Fremington Fossil Hunting

1864 – Coastal gravels enter the ice-age debate
George Maw described the gravel at Fremington Quay and followed comparable material through the railway cutting west of the Pill. His proposed connection with clay deposits farther inland helped establish a question that would occupy later researchers: whether the coastal gravel was an old beach, and how it related to glaciation.

1913 – Striated stones influence the interpretation
Henry Dewey described marine gravel overlain by head and stony clay at the quay. Scratched stones encouraged a link with glacial deposits inland. Later studies challenged that correlation, making the original observations part of a continuing scientific debate rather than a settled explanation.

1929 – Carboniferous ammonoids recorded near the Pill mouth
E. W. J. Moore recorded Reticuloceras from beds near the mouth of Fremington Pill. These finds identified younger Carboniferous rocks in the quay area, distinct from the Devonian–Carboniferous boundary beds along the shore to the north.

1955 – A reference locality for the Pilton B fauna
Roland Goldring established fossil divisions within the Pilton succession, with Fremington becoming the type locality for the B assemblage. Its brachiopods and trilobites gave researchers a practical framework for comparing the early Carboniferous rocks of north Devon.

1960 – Historic ammonoid sites relocated
Prentice relocated Moore’s two sites near the Pill mouth and documented ammonoids and a marine bivalve from the foreshore and small cliff. The work refined the biostratigraphic evidence for these younger Carboniferous beds.

1966 – A detailed account of the western Quaternary deposits
Stephens described the sequence of broken shale, pebbly silt, gravel and sand, and weathered stony clay at the quay. He favoured a glacial interpretation for the upper material, although its lack of shells and carbonate distinguished it from the inland Fremington clay deposits.

1970 – The Devonian–Carboniferous transition clarified
Goldring’s account related the fossil changes along the northern shore to folds and faults. The succession passes from latest Devonian A3 faunas into Carboniferous B faunas, with C and repeated B farther south. This showed why the section could not be read as a simple uninterrupted succession.

1972 – Reworking offers an alternative ice-age explanation
Edmonds emphasised the variability of the quay’s pebbly deposits, including clay lenses and stratified sand and gravel. He proposed that material derived partly from till had been moved and sorted again, and cautioned against assuming that the coastal gravel continued into the inland claypit sequence.

1974 – The locality gains statutory recognition
Fremington Quay Cliffs was first notified as a Site of Special Scientific Interest. Its importance embraces both the fossil-bearing bedrock and the western deposits that preserve evidence of much younger environmental change.

1974 – Coastal and inland gravels compared
Kidson and Wood compared the coastal sediments with the inland claypit deposits and disputed a direct correlation. Their interpretation favoured movement under cold, periglacial conditions for the quay’s stony clay, illustrating how the same exposures could test competing models.

1985 – Survey evidence brings the local geology together
A new geological memoir documented fossils from precisely located reefs and low cliffs, as well as folded bedrock, the narrow dyke and younger pebbly deposits. Its locality-by-locality records remain valuable for distinguishing the northern fossil shore from the Pill-mouth and inland exposures.

1993 – Plant spores narrow the system boundary
Microscopic plant spores from six mudstone samples provided an independent test of the larger-fossil evidence. Ó Liatháin reported that the Devonian–Carboniferous boundary lay within a 5.6-metre interval between sampled horizons near the northern part of the section.

1996–1998 – Displaced bedrock prompts renewed discussion of ice
A 1996 field meeting examined disturbed bedrock at the eastern end of the western Quaternary section. A four-bed log published in 1998 recorded displaced rock, deformed laminated silt, a thin silty layer and sandy gravel. Small-scale deformation was interpreted as possible evidence for a thin ice margin, with meltwater deposits above.

2005 – Crinoid stem fragments receive specialist study
Research on disarticulated crinoid stems included material from Fremington Quay. It showed the scientific value of the small skeletal fragments that are more likely to be encountered than complete crinoids, and helped compare Devonian and Carboniferous faunas within north Devon.

