1. The mode of life of Bohemilla

In his insightful review of the evidence for recognising a pelagic mode of life for some trilobites, Fortey (Reference Fortey1985, pp. 222–23) considered that the Bohemillidae is one of the groups of trilobites that had a mesopelagic habit (see also Fortey & Owens Reference Fortey and Owens1999, pp. 452–53, fig. 19). He noted (Fortey Reference Fortey1985, p. 223) the occurrence of bohemillids with the cyclopygid biofacies close to the edges of Ordovician continental shelves at or about the lower limits of the photic zone, around 200–700 m (but see Zhou et al. Reference Zhou, Zhou, Siveter and Yuan2003 for suggestions of somewhat shallower depths), possibly migrating through the water column diurnally. Fortey & Owens (Reference Fortey and Owens1999, pp. 452–53, fig. 19) suggested that Bohemilla and other poorly streamlined mesopelagic trilobites may have fed on smaller and slower-moving elements of the plankton.

Fortey & Owens (Reference Fortey and Owens1987, pp. 105–07) described the occurrence of the pelagic cyclopygid fauna (including Bohemilla) in the Arenig of South Wales in association with a benthic fauna of trilobites with no or shrunken eyes which they termed an atheloptic assemblage. These co-occurring assemblages of trilobites (with or without Bohemilla) have been recognised from outer shelf–upper slope settings from much of the Ordovician from many parts of the world (e.g., Zhou et al. Reference Zhou, Zhou, Siveter and Yuan2003, pp. 288–89; Fortey in Adrain et al. Reference Adrain, Edgecombe, Fortey, Laurie, McCormick, Owen, Waisfeld, Webby, Westrop, Zhou, Webby, Droser and Paris2004, pp. 250–51, 253; Owen & Bruton Reference Owen and Bruton2012, p. 964; Fortey Reference Fortey2014, pp. 253–54) including the Upper Whitehouse Subgroup at Girvan from where the present material of Bohemilla is described (Ingham & Fortey Reference Ingham and Fortey2023, pp. 314–15).

Fortey & Owens (Reference Fortey, Owens and McNamara1990, pp. 136–37, fig. 5.5) noted that trilobites with large eyes, reduced pleurae and occurrence in offshore biofacies represent a recurring morphotype through the Ordovician. They included, with some uncertainty, some Cambrian forms. With its highly reduced thoracic pleurae, Bohemilla conforms to this morphotype and Fortey & Rushton (Reference Fortey and Rushton2007) described a species of the middle Cambrian trilobite Beishanella Xiang & Zhang, Reference Xiang and Zhang1985 from South Wales, which they considered to show considerable morphological convergence with Bohemilla, thus confirming the likely development of a pelagic life habit in the Trilobita prior to the Ordovician (see also Fatka et al. Reference Fatka, Szabad, Budil, Micka, Rábano, Gozalo and García-Bellido2008, p. 120).

2. Bohemilla in space and time

Given its deep-water habitat, there is an inevitable bias in the record of Bohemilla to those sites where such environments are reflected in the shelly-fossil-bearing rock record. That said, for much of its history, Bohemilla – both the nominate subgenus and Bohemilla (Fenniops) Fortey & Owens, Reference Fortey and Owens1987 – seems to have been restricted to the fairly high latitude margin of Gondwana and, at slightly lower latitudes, Avalonia. The exception may be the intriguing record of Bohemilla (Fenniops) (along with the cyclopygid Prospectatrix) in association with more typical Baltic faunas in the Lower Ordovician (‘Tremadoc-Arenig’) at the southern end of the Urals noted by Cocks & Fortey (Reference Cocks and Fortey1998, p. 150). Only towards the end of its known history (middle Katian = upper Caradoc to lower Ashgill) did Bohemilla reach the margins of low-latitude Laurentia, where it is represented by the species of B. (Bohemilla) described herein and a probable species of the subgenus in the Central Terrane in Ireland. Bohemilla is not known from China, despite the widespread occurrence of the cyclopygid biofacies at various levels in the Ordovician there (e.g., Zhou Z.-Y. et al. Reference Zhou, Zhou, Siveter and Yuan2003; Zhou Z.-Q. et al. Reference Zhou, Zhou and Yuan2007). The reasons for this are not clear.

The oldest occurrences of Bohemilla are of an indeterminate species of B. (Fenniops) listed from the upper Tremadocian of the Montagne Noire, France (Vizcaïno et al. Reference Vizcaïno, Álvaro and Lefebvre2001; Vizcaïno & Álvaro Reference Vizcaïno and Álvaro2003) and material described as Bohemilla (Bohemilla) sp. by Rushton (Reference Rushton1988) from the upper Tremadoc of the English Lake District that may be intermediate between B. (Fenniops) and B. (Bohemilla) (see discussion of subgenera below). Elsewhere on Avalonia, B. (F.) sabulon Fortey & Owens, Reference Fortey and Owens1987, the type species of the subgenus, occurs in the upper Arenig, Fennian Regional Stage (= upper Dapingian–lowest Darriwilian) of Wales and B. (Bohemilla) tridens Rushton & Hughes, Reference Rushton and Hughes1981 is known from the Llanvirn (middle to upper Darriwilian) of England.

