands contributed to their preservation. They are not simply lava flows with leaves enclosed in them. BGS gives a regional member thickness of 0–220 m. The older names North Berwick Member and informal Oxroad Group occur in the literature, but are not additional units stacked above or below the modern member.

Local thickness and measured sections. The palaeobotanical Geological Conservation Review describes an approximately 38 m succession at Oxroad; Bateman and Scott’s 1990 account gives approximately 45 m. The difference remains unresolved. Their generalised column gives very approximate thicknesses of 3 m for each of the two lowest mapped units, 10 m for unit 3 and 25 m for unit 4. These estimates must not be used to scale every individual fossil bed. Measured sections document parts of exposures A, B, C, D and F, while faulting, lateral changes, cliff inaccessibility and covered intervals prevent a single continuous bed-by-bed reconstruction of the whole bay.

Age and revised spore evidence. These are Early Carboniferous, or Mississippian, rocks. Older studies commonly assigned the plant beds to the late Tournaisian, also described as late Courceyan, using the CM spore zone. More recent direct sampling at Oxroad yielded Lycospora pusilla from Exposure B, sample Tantallon 2, and Lycospora noctuina from sample Tantallon 3B close to Exposure C. Sample Tantallon 5B at D yielded figured Spelaeotriletes pretiosus and Radiizonates mirabilis. The rare, imperfectly preserved Lycospora records support a Viséan interpretation. Their first appearance is not a reliable exact marker of the Tournaisian–Viséan boundary, and the near-C sample is not demonstrated to be the same small lens as the anatomically preserved C flora.

Marshall and colleagues’ conclusion that the correlated vertebrate horizon was most likely early Viséan concerns Gin Head, where the tetrapod was found. It does not directly fix the age of every Oxroad plant horizon, especially A. Their cited radiometric dates, 343 ± 1.0 Ma and 339.2 ± 3.9 Ma, concern the regional Garleton Hills volcanic succession, not dated samples from the individual plant lenses. A precise numerical age for all of the fossils here would therefore imply more certainty than the evidence provides.

A changing volcanic lowland. Bateman and Scott identified five successive depositional settings: a braided floodplain; a shallow lake receiving volcanic debris; terrestrial mass-flow and base-surge deposits; a shallow lake with an important biological contribution to its sediment; and predominantly terrestrial reworked ash. Their interpretation describes an ocean-marginal lowland increasingly affected by basaltic tuff-ring volcanoes. Eruptions, sediment flows and fire repeatedly disturbed soils and vegetation. The palaeobotanical GCR’s four plant-bearing terrains, three mainly lacustrine and one associated with volcanic mass-flow, group the evidence differently; they are not four formal members or a competing count of the five sedimentary facies.

Transport and preservation. The plant assemblages are transported in the broad taphonomic sense, even where parts remained attached or a rooted plant system survived. They do not represent an entire undisturbed forest preserved exactly where every plant grew. Adpressions are flattened plant remains; permineralisations retain mineral-filled tissues, sometimes with cellular detail; fusainised material has a charcoal-like preservation, although a burnt origin requires supporting evidence. The differing mixtures of these preservation types are essential to understanding the exposures. Bateman and Rothwell’s 1990 study distinguished 43 anatomically preserved organ categories and 19 adpressed organ-species, representing at least 11 whole plants. Separate names for stems, leaves, cones, spores, cupules and ovules do not make each one a separate biological species.

Eight exposures at four probable levels. The research letters A–H identify geographical exposures, not formal beds and not an alphabetical upward succession. The probable correlation, from oldest to youngest, is B + C + F; A; D + G + ?H; E. The question mark belongs to H, whose equivalence with D and G is uncertain. Several horizons occur within individual exposures. The descriptions below retain that distinction and do not assign equal thicknesses or invented contacts to the four levels.