2024 – The coastal gravels remain open to interpretation
A new review of Devon’s Quaternary geology noted the shape and imbrication of gravel clasts at Fremington Quay and their position above laminated silt. These observations supported consideration of river transport and reinforced the need for caution when assigning one age or origin to the whole coastal sequence.

The detailed account below distinguishes the north-running Pilton fossil shore, from approximately SS 517 340 to SS 517 335, from the Quaternary cliffs west of Fremington Pill, approximately SS 514 332 to SS 509 331. The neighbouring bedrock around the quay and Pill mouth is described separately. The beds are folded, faulted and unevenly exposed: a walking direction is not a reliable substitute for stratigraphic order.

EXMOOR GROUP

Pilton Mudstone Formation

The formation spans the Famennian and Tournaisian stages, crossing the Devonian–Carboniferous boundary. Fremington exposes its mudstone-rich upper part and is a reference section for that interval. Thin, commonly lenticular sandstones, limestone lenses, calcareous or siliceous concretions and locally phosphatic nodules occur among grey shales and silty mudstones. No formal member is securely identified for this shoreline in the sources used here. The historical Pilton A, B and C divisions are fossil assemblage zones, not members or individually named rock beds.

Northern shales: Pilton A3, latest Devonian. Opposite the old limekiln and for roughly 150 metres southwards, folded and cleaved blue-grey shales contain sparse sandstone lenses and phosphatic nodules. Tidal mud frequently conceals their surfaces. The sparse larger fossils include Whidbornella caperata and Phacops accipitrinus. The assemblage places these rocks in the latest Devonian part of the succession. The surveyed reef at SS 5175 3384 also yielded brachiopods, bivalves, bryozoans and crinoid stem pieces, although the survey assigned that collection only tentatively to Pilton A.

Survey collection N: reef at SS 5175 3384. The survey assigned this collection only to “probably Pilton A”. It records the bryozoan Fenestella; brachiopods Aulacella interlineata, cf. Dalmanella whidbornei, Derbyia?, Chonetes sauntonensis, cf. Composita strunzensis (also marked questionable in the occurrence cell), Crurithyris unguiculus and cf. Cyrtospirifer grabaui; bivalves Aviculopecten transversus, Pteronites and ?Streblopteria lepis; an indeterminate trilobite; and crinoid columnals. These are published collection records, not a promise of present-day abundance. The survey’s names and qualifications are retained.

The first prominent shale reef and overlying grey-green shales: Pilton B. Southwards, a more diverse fauna occurs in blue-grey shale beneath the low cliff and the grey-green shales above. Published records include Avonia schmidti, Cleiothyridina roysii, Ovatia spinulifera, Productinella fremingtonensis and Strophonema paeckelmanni, with the trilobites Phillibole duodecimae and P. hercules. These are early Carboniferous assemblages. The important change is palaeontological: there is no comparably sharp change in the general rock type at the system boundary.

The spore-bearing boundary interval. A 1993 study sampled six mudstone horizons using distances south of the old limekiln. These are laboratory-extracted terrestrial microfossils preserved in marine sediment, not hand-specimen collecting targets. The author’s section was generalized: the historical distances and relative sample positions are not a modern measured bed log or a set of collecting waypoints.

Frem 1: brown-grey mudstone, LN Biozone. The lowest studied Fremington sample was taken approximately 122 metres south of the limekiln. Its latest Devonian assemblage includes Retispora lepidophyta, Verrucosisporites nitidus, Crassispora catenata, Vallatisporites hystricosus and Hymenozonotriletes explanatus. The study’s first appearance of Cyrtospora cristifer was recorded here.

Frem 4: grey mudstone, LN Biozone. At approximately 148 metres south of the limekiln, the grey mudstone yielded a relatively poor, low-diversity assemblage dominated by Retispora lepidophyta and Verrucosisporites nitidus. It remains within the latest Devonian LN Zone.

