A single genus represents this family, which is distinguished by its distinctive marginal processes. Species are known from the Upper Cretaceous (Campanian) to the present day. Modern species are attached, epiphytic and indicative of shallow marine conditions. (Round et al., 1990, p. 188).

Aulacodiscus Ehrenberg, 1844

Cells are solitary, mostly circular in valve view, rarely quadrate, with conspicuous marginal nodes or processes (rimoportulae) usually extending on the valve exterior as hyaline tubes and on the interior as horseshoe-shaped slits. Valve with a circular hyaline area at the valve centre from which rows of mostly loculate areolae radiate in sectors with a hyaline ray leading to each marginal process. Valves often dissimilar, with one valve lacking the conspicuous marginal processes. The valve surface is often corrugated and the valve mantle narrow and mostly separated from the valve face by a marginal ridge. The areolae are closed by cribra externally and open internally by large foramina. Some species have a single layer pierced by areolae with cribra.

A benthic genus with many fossil species, ranging from Upper Cretaceous (Campanian) to Holocene. Difficulties in distinguishing species were commented on in the extensive review of Burke and Woodward (1963–1974), in particular the fact that some species, including those encountered in North Sea Palaeogene and Neogene sediments, possess dissimilar valves (heterovalvy), which has caused taxonomic problems where whole frustules are not preserved intact. A recent morphological study of some of these taxa has resolved some of these issues (Strel'nikova et al., 2004). A number of species of Aulacodiscus occur in the Palaeogene of northwest Europe (including the North Sea, Norwegian Sea and West Shetland Basin) in pyritised form, some of which form useful stratigraphic markers. The taxonomic identity of these taxa was revealed through SEM studies and comparisons with specimens held in the collections of the Natural History Museum by Mitlehner (1994), refining the earlier work of King (1983) which highlighted their importance as stratigraphic markers in the Oligocene and Miocene but which had them left in open nomenclature (e.g. “Diatom spp. 3, 4 & 5”).

Late Cretaceous (Campanian: Tapia and Harwood, 2002) to Holocene. Although long-ranging, important stratigraphic marker species occur in the Eocene, Oligocene and Miocene of NW Europe and elsewhere, several are described more accurately in pyritised form for the first time.

Aulacodiscus allorgei Chenevière (1934: 106, plate 8, fig. 2). Plate 1, figs. 3–4.

Valves are circular, diameter 100–150 µm, with four (rarely five) broad, ligulate rimoportulae projecting radially outwards from dilated clear central areas at the angles of the polygonal elevation, with the outer part nearly horizontal and extending generally as far as the margin of the valve. Surface with a flat marginal zone thence sloping sharply to a more or less elevated large polygonal (usually quadrate) central zone with slightly convex sides, rounded angles somewhat distant from the margin, and sometimes a slight central depression. Markings small polygonal cellules, five to six in 10 µm.

Ross and Sims (1970) conducted a scanning electron microscope study of A. allorgei and described the fine structure of the valve. Burke and Woodward (1963–1974, p. 323) commented on the variation between specimens of the outline of the central elevation, which may be straight, convex or concave. They also remarked on the general difficulty in distinguishing A. allorgei from A. hirtus Barker and Meakin, with which it often co-occurs in the “Kamishev” deposit (now Kamyshlov, Sverdlovskaya District, late Palaeocene to middle Eocene). Strel'nikova et al. (2004) conducted a detailed study of diatoms from the Russian Collection in the California Academy of Sciences and also figured and described A. allorgei from the Sengilei diatomite (considered to be earliest Eocene; Oreshkina and Oberhänsli, 2003).

Aulacodiscus heterostictus Barker et Meakin (1944: 251, plate 38, fig. 2). Plate 1, fig. 2.

Valves are circular, diameter 150 µm, with three rimoportulae, each forming tubes with a slender stem and an enlarged, widened terminal portion. Surface depressed at centre, rising gently towards rimoportulae, descending to a flat border with centre. Markings different in three concentric zones: (1) a small rounded central zone with few conspicuous rounded areolae scattered irregularly in translucent furrows; (2) areolae being closer together in outer half of row, remotely spread in inner half; (3) border with radiating rows of confluent quadrate cellules extending more or less over edge of raised surface (Barker and Meakin, 1944, 251).