The mapped succession. Bateman and Scott’s informal unit 1 consists of red-grey siltstones with thin sandstones and yielded no plant remains at Oxroad. Unit 2 combines laminated grey siltstone, fine sandstone and cementstone with volcanic material, and contains B, C and F. The coarser, largely reworked tuffs of unit 3 include the channel deposit at A. More evenly bedded tuffs and impure limestones form unit 4, containing D, G, ?H and E. Their unit 5 is the coarse fill of the Tantallon volcanic conduit, cutting the other units; it is not a sedimentary bed above E. These numbered mapping units are distinct from both the modern formal stratigraphic ranks and the five reconstructed depositional settings.

Exposure B: mainly flattened plants in the lowest correlated level. B is probably a lateral equivalent of C and F. Adpressed plants predominate, with rare permineralised and fusainised remains. The original flattened megaspores later named Setispora pannosa came from B, and this spore also occurs in F. A fusainised Cladoxylon axis and petioles assigned to Cladoxylon cf. waltonii provide anatomical evidence alongside the compressions. Comparisons between these remains and the adpressed ?pteropsid stem and foliage are tentative, rather than demonstrated connections. B also yielded Sphenopterid foliage 2 and adpressed cupule aggregates, without a usable finer horizon assignment here. The modern Tantallon 2 spore sample belongs to B, but does not supply a measured thickness for its fossil beds.

B measured section. Seventeen horizons, some composite, were logged through 0.46 m of sediment. Thin, mainly plane-laminated siltstones and sandstones vary in hardness; individual horizons are 1.5–6 cm thick. Rippled surfaces and a fluted surface record water movement, while desiccation cracks were recognised in B10, B14, B16 and B17. Seven horizons contain flattened plants. These measurements describe the sampled B section, not the thickness of all plant-bearing rocks at Oxroad.

B8. The 1990 figure caption assigns an example of adpressed Ovule 2, specimen OBB(8)16.1, to this horizon. Ovule 2 is a descriptive category with a resemblance to Tantallosperma, potentially encompassing more than one species. B8 is also one of the coarse horizons in which Ovule 3 is especially common in the studied material; these isolated megaspores were interpreted mainly as probable Eosperma, possibly including small Stamnostoma.

B12 and ?B12. Secure B12 captions include a ?pteropsid stem, a flattened megaspore probably attributable to Eosperma oxroadense, Ovule 2 and Indeterminate organs 1 and 2. The last two names describe structures whose biological function remains unproved. Captions with an uncertain ?12 assignment include Sphenopterid foliage 1, ?pteropsid foliage, a ?pteridosperm stem with a petiole and Pollen organ 1. The source uses different stem and foliage labels in its prose and captions; those labels do not establish additional species. B also yielded two adpressed petiole size categories, for which possible correspondences with Lyginorachis waltonii and the unnamed Lyginorachis petiole remain unconfirmed.

Fish and coprolites at B. Bateman and Scott reported two complete, compressed palaeoniscoid fish from B12, without associated isolated scales. One is illustrated as specimen OBB(12)141. Their intact preservation was interpreted as evidence of quiet, oxygen-poor conditions on the lake floor. Coprolites occur on several B bedding surfaces, with the largest reported examples measuring 20 × 7 mm. These records establish animal remains at Oxroad, but do not show that fish fragments are common or identify the fish to species.

B13/B14 discrepancy and B16. The text places a concretion bearing Sphenopterid foliage 3 at B13, whereas its Figure 7 caption gives B14. Likewise, the best Indeterminate organ 3 specimen is described as B13 in the text but B14 in the Figure 8 caption. These are secure B records with unresolved finer positions; neither pair is silently assigned to one horizon here. Independently, B14 and B16 join B8 as coarse horizons rich in Ovule 3 in the study. These numbered occurrences do not supply bed thicknesses.