Bohemilla is known from the upper part of the Middle Ordovician (Darriwilian–uppermost Arenig) to the middle Katian (upper Caradoc to lower Ashgill) in the Czech Republic (Mergl et al. Reference Mergl, Fatka, Budil, Rábano, Gozalo and García-Bellido2008; Fatka & Mergl Reference Fatka, Mergl and Bassett2009; Fatka et al. Reference Fatka, Budil and David2015). Most species belong in B. (Bohemilla) (see discussion of B. (B.) scotica below) with species of B. (Fenniops), B. (F.) praecedens Klouček, Reference Klouček1916 and B. (F.) klouceki Marek, Reference Marek1966 being restricted to the lower part of that range: lower Darriwilian and upper Darriwilian to lower Sandbian respectively (uppermost Arenig to lower Caradoc).

The youngest occurrence of B. (Bohemilla) in the Czech Republic is broadly coeval with the species from Scotland, B. (B.) scotica Reed, 1914 and B. (B.) xesta sp. nov. described herein and the likely species from Ireland (see discussion below). The Scottish and Irish species represent the extension of the range of the genus to the Iapetan margin of Laurentia. As with the other members of the hitherto highly successful, long-ranging, cyglopygid biofacies, Bohemilla did not survive into the Silurian (Fortey in Adrain et al. Reference Adrain, Edgecombe, Fortey, Laurie, McCormick, Owen, Waisfeld, Webby, Westrop, Zhou, Webby, Droser and Paris2004, p. 250).

3. Bohemilla in the Upper Ordovician of Girvan

Bohemilla appears in the Upper Ordovician of the Girvan district, SW Scotland, late in the history of the genus, at levels around the Caradoc–Ashgill boundary (middle Katian in terms of international chronostratigraphy). B. (B.) xesta sp. nov. occurs in the Red Mudstone Member of the Myoch Formation and B. (B.) scotica Reed, 1914 in the stratigraphically higher Gray Member of the Mill Formation. These deep-water formations are part of the Upper Whitehouse Subgroup of the Ardmillan Group (Fig. 1) within one of the classical Ordovician siliciclastic successions (Ingham in Fortey et al. Reference Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000, pp. 42–47; Harper et al. Reference Harper, Meidla, Servais, Harper, Lefebvre, Percival and Servais2023, p. 17). Bohemilla is a rare component of the shelly faunas, with their disarticulated sclerites comprising less than 1 % of about 2,500 trilobite specimens known from these formations.

Figure 1

Simplified representation of the lithostratigraphy and its correlation of part of the middle Katian Whitehouse Subgroup exposed along a 2 km stretch of foreshore southwest of Girvan, Ayrshire, SW Scotland (adapted from Ingham & Fortey Reference Ingham and Fortey2023, fig. 1). Numbered sample horizons, including those yielding specimens of Bohemilla, are indicated; their detailed locations on geological maps will be published elsewhere.

Figure 2

Bohemilla (Bohemilla) scotica Reed, Reference Reed1914. All figured specimens are from the Gray Member of the Mill Formation (lower Ashgill) at locality W9 on the Whitehouse foreshore. (a) Latex cast from external mould of lectotype cranidium NHMUK In21694, previously figured by Reed (Reference Reed1914, pl. 4, fig. 1). (b) Internal mould of cranidium NHMUK In36987, referred to by Whittard (Reference Whittard1952, p. 319). (c) Latex cast from external mould of almost complete paralectotype cranidium NHMUK In21690; a reconstruction, certainly based on this specimen was figured by Reed, Reference Reed1904, pl. 8, fig. 4) as Bohemilla sp. (d) Latex cast from external mould of small cranidium NHMUK In36986. (e) Latex cast from external mould of small, partial cranidium NHMUK In21697. (f) Latex cast from external mould of damaged cranidium NHMUK In21695, showing anterior margin particularly well. (g) Internal mould of partial cranidium GLAHM A10237, clearly showing mesial furrow on frontal glabellar lobe. (h) Broken and somewhat crushed internal mould of cranidium GLAHM A6450. (i) Internal mould of hypostome NHMUK In20978, mistakenly included with Dionide richardsoni by Reed, Reference Reed1903. p.27, described but not figured by him). Scale bars represent 2 mm (a, b, c, f, g, h); 1 mm (d, e, i).

Figure 3

Bohemilla (Bohemilla) xesta sp. nov. All figured specimens are from the Red Mudstone Member of the Myoch Formation (upper Caradoc) of the Whitehouse (W localities), Port Cardloch (P localities) and Myoch (M localities) foreshores. (a) Latex cast from external mould of broken holotype cranidium GLAHM A6538b, locality P1. (b) Silicone cast from external mould of paratype cephalon, GLAHM A132434b, locality W3; anterior part of cranidium has been pushed rearwards giving an unusually blunt outline. (c) Internal mould of frontal part of paratype cranidium GLAHM A9215, locality M4, a few centimetres below the Trilobite Bed at the base of the Wall Member. (d) Internal mould of broken paratype cranidium GLAHM A6189a, locality W8. (e) Silicone cast from external mould of paratype partial hypostome, GLAHM A132599, locality W3. Scale bars represent 2 mm.