Barker and Meakin (1944) made comparisons with A. schmidtii Witt and A. septus A. Schmidt. Aulacodiscus heterostictus is distinguished by the markings of the median zone and the round instead of triangular central zone. One specimen of A. schmidtii figured by Glezer et al. (1974, plate XIV, fig. 3) has striking similarities to A.heterostictus; A. schmidtii may be a synonym of A. septus as specimens of all “taxa” appear in the same material (“Singliliewsky”, almost certainly the Sengilei diatomite; Strel'nikova et al., 2004). Strel'nikova et al. (2004) considered that these are synonyms and that variants (“formae”) with four, five or six rimoportulae are morphological variants within this assemblage as they occur in the same outcrop and there is no age variation between them.

Found in the middle Eocene of the North Sea in pyritised form. Known informally as “Sp. CN 22” (British Petroleum).

BP well 16/16a-3, 1322.7m (core), central North Sea from the BP collection of diatom morphotypes (now housed in the Jagiellonian University, Krakow, Poland). North Sea (Mitlehner, 1994; British Petroleum, unpublished data).

Aulacodiscus insignis Hustedt in A. Schmidt et al. (1944: plate 458, figs. 3–6). Plate 1, figs. 5, 6, 10–12.

Valves round or quadrate with convex sides and with a narrow hyaline margin, diameter 125–250 µm. Surface with large polygonal elevation, angles usually four, rounded, extending nearly to margin, sides arcuate, with steep slopes to margin. Rimoportulae four to five short stout tubes, projecting outwardly from small clear areas at angles of elevation. Markings on top of elevation subcircular or polygonal cellules, 2.5 in 10 µm, arranged in radiating lines. Cellules with internal pores, forming pseudoloculate structure, arranged in horizontal lines; larger cellules with eyespot and puncta, 20 in 10 µm.

Barker and Meakin (1947) were unaware of Aulacodiscus insignis, described by Hustedt (1944), when they described Aulacodiscus aemulans; this may have been due to a lack of scientific correspondence between them during World War 2. Later, Burke and Woodward (1971, p. 332) discussed the synonymy and used Hustedt's name Aulacodiscus insignis. A. aemulans Barker and Meakin is therefore a junior synonym. SEM studies of specimens of A. insignis yielded evidence that supports its subdivision into three variants. Further studies of well-preserved specimens (in original opaline silica) are required in order to formally erect varieties and so the most important of these are referred to here under morphotypes “aemulans” and “quadrata”.

Aulacodiscus insignis morphotype “aemulans”; Plate 1, figs. 5, 6.

Valves round, diameter 100–120 µm; large, polygonal elevation on valve face with four to five rimoportulae. Cellules often indistinct, with meshwork of small areolae visible, five to six in 10 µm.

King (in King and Hughes, 1983) named the pyritised form with five nodes or processes “Diatom sp. 4” and described it thus: “its valve is circular, very thin and slightly convex in girdle view, but with five small equally spaced projections present near the valve margin” (King. 1983, p. 20).

Valves with five processes commonly occur in the upper Oligocene and lower Miocene of the central North Sea (King, 1983; Schiøler et al., 2007).

BP well 3/25-1, 2162m (core), central North Sea from the BP collection of diatom morphotypes (now housed in the Jagiellonian University, Krakow, Poland).

Upper Oligocene to lower Miocene: North Sea (Mitlehner, 1994; Schiøler et al., 2007; this study). Sufficiently widespread to form a zonal marker for Zone NSP 10 in the North Sea microfaunal zonation of King (1983). Early Eocene: Russia, Volga Basin, Kamyshin Formation (Oreshkina and Aleksandrova, 2007); “Carlovo” and “Kamischev”, now Kamyshlov, Sverdlovskaya District (Barker and Meakin, 1947; Burke and Woodward, 1963–1974).

Aulacodiscus insignis morphotype “quadrata”; Plate 1, figs. 10–12.

BP well 21/296-7, 5280', central North Sea from the BP collection of diatom morphotypes (now housed in the Jagiellonian University, Krakow, Poland).

Valve quadrate in valve view, rectangular in girdle view, with heterovalvy. A clear angle of approximately 95∘ occurs between the sides, to form developed corners. Mantle narrow, with a distinct flange which extends around the valve circumference. Four processes, clearly produced, project at an inward angle of 45∘ at each corner, these are only present on one valve. Areolae large, hexagonal, five to six in 10 µm.

Valve diameter 120 µm (upper Oligocene and lower Miocene forms); 180 µm (upper Oligocene forms).