Exposure C: the foreshore lens. C includes a small plant-bearing lens in grey tuffites, with dolomitic permineralisation and often particularly good cellular preservation. It has been interpreted within a lake and braided-stream setting and lies below A in the probable local correlation. Its distinctive lycopsid was described in 1990 as Oxroadia sp. nov., subsequently named Oxroadia conferta. Vegetative axes, a rhizomorph, the reproductive structure then called Lycopsid strobilus B and associated Setispora pannosa megaspores contributed to that whole-plant reconstruction. These are associated organs of one reconstructed plant, not three separate plant species.

C measured section and C2 lens. The logged section spans about 1.5 m, with fine- to medium-grained sandstone and coarser tuffs concentrated towards its base and top. Ripple-marked surfaces occur at C14–C16; harder tuffs above protect the softer beds beneath. The plant-bearing grey tuffite at C2 varies laterally from 4 to 10 cm thick. Its anatomically preserved material was concentrated near injected sediment veins, which also disrupted plant fragments. The small permineralised zone sampled for the study was about 80 cm long and extended only 30–40 cm into the face, so it should not be envisaged as a continuous richly fossiliferous band around the bay.

C/2 and the rooted plant records. The C/2 specimen of Hestia eremosa was physically disrupted before burial and partly scorched at the outer cortex, yet preserves useful cellular detail. Other C material includes rooted vegetative systems later named Protocalamites longii, with stems, roots and rootlets, and reproductive structures named Protocalamostachys farringtonii. Their reconstruction supports a small heterosporous sphenopsid, producing both large and small spores. These other specimens are documented from C but are not all assigned here to horizon C/2. Stem bases and petioles of Cladoxylon cf. waltonii, and a single unnamed fusainised gymnosperm axis, Stem I, broaden the C assemblage.

C seed-plant organs. Eosperma oxroadense was the most abundant ovule in the studied C material. Hairy and apparently hairless forms were interpreted as potentially different preservation states. Pullaritheca longii cupules occur in directly sampled C material; Hydrasperma tenuis is the name of their ovules, also found dispersed. The older name Hydrasperma longii involved a debated separation of developmental stages and must not be counted as another securely distinct whole plant. Deltasperma fouldenense and cf. Sphaerostoma occur in C and D but were absent from intervening A in the 1990 study. A tentative in-situ C occurrence of cf. Eccroustosperma langtonense is independent of an older, probably misprovenanced loose-block report.

Loose block OBC084: probable C provenance. This particularly informative block was loose, and its C origin is inferred. It contains Ovulate cupule C with a demonstrably attached Dolichosperma cf. sexangulatum ovule, together with Hydrasperma tenuis, cf. Eurystoma burnense and material compared with Eccroustosperma langtonense. Its surface also preserves the adpressed Ovule 1, which resembles Hydrasperma but differs in attachment, curved Ovule 4 megaspores compared with Eccroustosperma, a figured possible Ovule 3 and Pollen organ 1. The block’s prefix does not turn these into precisely located in-situ C horizons. The anatomically preserved cupule letters and the adpressed cupule numbers belong to different descriptive schemes.

Exposure F: lateral adpression assemblage. F probably correlates with B and C and mainly preserves flattened plants. Records include Setispora pannosa, ?pteropsid foliage, Sphenopterid foliage 1 and 2, cupules, Ovules 2 and 3, rare Pollen organ 1 and an example of Indeterminate organ 3. The small-leaved and larger foliar forms cannot simply be assigned to named anatomical plants from their outline. Sphenopterid foliage 1 resembles Sphenopteris bifida, but the comparison is not a secure local identification of that named species.

F measured section. A 2.2 m section was divided into 40 thin horizons, typically 2–7 cm thick, mainly alternating siltstone and fine- to medium-grained sandstone. Many are graded or contain finer laminae; four tuff horizons occur within the upper 70 cm, with the uppermost strongly indurated. The authors compared the lower part of F with B and its upper part with C. F16, C2 and B12 plus B14 might be approximate lateral equivalents, but the individual fossil layers are too discontinuous and faulted for a confident bed-for-bed match.