The stratigraphy of the Upper Whitehouse Subgroup along the foreshore south of Girvan is summarised in Figure 1, which shows the sample horizons within the succession. Very detailed geological maps made by one of us (J.K.I.) which show the spatial distribution of the sample localities will be published elsewhere. Suffice it to note here that, as outlined by Ingham & Fortey (Reference Ingham and Fortey2023, pp. 314–15), the rocks are both tectonically complicated and show a great deal of syn-sedimentary complexity, including channel deposits, diachronous lithostratigraphical boundaries and evidence of local removal of sediment through penecontemporaneous down-slope displacement. The reddish and green silty mudstones of the Myoch Formation are interpreted as representing the over-bank fines of channel deposits. The overlying Mill Formation may be an even deeper water succession with the darker and lighter banded greenish-grey silty shales of the Gray Member yielding fairly common graptolites as well as shelly fossils.

Both the Myoch Formation and the Gray Member of the Mill Formation contain a mesopelagic cyclopygid fauna and an atheloptic benthic assemblage and thus the occurrence of Bohemilla conforms to much of its earlier history. The overall generic composition of the trilobite fauna was listed by Ingham & Fortey (Reference Ingham and Fortey2023, pp. 314–15), who noted that the atheloptic fauna of the Myoch Formation is thought to be essentially in situ whereas that of the Gray Member of the Mill Formation was largely washed in along with elements of the pelagic fauna.

4. Systematic palaeontology

The described material is held in the Hunterian Museum, University of Glasgow, Scotland, UK (specimen numbers prefixed GLAHM) and the Natural History Museum, London, England, UK (NHMUK). The illustrated specimens were lightly coated in magnesium oxide prior to photography.

The terminology applied to the dorsal exoskeleton is that advocated by Whittington & Kelley (in Whittington et al. Reference Whittington, Chatterton, Speyer, Fortey, Owens, Chang, Dean, Jell, Laurie, Palmer, Repina, Rushton, Shergold, Clarkson, Wilmot and Kelly1997, pp. 313–29). Some of that for the hypostome is more circumspect. The morphology of the very distinctive hypostome of Bohemilla is most completely displayed in a specimen of B. (B.) tridens illustrated by Rushton & Hughes (Reference Rushton and Hughes1981, pl. 5, fig. 10), but is consistent in those other species from which a hypostome is known, including those from Girvan (below). An undivided anterior swollen lobe of rounded pentagonal outline is defined posteriorly by a well-incised furrow and occupies about three-quarters of the length of the hypostome. Posterior to this is an area of much lower convexity bounded laterally by parallel or inwardly curved borders and posteriorly by a transverse or very gently forwardly curved furrow alongside a narrow (sag., exsag.) posterior border. A short posterolateral spine is present in most of the species from which the hypostome is known.

Rushton & Hughes (Reference Rushton and Hughes1981, p. 639) took a cautious view of the area posterior to the middle body in B. (B.) tridens and simply termed it the ‘posterior portion’ of the hypostome comprising a ‘U- or V-shaped ridge’, a dorsally directed ‘pair of depressed posterolateral processes’ and a ‘convex posterior border’. They also noted the presence of a pair of strong swellings near the anterolateral ends of the ridge. Strong swellings are also present in B. (B.) roxana Šnajdr, Reference Šnajdr1982. They are less well developed in B. (F.) klouceki Marek, Reference Marek1966 and the Girvan species of Bohemilla (Bohemilla) described below; they seem to be absent in B. (B.) pragensis Marek, Reference Marek1966.

The presence of a discrete posterior and lateral border precludes the whole area posterior to the anterior swollen area of the hypostome of Bohemilla belonging entirely to the posterior border (cf. Marek Reference Marek1966 and Šnajdr Reference Šnajdr1982). It is not clear, however, whether it is a highly unusual development of the posterior lobe of the middle body of the hypostome or whether it is an entirely separate structure. If it reflects the posterior lobe in some way, the swellings may be equivalent to the maculae in ‘typical’ hypostomes, although they are a little more posteriorly placed in terms of what would be the middle furrow of the middle body. In view of the uncertainty surrounding the posterior part of the hypostome, the cautious approach to the terminology applied by Rushton & Hughes (Reference Rushton and Hughes1981) is adopted here.

Superfamily Cyclopygoidea Raymond, Reference Raymond1925

Family Bohemillidae Barrande, Reference Barrande1872

Remarks. Reed (Reference Reed1914, p. 23) speculated that Bohemilla might not be a trilobite and may belong to ‘some primitive section of the Arachnida’ and Whittard (Reference Whittard1952, p. 320) concluded that it should be excluded from the Trilobita although it could not readily be assigned to any other group of arthropods. Henningsmoen, in the first edition of the ‘Trilobite Treatise’ (Moore Reference Moore1959, pp. O525–26), considered the Bohemillidae under ‘Supposed Trilobita here rejected from Class’. Conversely, Marek (Reference Marek1966) concluded that Bohemilla is a trilobite; a view followed by all subsequent authors. He considered what he termed the ‘Superfamily Bohemillacea’ to be related to the Remopleuridacea Hawle & Corda, Reference Hawle and Corda1847, with closest affinity to the subfamily Remopleuridinae. Fortey (Reference Fortey1974, p. 120) considered the Bohemillidae to be a family within the Remopleuridacea and Rushton & Hughes (Reference Rushton and Hughes1981) argued that Bohemilla evolved from an early genus of remopleuridid by ‘forward and inward migration of the eyes, reduction of the preocular cheeks and the forward swivelling of the postocular cheeks’. Similarly, Fortey & Owens (Reference Fortey and Owens1987, p. 129) favoured a likely remopleuridacean origin for Bohemilla and noted that ‘there is no feature of Bohemilla which indicates that it should be included in the Cyclopygacea’. Fortey & Chatterton (Reference Fortey and Chatterton1988, pp. 202–03) provisionally included the Bohemillidae within the Remopleuridacea, an assignment adopted in the revised edition of the Treatise (Fortey in Whittington et al. Reference Whittington, Chatterton, Speyer, Fortey, Owens, Chang, Dean, Jell, Laurie, Palmer, Repina, Rushton, Shergold, Clarkson, Wilmot and Kelly1997, p. 301).