Frem 5: the highest sampled LN assemblage. The sample was described as approximately 152.2 metres south of the limekiln and 0.5 metre below a thin limestone. Its abundant and varied spores include Retispora lepidophyta, Verrucosisporites nitidus, Diducites versabilis, Tumulispora malevkensis and Corbulispora cancellata.

Unnamed limestone and the boundary bracket. The thin limestone between Frem 5 and Frem 7 is an unnamed bed. Its conodonts did not resolve the system boundary. The spore study reported a 5.6-metre bracket between the highest sampled Devonian LN assemblage and the lowest sampled Carboniferous VI assemblage; this is the author’s reported interval, not a new calculation from the rounded shoreline positions.

Frem 7: grey mudstone, VI Biozone. Approximately 156.9 metres south of the limekiln, a grey mudstone described as 4.7 metres above the limestone yielded a restricted basal Carboniferous assemblage. Retispora lepidophyta, Vallatisporites hystricosus and the Diducites complex were absent. Smooth-spored Retusotriletes and Punctatisporites dominate.

Frem 8: dark-grey mudstone, VI Biozone. The sample at approximately 166.8 metres south of the limekiln gave a poor, limited assemblage dominated by Retusotriletes and Verrucosisporites nitidus. Its full assemblage, rather than one surviving species alone, supports the Carboniferous VI-zone assignment.

Frem 12: thick mudstone, VI Biozone. The highest sampled VI assemblage came from a thick mudstone at approximately 202.3 metres south of the limekiln. Preservation and diversity were poor, with Punctatisporites and Retusotriletes dominant. The study placed it 35.5 metres above Frem 8; that relative sample spacing is not the thickness of this mudstone bed.

Siliceous shales near the dyke: historical Pilton C. Near SS 517 336, faulting disrupts folded siliceous shales containing calcareous and siliceous concretions. The Geological Conservation Review lists Brachymetopus maccoyi, Piltonia salteri, Productinella fremingtonensis and Unispirifer tornacensis. The survey memoir instead retains the historical identification Brachymetopus woodwardii for the C fauna near the dyke. These differing published names are reported without assuming synonymy or claiming two newly verified species. Delicate bryozoan fronds and crinoid debris indicate little transport before burial. A later Tournaisian age has been proposed for this fauna, but remains qualified. No conformable transition from B to C has been demonstrated here, and whether C lies within or above B has not been securely resolved.

Repeated Pilton B towards the quay. The southern shore contains increasingly imbricated shales, lenticular sandstone and black, organic-rich shale with crinoidal bands. Brachiopods include many of those found farther north in B, together with Brachythyris ratingensis and Derbyia steinhagei. The survey also records Gattendorfia at genus level, Archegonus (A.?) porteri and the historical identification Brachymetopus woodwardii from the southern beach. The return to a B assemblage reflects the structurally complicated outcrop, not a second Devonian–Carboniferous boundary. Fremington is the type locality for the historical B assemblage zone.

Survey collection V: low cliff at SS 5175 3349. The Pilton B collection includes an indeterminate solitary coral, Fenestella plebeia, Leptagonia analoga, an indeterminate productacean brachiopod, Reticularia? and an elythid brachiopod. Gastropods were recorded as “Euomphalus”, “Murchisonia”, cf. Naticopsis (N.) hallii and Platyceras?. The bivalves are Cypricardinia, Palaeoneilo, Paracyclas, Parallelodon, Polidevcia cf. sharmani and Pterinopectinella austeni; trilobites are Archegonus (Phillibole) and Archegonus (Angustibole?) porteri, with crinoid columnals also present. The digitized survey table also keys Imitoceras to this collection, although its displayed occurrence falls on the cephalopod heading row, so that record needs caution. Quotation marks and question marks preserve the survey’s historical identifications.

Survey collection W: low cliff at SS 5175 3350. The survey records a Lower Carboniferous fauna from beds assigned by Goldring to Pilton B. It includes an indeterminate solitary coral; brachiopods Schizophoria?, Schuchertella?, Rugosochonetes cf. hardrensis, ?Ovatia spinulifera, “Productellina” fremingtonensis, an indeterminate rhynchonellacean and Spiriferellina; an indeterminate gastropod; and crinoid columnals. This separately located collection should not be merged with V merely because the grid references are close.

