Freshwater East

Modern type section. BGS gives a 50.25 m section on the north side at SS021981. The formation begins at the base of a 1.8 m conglomerate-dominated heterolithic unit, locally scouring into Gray Sandstone; this 1.8 m is included in the 50.25 m total. Pebble and cobble conglomerate passes into buff, cross-stratified pebbly sandstone. Four grey-green heterolithic units are separated by finer red sandstone and mudstone with calcrete palaeosols, often tubular, and mottled red-green wetting-and-drying features.

The upper boundary is the top of the fourth grey-green unit, grading into red calcretised mudstone with wedge-shaped soil peds of the Moor Cliffs Formation. Older accounts used different boundaries: the 2005 GCR describes a 1.4 m basal “Basement Group” and places the four green complexes within lower Moor Cliffs. That older scheme is not a second modern thickness and does not move the plant horizon into the overlying formation. Old Red Sandstone is used here as a regional facies term; Milford Haven Group is the historical rank superseded by Subgroup.

Dixon’s ten-bed northern basal log, ascending, beds 1–5. Bed 1: 4¼ ft of conglomerate and grey quartzitic sandstone on an irregular sharp base. Bed 2: 3½ ft of green marl above hard grey sandstone, both with rod-like or nodular “race”, locally angular quartz. Bed 3: 40 ft of red marl like bed 9. Bed 4: about 5 ft of green-grey marl or mudstone, shale, sandstone and thin coarse breccia, with Cephalaspis and frequent Lingula. Bed 5: 25 ft of red marl like bed 9. These are historical lithological descriptions and identifications.

Beds 6–10. Bed 6: 5 ft of green-grey beds like 8 and 10, including dark plant-rich layers, abundant Pachytheca, occasional fish including Cephalaspis, and fairly frequent Lingula; sandstone becomes fine breccia westwards as colours change. Bed 7: 7½ ft of red marl with race and a green top. Bed 8: 3 ft of green-grey beds with a thin basal breccia. Bed 9: 43 ft of red marl with race, red conglomeratic cornstones and a green top. Bed 10: 20 ft of green-grey mudstone, shale and finely laminated sandstone, with plant layers, rare fish, lenticular feldspathic grits, pale and pink rhyolitic breccias up to 6 ft thick, and small phosphatic concretions.

The row values sum to 156¼ ft, while Dixon printed an approximate total of 160 ft. The two definitions are retained without forcing equality or substituting them for the modern 50.25 m section. Bed 10’s breccia has an upper value only: the source’s nearby “p. 32” is a page reference, not a lower thickness in inches. Edwards later placed the approximately 0.3 m plant layer within bed 8, the third grey sandstone complex.

The unconformity can appear deceptively transitional where basal sandstone resembles the marine sandstone beneath, or where strike faults repeat beds. The conglomerate includes angular quartz and quartzite pebbles up to 6 in. Bands 4, 6, 8 and 10 change laterally; malachite coats some joints. At western localities 7 and 8, thin sandstone fills small potholes and sends narrow “veins” into marl above or below. These are physical sedimentary relationships, not named trace fossils.

The early land-plant horizon

Edwards (1979) described patches of drifted plant and animal fragments in an approximately 0.3 m grey-green micaceous sandstone on the north side, within the third grey sandstone complex, Dixon’s bed 8. The source prints grid reference SS09 0236 9812. Most remains are coalified compressions, sometimes accompanied by iron oxides or sulphides; no completely petrified specimens were reported. Hostinella-type sterile axes are common, with Nematothallus, Prototaxites and Pachytheca. Some films are animal fragments, including eurypterids. This is transported material, not an in-situ forest.

Freshwater East is the type locality of Cooksonia cambrensis. The alpha holotype is NMW 77.6G.105, figured in Plate 2 figure 1. The source’s beta holotype designation is internally inconsistent: prose gives 77.6G.34, while Plate 2 figure 8 and text-figure 1E captions give 77.6G.13. Both readings are preserved rather than choosing an accession without a museum or erratum check. The plants have smooth axes, occasional dichotomy and terminal circular to elliptical sporangia.