Aulacodiscus insignis morphotype “quadrata” has never been recovered in original opaline silica (Philip Copestake, personal communication, 1995). However, a remarkably similar form was illustrated in Schmidt's Atlas (1874–1959, plate 99, fig. 24). It was collected during the Gazelle expedition of 1874–8 by A. Janisch, who figured a drawing of a diatom which he referred to as Aulacodiscus subrotundata in a set of plates, distributed privately in 1904 and without any text. Other diatom taxa depicted from the same expedition have a distinctly modern aspect to them (Patricia A. Sims, personal communication, 1994) although it is possible that the highly robust frustules were reworked from older sediment, as some species from the Gazelle expedition are also known from the upper Eocene of Barbados (Rattray, 1889). In the absence of an accurate and valid description, it was decided to formally erect a new variety herein.

BP well 15/28a-3; BP well 21/296-7; BP well 21/10-1, central North Sea, from the BP collection of diatom morphotypes (now housed in the Jagiellonian University, Krakow, Poland).

Lower Oligocene to lower Miocene: North Sea (Mitlehner, 1994; Schiøler et al., 2007; this study). This highly distinctive diatom has a sufficiently widespread occurrence to form a zonal marker (Zone NSP 9) in the North Sea microfaunal zonation of King (1983). Larger specimens (up to 180 µm) are restricted to the upper Oligocene; smaller specimens range throughout the upper Oligocene and lower Miocene (Schiøler et al., 2007).

Aulacodiscus singiliewskyanus Barker et Meakin (1944: 251, plate 38, fig. 3.). Plate 1, fig. 7.

Valve circular. Processes five, rarely four and six, projecting from small circular clear areas on broad radial inflations extending to the border from the angles of a polygonal elevation. Primary rays distinct, extending from the clear areas on the inflations to the sides of a small roughly outlined polygonal clear area at the centre of the surface. Markings small rounded granules, close together in radial rows, confused near the centre. In the compartments between the processes, the surface near the border is depressed and then rises into an inwardly curved narrow mound.

A form always encountered in pyritised form in the North Sea and adjacent onshore deposits (Mitlehner and Hart, 2010), usually with five rays but sometimes with six. All of the features described by Barker and Meakin (1944) are clearly visible except for the primary rays, which extend along the inflations in unpyritised specimens.

Barker and Meakin (1944) and Ross and Sims (1970) commented on the close resemblance of this species to A. tuberculatus Pantocsek, a form found in the same material, from Sengilei in the Volga Basin, Russia. Ross and Sims (1970) considered that the two may be endmembers of a continuous series, a phenomenon often encountered in diatom populations. Strel'nikova et al. (2004) regarded A. tuberculatus and A. archangelskianus Witt as senior synonyms of A. sengiliewskyanus Barker et Meakin. However, A. archangelskianus ranges from the Late Cretaceous to the late Palaeocene whereas A. singiliewskyanus has been found exclusively in deposits of middle Eocene age (Mitlehner, 1994; Mitlehner and Hart, 2010). Accordingly, the latter designation is kept for the morphologies encountered in this interval during the present study.

BP well 208/15-1, 1315 m, northern North Sea in the BP collection of diatom morphotypes (now housed in the Jagiellonian University, Krakow, Poland).

Middle Eocene: Russia, Volga Basin, Sengilei (Barker and, Meakin, 1944; Ross and Sims, 1970). Middle Eocene: Marsh Farm Formation, Isle of Wight, UK (Mitlehner and Hart, 2010).

Aulacodiscus subexcavatus Hustedt (1944: plate 460, figs. 3–13). Plate 1, Fig. 1.

Forms with three radial elevations, at the summits of which are processes. Valve margin has small radial striae, central area plain. Surface with radial rays of round, free pores, five to six in 10 µm. This morphology has gross similarities with A. excavatus A. Schmidt. It is distinguished by its more prominent valve relief (strongly indented between the elevations) and the situation of its large, pear-shaped processes. The latter are positioned close to the valve margin (translated from Benda, 1965, p. 171–2).

Valve diameter 85–120 µm.

A. subexcavatus appears to be conspecific with Craspedoporus actinoptychoides Glezer. The latter was figured from the Eocene of Russia by Glezer et al. (1974) but not described in detail. Glezer's figured specimen is strongly suggestive of A. subexcavatus, the type specimen of which came from the area of the Volga Basin (Hustedt, 1944). The name appears to be based on Auliscus actinoptychoides published 10 years earlier by Glezer (Glezer and Posnova, 1964, p. 64, Taf. III, fig. 1, which is the same illustration as Glezer, 1974, plate XXXII, fig. 10). However, there is no publication that makes the formal transfer from Auliscus to Craspedoporus and so this name is thus considered invalid (David M. Williams, personal communication, 2013).