F16 and ?F16. Figure captions assign Sphenopterid foliage 2, specimen OBF(16)6.6, and cupule aggregates OBF(16)1.3 and OBF(16)1.4 to F16. Sphenopterid foliage 1, OBF(?16)8.1, retains an uncertain horizon. The adpressed cupules are called Cupule 1 in the prose and Cupule 2 in the captions and table; their suggested relationship to Pullaritheca is not demonstrated. No second F horizon identifier or numerical thickness is reconstructed from the incomplete table layout.

Exposure A: the classic cliff assemblage. Gordon’s original green-ash block came from immediately above an angular unconformity within a large, over-steepened channel. Medium- to coarse-grained grey-green tuff contains predominantly permineralised plants, including relatively large fragments with attached petioles. The assemblage has been interpreted in relation to terrestrial mass-flow and/or volcanic base-surge deposition. Permineralisation is largely dolomitic but complex, and tissues can be distorted. The angular surface is a local relationship within the exposure; it is not established here as a formation boundary.

A measured section and channel. The channel was described as approximately 25 m wide and 10 m deep, dipping about 30° south-westwards into the cliff. Its published 4 m log groups the deposits into nine generalised beds, rather than tracing every discontinuous lens. Erosional surfaces, cross-lamination, slumping and dewatering structures record repeated deposition and disturbance. Plant-bearing horizons occur through much of the lower 3 m, but their thickness and lateral extent vary. The channel dimensions and logged thickness describe different scales and must not be added together.

A/?5. The Hestia eremosa holotype, recovered from Gordon’s original block and formerly called Lycopsid stem A or cf. Paurodendron, has this explicitly uncertain horizon assignment. Its dark tissues lack the high reflectance typical of the strongest charcoal examples, and the 2007 study suggested possible charring during the depositing base surge. Dolomitic mineralisation and compression have distorted the cortex, leaving its cellular preservation poorer than that of the C specimen. The 2007 naming of Hestia concerns previously collected material, rather than three new field finds made that year.

A stems and foliage. Tetrastichia bupatides is firmly documented in Gordon’s in-situ A material and remained confined to A in the 1990 inventory, becoming dominant in some sampled horizons. A also supplied in-situ Eristophyton beinertianum, Buteoxylon gordonianum and Triradioxylon primaevum material. The unnamed Stem H includes A axes once placed in Tristichia ovensii; that old name covered unlike stems and is not retained as an unqualified identification. The small, asymmetrical foliar axes called Lyginorachis ‘sp. A’ probably represent secondary rachises. Amyelon bovius-like roots occur in A, but the original type block’s matrix was later considered more consistent with Castleton, so its type provenance should not be treated as securely Oxroad.

Stems attributed to A from loose blocks. The type of Oxroadopteris parvus, a stem with attached petioles, came from a loose block attributed to A. The single original Calathopteris heterophylla stem, bearing two petiole types, was likewise in loose greenish-grey tuff considered almost certainly to be from A. Their inferred origin must remain distinct from Gordon’s directly sampled material. Historically proposed combinations of these stems with particular cupules and ovules are not established merely by their occurrence together in a rock block.

A cupules and ovules. Gordon’s material yielded Calathospermum fimbriatum cupules, Salpingostoma dasu and Eosperma oxroadense ovules. Tantallosperma setigera was described from in-situ A material associated with Buteoxylon, but their same-block occurrence does not by itself unite them as one plant. Pullaritheca longii is represented by directly sampled A cupules as well as historical loose material attributed to A. Mature Hydrasperma tenuis ovules have been found within A Pullaritheca cupules; dispersed examples are relatively uncommon there. A rare historical Pullaritheca cupule bearing both ovules and pollen-producing sporangia was in a loose block attributed to C. It illustrates a developmental anomaly, not the presence of flowering plants.