The GB3D type-fossil catalogue records NMW 77.6G.10 as a Cooksonia cambrensis paratype from Freshwater East, held by Amgueddfa Cymru–National Museum of Wales, citing Edwards’s Plate 2 figure 4. The same block number also appears for cf. C. pertonii in Plate 4 figure 12. More than one fossil can occur on a block, so these records are kept distinct rather than silently changing an identification or inventing another accession.

The historically identified local C. hemisphaerica material includes axes 0.25–1.6 mm across and sporangia 0.3–2 mm wide and 0.2–2 mm high. Its holotype V58012 (Lang 181) is from Targrove, not this bay. Local cf. C. pertonii includes NMW 77.6G.10; the definitive specimen 114 shown for comparison is from Perton Lane. Specimens 33 and 113 were only cf. C. caledonica, and 113 was destroyed during preparation. Specimens 3, 74, 27a/b and 35 include material kept in open identification; apparent overtopping or unusual outlines do not establish a new local species.

The study sorted 83 more complete fertile specimens into six practical morphology groups, not 83 species; 77 axes could be measured. Six sporangia supplied the separate spore measurements. A printed conflict remains between the amplified C. hemisphaerica spore diagnosis, 22.5–34.5 micrometres, and Table 2 specimen 111, 25.5–37.5 micrometres. Other figure and accession disagreements also occur in the original paper. Later restricted concepts of Cooksonia, discussed by Gess and Prestianni (2021), complicate the historical use of the rounded-sporangium names C. hemisphaerica and C. cambrensis; no automatic renaming of these local specimens is made here.

Tortilicaulis transwalliensis was named from Freshwater East and is the type species of its genus. Holotype NMW 77.6G.2, Plate 5 figure 1, has a 3.9 mm high, 0.95 mm wide terminal body and a 0.4 mm wide axis. The material has elongate terminal bodies interpreted as sporangia and twisted axes, generally 0.1–0.4 mm wide. Seventeen more complete specimens were grouped by shape, not into five species. The observed material was unbranched, but fragmentary; no spores were isolated and no tracheids demonstrated in the original study. Repeated block accessions in the table do not necessarily represent duplicate individual fossils.

The affinity of Tortilicaulis was originally uncertain. Morris and colleagues (2024) demonstrated vascular tissue in the related T. offaeus from Brown Clee Hill in the Welsh Borderland, strengthening a vascular-plant interpretation of the genus. That is comparative evidence, rather than a new anatomical observation of the Freshwater East type material. Later branching T. transwalliensis and trilete spores were documented at Targrove. Neither those younger specimens nor living Pellia illustrated for comparison are Freshwater East fossils.

Other local remains include provisional cf. Salopella sp., NMW 77.6G.109a, figured with spore fragments in Plate 2 figures 13–14; an unnamed elongate body 108; and uncertain material with conflicting printed accession references. Hostinella is a form genus for sterile branching axes, not proof of Zosterophyllum. Spinose material called Psilophytites sp. may represent at least two taxa, but does not establish Psilophyton or Sawdonia. The paper’s regional South Wales plant table does not add Steganotheca, Powysia, Inopinatella or regional tracheid records to this beach.

Edwards interpreted a drifted coastal assemblage, possibly accumulated among channels and sandbanks with sea influence. Saltmarsh vegetation was considered, as were riverbank or lake-margin plants carried into coastal mudflats. The modern sedimentary framework likewise allows periodic marine or brackish influence among alluvial and soil-forming environments. Neither account proves one exclusive habitat for every plant fragment.

Microscopic spores and age evidence

Higgs’s preliminary 2015 conference account reported seven productive grey-mudstone samples from the type section’s green-grey heterolithic units. Cryptospores made up more than 67% of those assemblages, whereas ornamented trilete spores comprised less than 10%. Comparison with the Spanish spore succession suggested an early–mid Přídolí age, and rare prasinophytes and acritarchs indicated restricted marine or brackish influence. These are the preliminary account’s results: the seven samples are not seven independently established fossil horizons or a substitute for the later paper’s sample inventory.