More recently, Adrain et al. (Reference Adrain, Peters and Westrop2009, p. 39) argued that supposed remopleuridid characters of Bohemilla are not primitive features of the remopleuridids, rather that they are present in highly derived forms. Moreover, they noted that while remopleuridids have a thorax of 11 or 12 segments with a median spine on the eighth, it seems likely that the thorax of Bohemilla (Fenniops) sabulon Fortey & Owens, Reference Fortey and Owens1987 lacks a median spine and apparently has only seven segments, although the most complete known thorax lacks an attached pygidium (see Fortey & Owens Reference Fortey and Owens1987, p. 131, figs 22e, 23). It should also be noted that the lectotype specimen of the type species of Bohemilla, B. (Bohemilla) stupenda Barrande, Reference Barrande1872 has six segments posterior to the cephalon, the small, incompletely preserved, posterior one possibly being the pygidium (Whittard Reference Whittard1952, p. 318, pl. 33, figs 11, 12; Marek Reference Marek1966, p. 148, pl. 1, fig. 1) and there is no median spine. Adrain et al. (Reference Adrain, Peters and Westrop2009, p. 39) suggested that bohemillids are unlikely to be remopleuridoideans and may instead be cyclopygoideans, a view cautiously adopted here. Adrain (Reference Adrain, Harper and Servais2013 pp. 318–19) subsequently went further and considered the group to be a subfamily of the Cyclopygidae. Bohemilla was not included in the recent phylogenetic analysis of the Cyclopygacea by Braddy et al. (Reference Braddy, Dale and Wang2025).

Genus Bohemilla Barrande, Reference Barrande1872

Type species. Bohemilla stupenda Barrande, Reference Barrande1872, from the Dobrotivá Formation (upper Darriwilian–lowest Sandbian = uppermost Llanvirn–lowest Caradoc) of the Czech Republic. Original designation.

Subgenus Bohemilla (Bohemilla) Barrande, Reference Barrande1872

Remarks. In establishing the subgenus Bohemilla (Fenniops), Fortey & Owens (Reference Fortey and Owens1987, p. 128) noted that in Bohemilla (Bohemilla) the anterior part of the glabella is somewhat constricted in front of the S2 furrows and framed by long palpebral lobes that extend from adjacent to L3 to a level almost at the front of the glabella, the glabellar furrows extend further toward the glabellar midline such that the median glabellar lobe is less than one-third the glabellar width and the postocular fixigenae are acutely angled forwards and are thus very narrow. In contrast, Bohemilla (Fenniops) (type species Bohemilla (Fenniops) sabulon Fortey & Owens, Reference Fortey and Owens1987 from the Pontyfenni Formation, upper Arenig, Fennian Regional Stage (= upper Dapingian–lowest Darriwilian) of South Wales) was characterised by a wide frontal lobe posteriorly equal in width to the occipital ring, much shorter (exsag.) palpebral lobes symmetrically disposed about the S3 furrows and relatively wide postocular fixed cheeks. Fortey & Owens (Reference Fortey and Owens1987, p. 128) considered B. (Fenniops) to be the more primitive of the two subgenera on the basis of the ‘less specialised’ occipital ring and glabellar furrows and its wider fixed cheeks being ‘more comparable with other ptychoparioid trilobites’.

The differences between the subgenera noted by Fortey & Owens (Reference Fortey and Owens1987) are consistent in the named species of Bohemilla. Poorly preserved material described by Rushton (Reference Rushton1988, pp. 690–692, fig. 4c, pl. 68, figs 4, 7) as Bohemilla (Bohemilla) sp. from the upper Tremadoc Skiddaw Group of the English Lake District is more problematic. As noted by Rushton (Reference Rushton1988, p. 692), the free cheek has a small eye, as in B. (Fenniops). The partial cranidium is much harder to interpret. Only the left side is preserved and the frontal lobe of the glabella is unusually short (sag., exsag.) compared to the lengths of the lobes posterior to it. Together with what was interpreted as L3, the frontal lobe is bounded laterally by a ridge-like structure that is concave outwards and curves around the front of the glabella. The direction of curvature of the ridge alongside the anterior part of the glabella precludes it from being interpreted as a long palpebral lobe as in B. (Bohemilla). A slight swelling at the posterior end of the ridge might be interpreted as a short palpebral lobe as shown on the reconstruction illustrated by Rushton (Reference Rushton1988, fig. 4c) and is a feature of B. (Fenniops), but more posteriorly placed than is typical, as interpreted by Rushton in terms of the free cheek (1988, p. 692). There is a suggestion of a structure adjacent to the posterior part of the glabella, interpreted by Rushton (Reference Rushton1988, p. 692) as a postocular fixed cheek of similar shape to that of B. (Fenniops). Rushton considered that the only ‘comparatively advanced character’ is the anterior narrowing of the glabella, although this is more gradual along L3 rather principally occurring at the S2 furrow. It is tempting to consider this material to be intermediate between typical B. (Fenniops) and typical B. (Bohemilla), although it has more features of the former than the latter. Given its poor preservation it may be safest to refer to it as Bohemilla sp., pending the discovery of more convincing material.