A block ACS652 and Ovulate cupule D. This A block contains terminal and overtopped cupules on branching axes, each bearing one to three ovules described in 1990 as Stamnostoma sp. nov. The subsequently named Stamnostoma oliveri cupule system is part of this research history. Its earlier identification as cf. Sphaerostoma must not be confused with the separate cf. Sphaerostoma organ category recorded from C and D. Stamnostoma ovules occur in A, C and D and can form thin lenses in A, but those lenses have no verified numerical thickness here.

Exposure D: impure limestone and charcoal-bearing assemblages. D lies above A in the probable correlation. Its greenish-grey impure limestone preserves plants mainly through calcitic permineralisation, with substantial fusainised material. Black, angular, reflective fragments on bedding planes provide particularly strong evidence of charcoal. Fusainised fragments assigned to Cladoxylon cf. waltonii also occur here, although that comparative identification may encompass a broad range of forms. The limestone has been interpreted as the deposit of a transient, nutrient-rich lake supporting algal blooms. The study distinguished three local measured sections and 16 correlated plant-bearing horizons within D; these are finer research subdivisions, not sixteen named formal beds of known equal thickness.

D measured horizons and rooted surfaces. The composite log combines parts of three sections across a 10 m horizontal length of exposure. The authors resolve 6–8 m of deposits into 49 horizons, with individual thicknesses ranging from 1 cm to 3 m. Of these, 16 contain permineralised plants and most are less than 8 cm thick; D2.5 reaches 22 cm but contains few anatomically preserved plant fragments. D1.5 and D1.9 contain abundant roots preserved in place and were interpreted as palaeosols, recording brief emergence and colonisation between depositional episodes. D2.22 is a siltstone about 30 cm thick; D2.31 is at least 3 m thick and is truncated by a bounding fault. Both upper siltstones contain nodular calcareous concretions; the lower, D2.22, also yields sparse, highly fragmented compressed plants. The section numbers and horizon numbers are identifiers, not measurements in metres.

D research-horizon register. The published table’s textual record retains the section/horizon identifiers 1.29, 2.30, 1.28, 2.28, 2.26, 2.23, 1.23, 2.17, 1.19, 2.15, 1.17, 2.12, 2.9, 2.7, 2.5, 1.13, 1.9, 1.5, 2.2, 3.11, 3.8 and 3.6. These 22 section-specific labels represent 16 correlated levels. Their correlation brackets and the original fossil-distribution matrix are not reconstructed here, so this register is neither a newly verified order of 22 separate beds nor a metre scale. In an identifier such as D2.30, 2 means section 2 and 30 is that section’s horizon number. The following individual records are supported independently by explicit text or figure captions.

D1.23. The largest described adpressed axis compared with Oxroadia gracilis, 12 mm wide and showing leaf scars, was recorded here. This is section 1, horizon 23, and must not be confused with the separate D2.23 identifier.

D?2.12. A figured adpressed axis identified as cf. Oxroadia gracilis has this uncertain horizon assignment. Both the comparative identification and the question mark are significant.

D?2.15. Figured Lycopsid strobilus A and Setispora subpalaeocristata material retain an uncertain section 2, horizon 15 assignment. Strobilus A is the reproductive structure associated with Oxroadia gracilis; the separately named megaspore is part of its reproductive evidence, not an additional whole plant.

D2.17 and D?2.17. Secure D2.17 records include a fusainised Oxroadia gracilis leaf and an enigmatic axis interpreted as a possible cladoxylalean petiole, both from block 176. Separately, Frond A and Stauropteris cf. berwickensis are figured with the uncertain assignment D?2.17. Frond A comprises three-ribbed foliar axes formerly included in Tristichia ovensii; it is an organ category, not a demonstrated connection to Stem H.

D2.22. The adpression account records a small example of Sphenopterid foliage 1 and Sphenopterid foliage 4 at this horizon. Foliage 4 may be an architectural variant of Foliage 3, but that relationship is unresolved. D2.22 comes from the descriptive foliage text, not the separate Table 4 register of anatomically preserved assemblages.