BP well 208/15-1, 1220 m, Horda Formation, northern North Sea from the BP collection of diatom morphotypes (now housed in the Jagiellonian University, Krakow, Poland).

Lower Eocene: Denmark, Fur Formation (Benda, 1972); North Germany (Benda, 1965); Russia, “Kamischev” now Kamyshlov, Sverdlovskaya District (Hustedt, 1944; Glezer et al., 1974); Eocene: Horda Formation, North Sea Basin (Mitlehner, 1994); middle Eocene: Marsh Farm Formation, Isle of Wight, UK (Mitlehner and Hart, 2010).

Aulacodiscus suspectus A. Schmidt (1876, plate 36, figs. 17, 18). Plate 1, figs. 8, 9.

Valve outline circular, flat or slightly convex. Central area poorly defined, with no central rosette. Areolae polygonal, arranged in radial lines away from the centre (translated from Homann, 1991, p. 38). Shape flattened, discoidal with a low mantle. Central area hyaline. Areolae arranged radially in rows, two to seven in 10 µm near the valve centre, declining markedly in size at the margin (9–10 in 10 µm). An area of larger areolae occurs mid-valve. Between two and six hyaline rays are visible between the rows of areolae, at the end of each is a labiate process. Micro-labiate processes are sometimes visible both at the valve margin and on the valve face (Plate 1, fig. 9).

Valve diameter 140 µm.

A. suspectus resembles Aulacodiscus probabilis A. Schmidt by the lack of valve face undulation but differs from it by the smaller size of the external expression of the rimoportulae. The synonymy of this taxon reflects the fact that the true delineation between Coscinodiscus and Aulacodiscus is not perfectly clear at present and requires further research (Strel'nikova et al., 2004). To date, no formal comparison of the holotypes of Aulacodiscus suspectus and Coscinodiscus josefinus has been carried out. Strel'nikova (1997) erected a new genus, Moisseevia, based on a superficially similar taxon from Palaeogene deposits in Russia. However, this is morphologically much closer to Aulacodiscus probabilis as it has larger expression of external fultoportulae (see Oreshkina and Aleksandrova 2007, plate II, 11).

In pyritised specimens, the areolae are filled in with pyrite to form raised “bosses” or papillae, arranged radially over the valve surface in the same manner as the original areolae. This is a useful feature for identification during the picking of microfauna from micropalaeontological residues. Other features, such as the micro-labiate processes, are only readily observable via SEM.

BP wells 15/28a-3 and 21;/9-1, central North Sea, Sele Formation; Knudeklint Member, Fur Formation, Denmark.

Lower Eocene: Denmark, Fur Formation (Benda, 1972; Homann, 1991; Mitlehner, 1996); North Sea, Sele and Balder formations (Mitlehner, 1994, 1996); Russia, Barents Sea (Grunow, 1884); Sengiley (Oreshkina and Oberhänsli, 2003).

Order Hemialulales Round et Crawford, in Round et al. 1990

A very characteristic triangular or quadrate genus, often confused with Triceratium, from which it is distinguished by its corner spines, larger size and radial areolae. Some species of Trinacria display considerable morphological variation, in both valve shape and ornament. The extensive and varied synonymy for some species reflects this, and it is only recently that these variations have been studied in detail. SEM examination of well-preserved, non-pyritised specimens from the Fur Formation by Mitlehner (1994) showed that many of the “varieties” of the species figured by Grunow, Schmidt and others and discussed by Homann (1991) are in fact different valves which interlock within a chain. An exclusively fossil genus, with species ranging from the Upper Cretaceous (Maastrichtian) to upper Miocene (Sims and Ross, 1988). Following the investigations of Mitlehner (1994), both the description and synonymy of Trinacria are emended herein.

Trinacria regina (Heiberg) Homann, emend. Plate 2, figs. 1–11. Heiberg (1863, 50, plate 3.)