Survey collection S: historical railway cutting at SS 5170 3339. This nearby cutting provides a separate provenance record, not a recommendation to enter a cutting or leave public routes. The survey’s age assignment is “Lower Carboniferous?”. The collection contains Palaeacis?, Syringaxon?, a questionable indeterminate chonetacean brachiopod, an indeterminate productacean brachiopod, Crurithyris and crinoid columnals. Its uncertain age and identifications are retained.

Lamprophyre dyke: width and direction. The intrusion at approximately SS 5175 3365 is about 0.8 metre wide and trends 290° in the published survey. It cuts shale and silty shale with thin limy lenses on the western side of the Penhill promontory. Its rusty-brown, pitted surface contrasts with the sedimentary rocks; the width is not a sedimentary bed thickness.

Dyke mineralogy and weathering. Petrographic examination identified an autolytically altered kersantite lamprophyre. Former olivine phenocrysts are replaced mainly by calcite, with some quartz and chlorite; the fine matrix contains oligoclase feldspar, biotite, secondary chlorite and carbonate, with accessory ilmenite and traces of pyrite. Irregular rounded calcite–quartz aggregates, typically up to 5 mm across, were interpreted as altered former vesicles. Dissolution of their filling minerals produces pits, while iron alteration contributes limonite staining. The proposed affinity with Permian lamprophyres near Dartmoor is a correlation, not a direct age determination for this dyke.

TEIGN VALLEY GROUP

Codden Hill Chert Formation

Cherty shales by the former station, SS 5156 3329. The 1985 survey described about 70 metres of mainly shale and cherty shale, dipping approximately 45° towards 200° on the northern limb of a syncline. The foreshore yielded Posidonia becheri. The thickness is a published local succession estimate, not the height of today’s exposure. The modern Codden Hill Chert Formation is assigned to the Teign Valley Group, with a regional Arundian–Brigantian age range; the following remapping caveat remains important for the old shale-rich local assignment.

Southern syncline limb and Pill-bank counterpart. Shale, mudstone and chert occur on the estuary shore about 300 metres west-south-west of the Pill mouth. Comparable rocks immediately south of Pill Cottage at SS 5134 3306 show fractures, overfolds and contortions. This southern outcrop is narrower than the corresponding station-side belt. It is a separately located structural counterpart, not an extra thickness to add to the northern section.

The historical “G. spiralis” Bed and associated shell-bearing shales. Work summarized by the survey recorded Posidonia becheri, P. membranacea and repetition of a horizon called the “G. spiralis” Bed on the western bank of Fremington Pill. This is an old biostratigraphic label retained to explain the published record. Its exact modern exposure and correlation are not established by the sources used here. It should not be drawn as a continuous, readily collectible bed along the north-running Pilton shore.

Revised regional nomenclature. Older descriptions divided these Carboniferous rocks differently. The Doddiscombe Formation, beneath the chert, and the Dowhills Mudstone Formation, above it, are now recognized regionally within the Teign Valley Group. The BGS Lexicon warns that their distribution has not been fully mapped in this western district. It also reports Jackson’s 1991 reinterpretation of Prentice’s limestone- and mudstone-dominated “Fremington facies” as the upper half of the Codden Hill Chert Formation, with the lower half thrust out. That reinterpretation concerns the Fremington area; it does not independently reidentify every interval in the old station-side description. Some mudstones previously included in the Pilton, Codden Hill Chert or Crackington units therefore need reassessment. There is insufficient exact-shore evidence here to invent separate measured Doddiscombe or Dowhills sections, or to assign the old named fossil bed confidently to one of them.

HOLSWORTHY GROUP

Crackington Formation — historical shoreline mapping

Pill-mouth foreshore, SS 5135 3325. The memoir records Reticuloceras reticulatum from this foreshore locality, within rocks historically mapped as Crackington Formation. This younger Carboniferous record is separate from the Pilton boundary fauna north of the quay.