D?2.23. A figured ovule of Deltasperma fouldenense carries this uncertain assignment. The question mark cannot be removed simply because an unqualified 2.23 also appears in the horizon register.

D2.26. A Deltasperma fouldenense ovule from block 190 contains a microspore tentatively compared with Apiculatisporis. Pollen organ C is separately figured from block 188 at this horizon, including its contained spores. These observations document organs and their contents within a research horizon; they are not independently correlated sedimentary spore samples or proof that the ovule and pollen organ belonged to the same individual plant.

D2.28. The 1990 illustrations include an Oxroadia gracilis microspore attached to a Setispora megaspore from block OBD(2.28)085b. Megaspores from that same recorded block were used in Taylor’s 2019 study of spore-wall ultrastructure, providing unusually clear continuity between a named research horizon and later microscopic work. The separately cited microspore peel OBD084mT11 does not carry an equally explicit 2.28 horizon assignment in its identifier.

Probable scorpion cuticle: conflicting horizon records. Part 2 illustrates a fragment of probable scorpion cuticle, possibly from a tarsus, bearing several setal apertures. The Figure 9e caption (p. 170) identifies the illustrated specimen as OBD(2.28)085b, CP8, but the text on p. 178 reports several arthropod-cuticle fragments from D2.12 while referring to that figure. The horizon discrepancy remains unresolved. The interpretation is tentative: neither record establishes a complete scorpion or a named species.

D2.30. The illustrated anatomical assemblage includes the unnamed Lyginorachis petiole, an Amyelon root, Salpingostoma dasu, cf. Tantallosperma setigera, cf. Sphaerostoma and the small axis then called Lycopsid stem A, subsequently identified as Hestia eremosa. The Hestia specimen from block 192 is wholly fusainised and retains good cellular detail in the mainly calcitic matrix, although a later calcite vein cut away part of its cortex. These separate organs are recorded at the same horizon; their co-occurrence alone does not reconstruct one whole plant.

D records without a secure finer assignment. The 1990 text describes a terminal sporangium on a branching axis as Pollen organ D, but the corresponding Figure 6 caption calls it Pollen organ C. Its block, OBD(?)147, has no secure horizon. The discrepancy is retained rather than resolved by guesswork. Table 4 also lists cf. Oxroadopteris parvus among the D assemblage, but no finer horizon is assigned here. D1.19, used in the paper’s explanation of specimen notation, is only an illustrative identifier and does not establish a named fossil occurrence. No taxon-to-horizon assignments are inferred from displaced symbols in the table’s textual reproduction.

Exposure G: probable equivalent of D. G predominantly yields permineralised plants and is correlated with D. Oxroadia gracilis axes are explicitly documented. Its plant-bearing horizon is a well-indurated, impure limestone 5–7 cm thick, between much thicker tuffs. Only small test samples were collected, and no full log was made. The limited sampling and relatively poor preservation mean that its published fossil inventory is incomplete.

Exposure H: uncertain equivalent of D and G. H also preserves permineralised plants in a well-indurated, impure limestone horizon 5–7 cm thick, enclosed by tuffs. Its placement at the D–G level remains tentative, so it is shown as ?H in the correlation. As at G, only small test samples were taken and no complete lithological log was made. If H is younger than D, it would record another episode of lake deposition rather than the same horizon repeated.

Exposure E: youngest probable level. E is believed to lie above D within unit 4. The cliff exposure could not be logged or sampled in place; the study used recently fallen angular blocks directly beneath it. The 1990 work figures an Oxroadia gracilis axis and a fusainised petiole identified as cf. Triradioxylon primaevum from block E001. Part 2 describes the plant-bearing material as representing a horizon at least 30 cm thick, although this is not an in-situ measured log. Its host is predominantly dolomitic, with sparse, sometimes relatively large plant fragments. Lake or pond deposition was suspected, but the original carbonate and its subsequent alteration remain uncertain. Several metres of reworked tuffs overlie E; the final, mainly terrestrial reworked-ash setting should not be assigned wholesale to E itself.