The stratigraphical ranges of the subgenera of Bohemilla overlap, as is noted above. Unequivocal B. (Bohemilla) has a range from the Darriwilian to the middle Katian (uppermost Arenig to lower Ashgill in terms of the regional chronostratigraphy of Britain and Ireland) and B. (Fenniops) from upper Tremadocian to lower Sandbian (upper Tremadoc to lower Caradoc).

Bohemilla (Bohemilla) scotica Reed, Reference Reed1914

Figure 2a–i

Reference Peach and Horne1899 Bohemilla sp., Peach & Horne, pp. 517, 688.

1904 Bohemilla sp., Reed, p. 53, pl. 8, fig. 4.

Reference Reed1914 Bohemilla scotica sp. nov., Reed, p. 22, pl. 4, fig. 1.

Reference Whittard1952 Bohemilla scotica Reed: Whittard, p. 319.

Reference Morris and Fortey1985 Bohemilla scotica Reed: Morris & Fortey, p. 22.

Reference Morris1988 Bohemilla scotica Reed: Morris, p. 34.

Type material. Lectotype selected herein NHMUK In21694, one of two cranidia, taken as syntypes (see Morris Reference Morris1988, p. 34), known to Reed from the Gray Collection and figured by him in 1914. The second specimen, NHMUK In21690, figured by Reed, Reference Reed1904, pl. 8, fig. 4) as Bohemilla sp., becomes paralectotype. Both specimens were collected from the Gray Member of the Mill Formation of the Upper Whitehouse Subgroup (lowest Ashgill) at locality W9 on the Whitehouse foreshore (see Fig. 1).

Other material. All specimens from the Gray Member of the Mill Formation (earliest Ashgill) of the Whitehouse foreshore (locality W9): a cranidium (NHMUK In36987) referred to by Whittard (Reference Whittard1952, p. 319), three other cranidia (NHMUK In21695, In21697 and In36986) and a hypostome (NHMUK In20978) in the collections of the Natural History Museum, London, and four cranidia (GLAHM A6450, A6596, A10237, A10244/1) in the Hunterian Museum, University of Glasgow.

Diagnosis. Species of Bohemilla (Bohemilla) bearing a dense, coarsely granular glabellar sculpture, a distinct groove along the mid-line of the glabella, a robust spine on the posterior part of the occipital ring flanked by a transverse ridge bearing three short spines on each side, very prominent ridge- or spine-like tubercles on L2 and L3, a well-developed lenticular area on the anterior edge of the mid-part of L1, S1–3 extending across the glabella, frontal lobe with an evenly curved lateral and anterior margin extending only a very short distance in front of the evenly arched, narrow (tr.) palpebral lobes.

Description. All known cranidia are less than a centimetre long (sag.) and are somewhat crushed. Glabella transversely convex, occupies at least 95 % of the length of the cranidium and has a maximum width (at anterior part of L2) equal to about three-quarters of its sagittal length. Glabella expands very gently forwards to anterior part of L2 before narrowing abruptly in front of S2 and over the anterior half of L3, adjacent to the posterior part of the palpebral lobes, such that the width at the posterior edge of the frontal lobe is about four-fifths the maximum glabellar width. Occipital ring occupies almost a quarter of glabellar length. Posterior edge transversely directed or arched very gently rearwards mesially, curving gently forwards over most of the width of the ring before turning to become almost parallel to the sagittal line laterally. Ring delimited anteriorly by sinuous occipital furrow that is flexed forwards mesially and traces a broad, forwardly concave curve laterally. Ring bears a broad (tr.) median ridge with a small anterior swelling and a robust, postero-dorsally directed spine posteriorly (Fig. 2a, b). A transverse posterior ridge extends abaxially from a short distance beyond the median ridge, tapering out distally; bearing three short spines and bounded anteriorly by a well-incised, transversely directed furrow. A much shallower transverse furrow is developed at nearly half the distance between the posterior furrow and the deep occipital furrow, effectively separating two weakly swollen transverse ribs, the outer part of the anterior one of which expands (exsag.) markedly abaxially. Preoccipital glabella bears a median furrow that is well incised between the lateral glabellar lobes; less so on the anterior lobe of most specimens (e.g., Fig. 2c). L2 and L3 bear short, robust spines adjacent to this furrow. L1–3 transversely directed, together occupying about two-fifths of the sagittal length of the glabella and defined by furrows that extend across the whole glabella. L1 bears a transverse lenticular region of low convexity on the anterior edge of its mid-part, delimited by a very shallow furrow posteriorly and the deeper S1 anteriorly. S1 shallow and transversely directed mesially, deepening and arched forward around the lenticular region of L1; deep and transversely directed distally. L2 broadest (sag., exsag.) mesially, narrowing in front of the lenticular region of L1 before broadening slightly beyond this. S2 transversely directed, shallow mesially, deepening markedly abaxially. L3 of constant width (sag., exsag.) over all but the outermost part in most specimens. A shallow, transversely directed furrow is present over all but the innermost quarter of L3 in the lectotype cranidium (Fig. 2a) at about one-third the distance from its posterior edge. Anterior half of the outer edge of L3 curves adaxially posterior to the posterior tip of the palpebral lobe. S3 transversely directed, very shallow mesially, more deeply incised laterally. In one cranidium (Fig. 2b) a swollen median area on the posterior of the frontal lobe protrudes into L3 which narrows (sag., exsag.) posterior to it. S3 is very shallow and transversely directed posterior to this swollen area across the sagittal line before deepening and being directed abaxially forward at about 50 ^o and then turning sharply to become transversely directed again. There is a very weak transversely directed furrow across the middle of the swollen area, slightly further forward than the transverse outer parts of S3. Excluding this specimen, the frontal lobe occupies almost two-fifths of the glabellar length; sagittal length about two-thirds of posterior width; lateral margins gently curved adaxially; anterior edge almost transversely directed such that it extends only very slightly in front of the anterior ends of the palpebral lobes. A shallow furrow is present laterally on the posterior part of the frontal lobe on one specimen (Fig. 2c), but this may be a preservational artefact.