Higgs’s 2022 published abstract describes the Freshwater East type section and reports 11 cryptospore and 31 trilete-spore taxa, with cryptospores making up 67% and smooth monads and dyads predominating. It names the new species Cymbohilates richardsonii and Velamisporites edwardsae. The Chelinospora lavidensis assemblage is correlated with the upper Chelinospora hemiesferica Biozone and interpreted as early mid Přídolí. Spanish chitinozoans supply external age control, rather than another fossil occurrence at this beach.

The abstract also reports rare small spiny and sphaeromorph acritarchs and prasinophyte phycomata in two local assemblages. These microscopic remains support tidal influence in fluvial and estuarine environments. This is an abstract-level summary, not a complete species or specimen inventory: individual sample heights, vouchers and the remaining named taxa are not established by the evidence presented here.

Animal traces and specimen provenance

Morrissey and Braddy (2004) explicitly include Freshwater East at SS025980 and report Diplichnites gouldi in its grey Freshwater East Formation sandstones. Their Figure 5d illustrates an 8 mm externally wide, bilobed Diplopodichnus trail leading to an elliptical pit, suggested as a possible Beaconites entrance. The figure establishes no Freshwater East accession, no Diplopodichnus biformis determination and no confirmed Beaconites species. Manorbier population dimensions and Bristol vouchers must not be reassigned to it, and the trail is not allocated to plant bed 8 without evidence.

The 2015 BGS account separately states “At Freshwater East” before listing Diplichnites, Palmichnium, Rusophycus, Isopodichnus, Striatichnium, Arenicolites and Helminthopsis, citing the 2012 study by Morrissey, Hillier and Marriott. These are retained as BGS-reported local genera at formation level. The evidence used here does not resolve their individual bed heights, species, specimens or figure assignments.

Marriott and colleagues’ 2009 illustrations labelled Freshwater East Formation do not supply exact specimen sites or vouchers for this beach. They are not independent proof of local species records, even where their illustrated genera overlap the later BGS list. Generalized tuff profiles, proposed tsunami transport or colonisation models, and the Actinophyllum record from Little Castle Head are not added as events or fossils at Freshwater East.

Historical red-bed records. Dixon’s locality 6 supplied Pachytheca, Lingula squamiformis and Cephalaspis, with registers E.D.971, 972, 1039, 1040 and 1853–1867. Locality 7 supplied eurypterid fragments, E.D.1868–1869; locality 8 recorded uncertain “tracks of crustaceans?”, E.D.973 and Pl.1236–1237. Locality 10, 450 yd WNW of Trewent Point, recorded Lingula, E.D.874, 875 and 1439. The other numbered localities in that table belong to Manorbier, Caldey, Swanlake, South Studdock or Greenala; their fish, plants and accessions are not Freshwater East records. These old registers do not identify the modern illustrated trail specimen.

Above the fourth grey-green unit, red calcretised mudstone with wedge-shaped soil peds belongs to the Moor Cliffs Formation. Only part of the unit is exposed here; no complete local thickness is inferred. The Townsend Tuff lies within the wider Moor Cliffs succession and should not be used as a blanket base for all Milford Haven rocks or treated as a precisely demonstrated Silurian–Devonian boundary at this bay. Regional type-section thicknesses and tuff intervals from Freshwater West, Gateholm or Priest’s Nose are not local measurements.

Structure and superficial deposits

Dixon’s map of southern Freshwater East distinguishes faults, fossil localities, historical marine and red-marl subdivisions, blown sand and several separate dip or cleavage symbols. Its 80 yd leftward heave is a mapped horizontal relationship, not a vertical throw or uniform bedding angle. The mottled beds at Trewent Point are fault-isolated, and their distance from historical group boundaries could not be determined. A basal conglomerate stack lies between branches of a major north-east fault; largely unexposed green beds prevented direct comparison with the full northern ten-bed log.

Dixon mentioned a submerged forest at Freshwater East on Leach’s authority, but it was not exposed during the Survey work. No measured local forest log or species list was supplied. The separately described Lydstep peat, pig remains, flints and shells cannot be transferred here. The photographed loose fossiliferous limestones likewise require their own provenance rather than being treated as an additional in-situ formation.