Plant organs shared among exposures. Some important anatomical records have secure exposure-level provenance but no usable fine-horizon assignment. Amyelon roots are widespread. Triradioxylon petioles were recorded in A, C, D and E, while Buteoxylon-type petioles also occur in C. Some Triradioxylon-type petioles from A and a few from D were associated with poorly preserved stems; these associations do not establish securely identified D stems. Frond A occurs in A, C and D; large petioles tentatively assigned to Lyginorachis waltonii are widespread in permineralisation-bearing horizons. The smaller unnamed Lyginorachis petiole, particularly common in C, is recorded in organic connection with Bilignea cf. solida. The latter’s stem material combines in-situ and loose specimens attributed to C, A and D, with some assignments uncertain. These qualifications are important: an attached organ is stronger evidence of a whole plant than proximity in a slab.

Ovule and pollen-organ distributions. Tantallosperma setigera was reported throughout the six predominantly permineralisation-bearing exposures examined, rather than being limited to its original A material. Pollen organs compared with Telangium and Melissiotheca occur in A, C and D; D examples of cf. Telangium may be fusainised, and particularly informative cf. Melissiotheca material includes probable-C block OBC084. Pollen organ C is frequent in A and some D horizons; a smaller, tentatively referred C example may be juvenile. The older use of Telangium for this organ must not merge it with the separate cf. Telangium category. Adpressed Pollen organ 1, also labelled cf. Telangiopsis bifida in the table, may encompass more than one anatomical form. These are research comparisons, not a field key for identifying every dark fragment.

What the organ associations establish. Anatomical studies support distinct whole-plant reconstructions of Oxroadia conferta, especially at C, and Oxroadia gracilis, especially at D and also documented at G and E. The original O. gracilis type block was loose and its cliff source inferred, but later in-situ specimens independently establish its Oxroad occurrence. The latter plant’s reconstruction revised its original simple herbaceous interpretation. Seed-plant reconstructions are more complicated: later work connected Calathospermum-like cupules with Tetrastichia frond bases, using both Oxroad and Irish material, while the identity of the associated ovules remained tentative. Proposed links with Salpingostoma or Eosperma should therefore not be presented as settled for every A specimen.

Oxroad Bay Fault and the adjacent Tantallon Vent. The south-west-trending Oxroad Bay Fault downthrows to the south-east and separates the green tuffites from coarser pyroclastic breccias of the Tantallon Vent. Between The Gegan and Oxroad, the downfaulted tuffites include dolostone, ripple-bedded sandstone and impure limestone. Fault disturbance produces dips in differing directions, locally reaching 60°; this is not a single dip for all plant beds. The vent is a separate volcanic structure beside the sedimentary succession, not an ordinary younger bed to place above E in a stratigraphic column.

Small-scale faults within the plant exposures. The broad vent-boundary description does not capture all of the bay’s structure. Detailed mapping found 12 faults between B and C over a distance of only 50 m. D occupies a small downfaulted block between two approximately parallel faults, with changing throws and evidence of rotational movement. Bateman and Scott interpreted some rocks previously mapped as a small subsidiary vent as displaced tuff blocks. This local complexity explains why nearby exposures need not join into one simple, uninterrupted layer.

Tantallon vent fill. Castle-side cliffs expose green, largely unbedded pyroclastic material with small basanite bombs. A sandstone lens high in the cliff has been interpreted as sediment accumulated during a dormant phase; no local thickness is verified here. Other vent-fill zones contain large reddish bedded-tuff blocks and basanite bombs, while the southern part includes green bedded tuffs. The excursion account also records carbonate-veined or impregnated agglomerate with amygdaloidal basanite bombs and a curved basic dyke. Red bedded tuffs and large basanite blocks outside the vent are distinguished from its fill. The southern margin is faulted and can be obscured by shingle on the foreshore.