Valve tripolar, with more or less concave, rarely almost straight sides with broadly rounded poles, drawn out into a short, narrow point. Valve centre flat to highly convex, specimens with the latter sometimes having a concave depression at top; rare teratological specimens with an eccentric curvature; central elevation substantially lower than the elevations which project from the corners. Elevations straight, narrow, usually tall but sometimes very short. Occasionally the apices of the elevations have a thickened, rounded top. Horn apices with two to three short strengthened spines, circa 8–15 µm long; these interlock with similar spines on adjacent cells to form an enclosed, interlocking chain. Valve margin sharply pronounced, distinctly alar (winged), proceeding more or less parallel to the base of the mantle valve. The “wings” are drawn up to the height of the elevations. Valve mantle low, 5–25 µm high, shaped around the base of the valve. The sharp edge of the mantle, combined with the extended alar mantle, often gives the impression of a “double edge” when seen in valve view. Detailed SEM examination of well-preserved specimens in original opaline silica from the Fur Formation of northern Denmark (see outcrop photographs, figs. 2-d) during the study of Mitlehner (1994) shows that many of the varieties of this species, figured in LM (light microscope) only by Homann, are in fact different valves which interlock within a chain (Plate 2, figs. 10–11). In addition, some of the areolae are seen to be ornamented with small spines (see Round et al., 1990, p. 269, fig. h). The “alar” projections of Homann are more properly termed marginal ridges (Patricia A. Sims, personal communication, 1994).

Length between poles variable, from 60 to 230 µm. Areolae three to six in 10 µm, becoming smaller up the elevations.

Trinacria regina displays considerable morphological variation, both in valve shape and ornament. The extensive and varied synonymy for this species reflects this, and it is only relatively recently that these variations have been studied in detail (Homann 1991, Mitlehner 1994; this study). Homann (1991) considered that specimens with an eccentric central areolation (see Homann 1991, plate 50, fig. 7) should be regarded as teratological (i.e. ecophenotypic) variants, whilst other morphologies formerly assigned species names (“jordanii”, “punctulata”, “wittii”) are now regarded as varieties of T. regina.

This species is one of the most abundant taxa in both offshore and onshore samples studied and often dominates assemblages in the lowermost Eocene Balder (offshore) and Fur (onshore) formations (see Mitlehner, 1996, for biostratigraphic correlation of these and other age-equivalent strata around the North Sea Basin). Although often pyritised or calcified, enough of the diagnostic characters described above are preserved to allow confident species designation, and although only the bases of the corner elevations are often preserved, in some cases complete projections are seen with the whole specimen completely replaced by pyrite (Plate 2, fig. 7; see De Jonghe et al., 2011, for a discussion on the pyritisation process). However, some detail such as rimoportulae (Plate 2, fig. 5) are not clear in pyritised specimens. In some cases the valves are deeply concave, which on closer inspection are separation valves, found at the terminal ends of each chain (Plate 2, fig. 10). In most cases the edge of the mantle is very prominent, appearing slightly concave in some specimens (Plate 2, fig. 1). Some specimens exhibit fine, radial striae on the valve face, originally named Trinacria jordanii by Schmidt and renamed “Typ jordanii” by Homann (Homann 1991, plate 51, fig. 5); others exhibit three radial, hyaline rays, which Homann termed “Typ wittii” after Benda's (1972) original description as Trinacria wittii. Malm et al. (1984, fig. 8j) figured the latter variant from an offshore oil well in the Norwegian Sector of the North Sea, in the Lower Eocene Balder Formation. In the lower part of the age-equivalent Fur Formation of Denmark (Knudeklint Member, Pedersen & Surlyk 1983), T. regina is extremely abundant in some samples examined by the present author, particularly around ash layers -20 to -6 of Bøggild (1918), equivalent to the Knudeklint Member; this may be due to blooms during a period of intense eutrophication during the Initial Eocene Thermal Maximum (Schoon et al., 2015). In pyritised form, specimens have hitherto been designated as Triceratium sp. 1 Bettenstaedt et al., 1962 (Malm et al., 1984, fig. 8i). Other morphotypes in use in the offshore oil exploration industry are variants of T. regina (for example, “Triceratium spp. 30, 31, 55, 58” of Robertson Research and others).

Lower Eocene: BP wells 15/28a-3 and 21;/9-1, central North Sea, Sele Formation; base of Ieper Formation, Knokke borehole, Belgium; Ølst Formation, Denmark; upper part of the Knudeklint Member, Fur Formation, Denmark.

Palaeocene: Cape Basin, South Atlantic (Gombos 1984, DSDP 73); Volga Basin, Russia (Glezer et al., 1974). Kerteminde Marl (Bignot 1983). Lower Eocene: Barents Sea (Grunow 1884); Denmark, Fur Formation (Benda 1972; Homann 1991); Sengiley, Middle Povolzhie (Oreshkina and Oberhänsli 2003); North Sea, Sele Formation (King 1983; Malm et al. 1984; Mitlehner 1996); England, London Clay, Harwich Member (Shrubsole and Kitton 1881; King 1981); North Germany, Unter Eozän (Benda 1965); Northern France, Sparnacian (Bignot 1983; Van Eetvelde and Cornet 2002); southern North Sea (Hughes 1981).