Postocular part of fixigena poorly known; developed as a narrow border bounded adaxially by a deep furrow adjacent to the occipital ring and L1–2, its connection to the palpebral lobe is not preserved. In small specimens, this border is essentially parallel to the sagittal line (Fig. 2d, e) but one of the larger specimens (Fig. 2f) shows a slight forward divergence. Gently arched palpebral lobes extend forward from adjacent to the anterior parts of L3 around the anterolateral parts of the frontal lobe such that their anterior ends are separated by a distance equivalent to almost two-thirds of the posterior width of the frontal lobe. Anterior border very narrow (sag., exsag.), very gently swollen laterally, more strongly so mesially where the preglabellar furrow deepens. Librigena not known.

Surface of glabella (including internal mould) densely covered by low granules except on the mesial, especially posteromesial, part of the frontal lobe where these swellings are considerably larger (Fig. 2a, b).

Hypostome (Fig. 2i) with anterior margin curved strongly forwards, broad (tr.) wings, lateral margin very gently adaxially curved and posterior border curved forward very gently with short posteriorly directed spines at its ends. Maximum width of hypostome, at anterior wings, very slightly less than its sagittal length. Strongly convex (sag., exsag.), well-rounded pentagonal middle body occupies a little over three-quarters of the sagittal length of the hypostome and has a maximum width equal to about three-quarters of its maximum width; defined by deep furrows posteriorly and alongside anterior of wings. Area between the middle body and the narrow posterior border gently swollen, defined laterally by a deep furrow and posteriorly by a furrow that deepens abaxially. Lateral borders narrow (tr.). Wings and borders smooth; middle body coarsely tuberculate, area between middle body and posterior border bears a few very weakly swollen tubercles including a slightly larger one close to the lateral furrow at a little less than half the distance between the middle body and posterior border.

Remainder of exoskeleton not known.

Remarks. The coarsely granular glabellar sculpture, sagittal glabellar groove, robust spine on the posterior part of the occipital ring, three short spines on a distinct ridge each side of the median occipital spine, very prominent tubercles on L2 and L3 distinguish B. (B.) scotica from the other described species of Bohemilla (Bohemilla).

The type species of Bohemilla, B. stupenda Barrande, Reference Barrande1872 from the Dobrotivá Formation (upper Darriwilian–lowest Sandbian = uppermost Llanvirn–lowest Caradoc) of the Czech Republic was redescribed by Whittard (Reference Whittard1952, pp. 316–20, pl. 33, figs 11, 12) and Marek (Reference Marek1966, pp. 147–48, pl. 1, figs 1–4, text-fig. 2); the latter also including a specimen from the underlying Šárka Formation (lower Darriwilian = uppermost Arenig–Llanvirn). In addition to the distinguishing features noted above, B. (B.) scotica differs from B. (B.) stupenda (which has one tubercle on the posterior of the occipital ring on each side of the sagittal line) in S3 extending across the glabella, in having a less strongly tapered frontal lobe and two transversely directed furrows on the lateral parts of the occipital ring compared to a very broad furrow defining the anterior edge of a weak ridge directed obliquely abaxially rearwards from near the sagittal line towards the posterior corner of the ring in B. (B.) stupenda.