Oxroad and Seacliff–Gegan sedimentary structures. The excursion describes planar and ripple bedding in the green tuffs and fine agglomerate on the south side of Oxroad Bay, together with contorted bedding tentatively interpreted as algal and networks of calcite and cementstone veins. On the approach at Seacliff Harbour and The Gegan, large-scale cross-bedding, fine agglomeratic bands with volcanic and marl or limestone clasts, rectangular joints and calcite veins can also be seen. The red–green colour boundary there cuts across bedding and is interpreted as secondary oxidation, possibly connected with basic intrusions. Colour alone therefore does not define a member boundary, and these approach features cannot each be assigned to a particular Oxroad fossil horizon.

References

British Geological Survey. Garleton Hills Volcanic Formation and North Berwick Pyroclastic Member, BGS Lexicon of Named Rock Units.

Bateman, R. M. and Rothwell, G. W. (1990). A reappraisal of the Dinantian floras at Oxroad Bay, East Lothian, Scotland. 1. Floristics and the development of whole-plant concepts. Transactions of the Royal Society of Edinburgh: Earth Sciences 81, 127–159. Exposure correlations, organ descriptions and explicitly stated horizon records are used here; the full presence/absence matrix is not reproduced.

Bateman, R. M. and Scott, A. C. (1990). A reappraisal of the Dinantian floras at Oxroad Bay, East Lothian, Scotland. 2. Volcanicity, palaeoenvironments and palaeoecology. Transactions of the Royal Society of Edinburgh: Earth Sciences 81, 161–194. The measured sections, mapping, animal records and environmental interpretations discussed here are documented in pp. 165–180 and 187–190, with the exposure logs in Figures 7, 8 and 13.

Bateman, R. M., Rothwell, G. W. and Cleal, C. J. (1995). Oxroad Bay, in Cleal, C. J. and Thomas, B. A., Palaeozoic Palaeobotany of Great Britain, Geological Conservation Review Series 9. JNCC.

JNCC. North Berwick Coast, in Carboniferous and Permian Igneous Rocks of Great Britain North of the Variscan Front, Geological Conservation Review Series 27. McAdam, A. D., North Berwick Excursion B: Tantallon to St. Baldred’s, Edinburgh Geological Society excursion account.

Bateman, R. M., Kenrick, P. and Rothwell, G. W. (2007). Do eligulate herbaceous lycopsids occur in Carboniferous strata? Hestia eremosa gen. et sp. nov. from the Mississippian of Oxroad Bay, East Lothian, Scotland. Review of Palaeobotany and Palynology 144, 323–335. The specimen descriptions and source table distinguish A/?5, C/2 and D/2.30.

Rothwell, G. W. and Wight, D. C. (1989), Pullaritheca cupule revision, Review of Palaeobotany and Palynology 60, 295–309; Rothwell, G. W. and Scott, A. C. (1992), Stamnostoma oliveri cupule systems, the same journal 72, 273–284; Taylor, W. A. (2019), Oxroadia gracilis spore ultrastructure, International Journal of Plant Sciences 180, 571–577.

Dunn, M. T. and Rothwell, G. W. (2012), Tetrastichia bupatides anatomical reassessment, International Journal of Plant Sciences 173, 823–834; Rothwell, G. W., Dunn, M. T. and Scott, A. C. (2022; published online 2021), Tetrastichia whole-plant reconstruction, Review of Palaeobotany and Palynology, article 104551. The later reconstruction is used at the level supported by its accessible abstract and indexed text.

Marshall, J. E. A., Reeves, E. J., Bennett, C. E., Davies, S. J., Kearsey, T. I. and Millward, D. Placing the Scottish Romer’s Gap tetrapods within a Tournaisian time frame. Earth and Environmental Science Transactions of the Royal Society of Edinburgh 115, 99–112; nominal 2024 volume, published online 20 October 2025. Its Oxroad samples and Figure 6 underpin the qualified age discussion.