Small-cliff foot, SS 5135 3324. A distinct collection yielded Reticuloceras gracile, an indeterminate Reticuloceras and the bivalve Dunbarella rhythmica. The close grid references do not justify merging its recorded fauna into the preceding collection.

Later collection, SS 5131 3321. The published fossils are Reticuloceras subreticulatum, R. todmordenense and R. aff. eoreticulatum, indicating an early Kinderscoutian interval. The affinity qualification is retained: this is not an unqualified species identification.

Turbidite sandstone beds near the estuary shore, SS 5128 3319. Beds reach up to 0.5 metre thick in the sandy development within the syncline beneath the Pill mouth. They are interpreted as deposits of sediment-laden submarine flows.

Thicker sandstone beds at the old tramway end, SS 5151 3313. A separate survey exposure has sandstone beds up to 1.5 metres thick. It lies at the end of the former tramway from Penhill limestone quarry; this is geological provenance, not a current access recommendation.

Thin sandstone in the surrounding mudstone-rich succession, SS 5130 3321. Similar sandstone beds reach up to 0.2 metre thick in the more argillaceous rocks around the sandy development. These are separately recorded beds, not another composite unit thickness. In modern terminology the Crackington Formation belongs to the Holsworthy Group, but the exact local distinction from Dowhills Mudstone has not been resolved in this evidence. The older assignment remains explicitly historical.

QUATERNARY DEPOSITS — WEST OF FREMINGTON PILL

The principal western section extends for roughly half a kilometre. Near its eastern end, about four metres of Quaternary sediment have been described above a four- to five-metre shale cliff; farther west the exposure falls to a low cliff about one to two metres high. Thicknesses and sediment types vary laterally. The deposits include gravel, sand, silt and stony clay, but their interpretation as raised beach, till, head or water-reworked sediment has long been disputed. They do not constitute a demonstrated continuation of the inland pottery-clay succession.

Eastern four-bed log: scope. The following bottom-to-top log was published in 1998 for SS 513 332. The quoted figures are maximum thicknesses of individual beds, not a single measured composite thickness applicable everywhere along the coast. Bed numbers are local to this published log, not formal regional units.

Bed 1 — displaced bedrock, up to 1.5 metres. Fragments range from small gravel to large cobbles. The published description records movement of bedrock blocks towards the south-east by as much as 0.5 metre. This disruption underpins the proposed glaciotectonic interpretation, although it does not by itself provide a numerical age.

Bed 2 — deformed laminated sandy silt, up to 1.4 metres. The laminated sediment contains inclusions of grey-blue silt and iron-stained banding shown in the published section figure. It is contorted and cut by reverse faults whose planes dip north-west. The deformation was interpreted as possible activity at a thin ice margin and the deposit as basal or subglacial. The investigator did not record erratics or evidence of overconsolidation in this bed, limitations that should remain part of the interpretation.

Bed 3 — poorly sorted silt, up to 0.2 metre. A thin interval of silt containing some sand, with lenses of finer material shown in the section figure, separates the more strongly deformed sediment below from the upper sandy gravel. Its description supplies a distinct horizon in the local log, but no separate age or diagnostic fossil assemblage was established.

Bed 4 — sandy gravel, up to 2 metres. The upper deposit is partly cross-laminated and more widely developed, with rounded pebbles illustrated in the section figure. It was interpreted as material deposited by meltwater in an outwash fan. That is a published environmental interpretation rather than a settled identification of every gravel exposure on the shore.

Unnumbered soil and slopewash cap. The published section figure shows a surface layer of soil and slopewash above Bed 4. It is unnumbered and has no separately stated thickness in this account. It should not be renamed Bed 5 or included in an invented total from the four quoted maxima.

Stephens’s 1966 log: a separate description. This older generalized account records the following four units from base upwards. Its numbering is independent of the later four-bed eastern log. Similar numbers do not establish identical boundaries or one-to-one correlation.