Two younger species of Bohemilla (Bohemilla) have been described from the Czech Republic: B. (B.) pragensis Marek (Reference Marek1966, pp. 151–52, pl. 2, figs 10–12, text-figs 5, 6) and B. (B.) roxana Šnajdr (Reference Šnajdr1982, pp. 227–28, pl. 1, figs 6–8). The former was described by Marek (Reference Marek1966) as being from the Zahořany Formation, but Vaněk & Valíček (Reference Vaněk and Valíček2001, p. 47) considered it to be from the overlying Bohdalec Formation (middle Katian = upper Caradoc to lower Ashgill), the same unit as that from which B. (B.) roxana was described. Vaněk & Valíček (Reference Vaněk and Valíček2001, p. 47) considered B. (B.) roxana to be a junior synonym of B. (B.) pragensis. The material of both ‘species’ is limited and not well preserved, but differences noted by Šnajdr (Reference Šnajdr1982, p. 228) seem valid, with B. (B.) roxana having more prominent tubercles on L2 and L3, a strongly narrowed anterior portion of the frontal glabellar lobe (in one specimen; see below), a broader lateral border on the hypostome and a finer tuberculation of the middle body of the hypostome. There is also a pair of swellings on the anterolateral parts of the area posterior to the middle body. The cranidium of B. (B.) scotica differs from the material of B. (B.) pragensis illustrated by Marek (Reference Marek1966) in the distinguishing characters noted above and in its proportionally slightly longer frontal lobe in relation both to its posterior width and to the length of the glabella. The shape of the frontal lobe in B. (B.) roxana is different in the two specimens illustrated by Šnajdr with one (Šnajdr Reference Šnajdr1982, pl. 1, fig. 6) showing its anterolateral edge curving forward to produce a short ‘tongue’ between and slightly in front of the anterior ends of the palpebral lobes. The overall shape of the frontal lobe is semicircular in the other specimen (Šnajdr Reference Šnajdr1982, pl. 1, fig. 7) and thus similar in outline but proportionally shorter than that of B. (B.) scotica, which also differs in its coarse surface granulation; more clearly defined lenticular areas on L1 and the well-developed pair of median spines on L2 and L3 (cf. large tubercles). The occipital ring is incompletely preserved in the specimens of B. (B.) roxana illustrated by Šnajdr (Reference Šnajdr1982, pl. 1, figs 6, 7), preventing comparisons but the hypostome of B. (B.) scotica differs in its narrower lateral border, coarser granulation on the middle body and in having a much less prominent lateral swelling posterior to the middle body.

The best-preserved described species of Bohemilla (Bohemilla) is currently B. (B.) tridens Rushton & Hughes (Reference Rushton and Hughes1981, pp. 638–40, pl. 5, figs 10–16) from Llanvirn (essentially middle–upper Darriwilian) strata in the Great Paxton borehole in eastern England. The cranidium of B. (B.) scotica differs from this English species in the distinctive features noted above (note that B. (B.) tridens, as its name suggests, has three spines along the posterior border of the occipital ring) and in the outline of the frontal lobe (that of B. (B.) tridens is arched abaxially over its posterior part before becoming almost parallel-sided forming a blunt-ended, tongue-like frontal portion), in S3 being continuous across the glabella, in the narrower, less strongly curved, palpebral lobes that terminate abaxial to the anterior half of L3 (cf. its posterior edge) and, possibly, in its less divergent posterior portions of the postocular fixigenae. The hypostome of B. (B.) scotica is more elongate than that of B. (B.) tridens, has a more oval and coarsely tuberculate rather than a sub-pentagonal, finely granular, middle body, less strongly curved lateral border and a less prominent pair of swellings posterior to the middle body.

Bohemilla (Bohemilla) xesta sp. nov.

Figure 3a–e

Derivation of name. From the Greek, xesta, smoothed or polished, in reference to the relative smoothness of the glabellar prosopon compared to that of other species, including Bohemilla (Bohemilla) scotica, which also occurs in the Upper Whitehouse Subgroup at Girvan.

Holotype. GLAHM A6538a, b, moulds of a cranidium from the Red Mudstone Member of the Myoch Formation (uppermost Caradoc) of the Port Cardloch foreshore (locality P1).

Paratypes. All from the Red Mudstone Member of the Myoch Formation: partial cephalon, GLAHM A132434a, b (Whitehouse foreshore, locality W3); cranidium, GLAHM A9215 (Myoch foreshore, locality M4 a few centimetres below the Trilobite Bed at the base of the Wall Member); cranidium GLAHM A6189a, b (Whitehouse foreshore, locality W8); partial hypostome, GLAHM A132599 (Whitehouse foreshore, locality W3).

Other material. All from the Red Mudstone Member of the Myoch Formation: a cranidium GLAHM A22122 (Whitehouse foreshore, locality W9); a crandium, GLAHM A10426, and two fragmentary hypostomes, GLAHM A7477 & GLAHM A132438a, b (Myoch foreshore, locality M4, the Trilobite Bed at the base of the Wall Member).

Diagnosis. A species of Bohemilla (Bohemilla) with a smooth glabellar surface, at least five small tubercles developed on each side of the medial tubercle in front of the markedly transverse posterior margin of the occipital ring, very prominent tubercles on L2 and L3, a distinct but subdued lenticular area on the adaxial part of the anterior of L1, S1–3 extending across the glabella, lateral margin of frontal lobe slightly sigmoidal leading to an anterior tongue-like part, between and slightly in front of the anterior ends of the narrow palpebral lobes.