Stephens unit 1 — distorted and brecciated shale. The lowest unit is broken and deformed shale. It records disturbance of the older substrate, rather than a newly named Devonian sedimentary bed.

Stephens unit 2 — silt with pebbles. Pebble-bearing silt succeeds the disturbed shale in this historical description. No separate thickness is supplied here.

Stephens unit 3 — gravel and sand. Gravel and sand form the next recorded unit. Their proposed marine or water-reworked origin belongs to the wider interpretative debate, rather than being established by grain size alone.

Stephens unit 4 — weathered pebbly sandy clay. The upper clay contains striated stones and erratic material in the published account. These observations helped motivate a glacial interpretation, subsequently challenged by workers favouring reworking and solifluction. Unlike the inland Fremington clay, Stephens described the coastal stony clay as non-calcareous and without shells.

Clay lens west of the Pill, SS 5111 3313. Pebbly clayey sand and silt enclose a clay lens approximately five metres long and one metre thick. This is the geometry of a local lens, not the thickness of the whole coastal deposit.

Stratified sediment west of the Pill, SS 5118 3315. The survey records slight stratification here and describes seams and lenses of clay, silt and gravelly sand in the western deposits. It also notes a poorly defined upper layer of weathered sandy pebbly clay, compared by Stephens with inland boulder clay. The similarity was part of a historical interpretation; it does not demonstrate continuity with the inland claypits.

Western vertical pipes and cracks. Towards the western end of the Quaternary exposure, near-vertical structures are filled with lighter-coloured silt and clay. They extend below the visible base of the low cliff. Frost cracking and desiccation have both been considered, so a single origin should not be assumed.

The wider drift surface. The survey reports pebbly drift rising to about 18 metres above Ordnance Datum around Fremington and Penhill. This is a surface elevation, not the thickness of a single coastal bed.

Station-side superficial cap east of the Pill. The survey recorded up to two metres of pebbly drift over Carboniferous rocks near the former Fremington Station. This is a separate eastern exposure, not the western four-bed log repeated unchanged.

Northern promontory cap, SS 5173 3403. On the western side of the promontory, the survey recorded approximately three metres of pebbly sand, with seams of fine sand showing some stratification. It is a separately located capping deposit above part of the northern fossil shore.

Origins and age remain qualified. Historical models included a Hoxnian beach overlain by Wolstonian till, an Ipswichian beach followed by Devensian solifluction deposits, and glacially displaced bedrock followed by meltwater sediment. A 2024 review highlighted imbricated, subrounded to subangular clasts at Fremington Quay as possible evidence for river transport, and noted gravel resting unconformably on laminated silt. The coastal sediments remain important precisely because their origin, correlation and age need careful interpretation. Shells, microfossils and numerical dates from the inland Fremington claypits cannot be assigned to this shore without an independently demonstrated connection.

References

British Geological Survey. Pilton Mudstone Formation; Codden Hill Chert Formation; Doddiscombe Formation; Dowhills Mudstone Formation; Crackington Formation. BGS Lexicon of Named Rock Units.

Edmonds, E. A., Whittaker, A. and Williams, B. J. (1985). Geology of the country around Ilfracombe and Barnstaple. Memoir of the British Geological Survey, sheets 277 and 293. Chapters 2–6 and Table 6.

Ó Liatháin, M. (1993). Stratigraphic palynology of the Upper Devonian–Lower Carboniferous succession in North Devon, southwest England. Annales de la Société géologique de Belgique, 115(2), 649–659; volume dated 1992, published 1993.

Campbell, S. and Croot, D. G. (1998). Fremington Quay, in Quaternary of South-West England. Geological Conservation Review Series, 14.

Cossey, P. J. and colleagues (2004). Fremington Quay, in British Lower Carboniferous Stratigraphy. Geological Conservation Review Series, 29, pp. 485–487.

Bennett, J. A., Cullingford, R. A., Gibbard, P. L., Hughes, P. D. and Murton, J. B. (2024). The Quaternary geology of Devon. Proceedings of the Ussher Society, 15, 84–130, particularly p. 87.