Description. All known cranidia are less than a centimetre long (sag.) and are somewhat crushed. The anterior of the paratype cephalon (GLAHM A132434b, Fig. 3b) has been ‘pushed in’, giving a much more blunt-ended outline than is seen in other specimens and so the most anterior part is not preserved. Glabella occupies a little over 95 % of the length of the cranidium and has a maximum width (at L2) equal to a little under two-thirds of its sagittal length. Glabella essentially flat mesially, curving steeply downwards along its lateral parts and along the anterior part of the frontal lobe. Glabella expands very gently forwards to the mid part of L2 before narrowing abruptly in front of S2 and then tapering gently forwards to a level at about two-thirds of the length of the frontal lobe, in front of which it tapers a little more strongly and then much more gently giving its anteriormost part a tongue-like outline with an almost transversely directed anterior edge (Fig. 3a). Occipital ring occupies almost a quarter of glabellar length. Posterior edge transversely directed except distally where it curves very sharply forwards through 90 ^o. Ring widest (tr.) at mid-length, bears a narrow (tr.) median ridge, the posterior part of which is developed as a posteriorly directed tubercle abaxial to which is a row of at least five small tubercles (Fig. 3a, d). A very shallow furrow extends obliquely abaxially rearwards across the outer third of the occipital ring. Occipital furrow curves gently forwards mesially, transversely directed or very slightly sinuous laterally. Preoccipital glabella bears a very weakly developed mesial furrow. L2 and L3 bear large, ridge-like, tubercles adjacent to this furrow (Fig. 3a); the posterior part of the frontal lobe on one specimen (Fig. 3c) bears a pair of very weakly swollen tubercles in this position. S1–3 essentially transversely directed and extend across the glabella. L1–3 together occupy about two-fifths of the sagittal length of the glabella. L1 bears a weakly differentiated transverse lenticular region of low convexity on the anterior edge of its mid part, delimited by little more than a break in slope posteriorly and the well-incised S1 anteriorly. S1 shallower than this mesially but well developed there none the less. L2 broadest (sag., exsag.) mesially, narrowing in front of the lenticular region of L1 before broadening slightly beyond this. S2 well developed mesially, deepening abaxially. L3 of largely constant width (sag., exsag.) except at its round-ended distal extremity. S3 transversely directed, very weakly impressed mesially. Frontal lobe occupies about a third of the glabellar length; sagittal length about two-thirds of posterior width; anteriormost tongue-like part, between and slightly in front of the anterior ends of the palpebral lobes has a width about two-fifths that of the posterior part of the frontal lobe.

Postocular part of fixigena poorly known; developed as a narrow border that expands in width (tr.) slightly from adjacent to the occipital ring to alongside L3, very slightly divergent from the sagittal line and bounded adaxially by a deep furrow that broadens (tr.) forwards at least to a level abaxial to L1 (Fig. 3c, d). Gently arched palpebral lobes extend forward from adjacent to the anteriormost parts of L3 around the anterolateral parts of the frontal lobe (Fig. 3a). Anterior border very narrow (sag., exsag.), decreasing in convexity (exsag) and tapering out abaxially (Fig. 3a). Details of eye not known but slender librigena transversely directed from abaxial to the anterior part of the glabella before curving rearwards in a wide arc through almost 90 ^o to become nearly parallel to the sagittal line and extending to an unknown level posterior to that of the occipital ring (Fig. 3b). Librigena narrows (exsag.) abruptly proximally before tapering only very slightly along most of its length; bears a broad groove that shallows gently and disappears towards the distal end of the cheek.

External surface and internal mould of glabella essentially smooth; a dense, very fine granulation may be no more than a reflection of the surrounding sediment grains.

Hypostome (Fig. 3e) incompletely known. A deep, straight, furrow directed forwards at about 55 ^o bounds the posterolateral part of the middle body; shallowing markedly posteriorly and curving round the transverse posterior edge of the middle body. Area between the middle body and the shallow, transversely directed border furrow gently swollen (sag., exsag.). Lateral border narrow (tr), gently curved concave outwards, poorly differentiated from the area posterior to the middle body. Posterior border narrow (sag., exsag.) over most of its length but developed distally as a very broad-based, posteriorly directed spine. Middle body and the swollen area between it and the posterior border densely covered in large tubercles and there is a rather diffuse, large swelling abaxial to the posteriormost part of the middle body. A few, more subdued, tubercles are present on the lateral border.

Remainder of exoskeleton not known.

Remarks. Although the tubercles on L2 and L3 are prominent in B. (B.) xesta sp. nov., they are less so than in B. (B.) scotica and while S1–3 extend across the glabella in both species, S3 is much less well impressed mesially than in the slightly younger Girvan species. B. (B.) xesta also differs in its smooth surface, markedly transverse posterior edge of the occipital ring in front of which there is only a small median tubercle (cf. robust spine) and at least five tubercles (cf. three short spines) on each side in a strip that is not ridge-like; more subdued lenticular area on L1 and in the shape of the frontal lobe with its short tongue-like anterior part.

The row of tubercles along the posterior edge of the occipital ring also differs from the condition in B. (B.) stupenda, which has one tubercle on each side of the median one. The more prominent pairs of tubercles close to the sagittal line on L2 and L3, less prominent lenticular area on L1, and much more strongly tapering frontal glabellar lobe also serve to distinguish B. (B.) xesta from B. (B.) stupenda, B. (B.) pragensis Marek (Reference Marek1966, pp. 151–52, pl. 2, figs 10–12, text-figs 5, 6) and, to some extent, B. (B.) roxana Šnajdr (Reference Šnajdr1982, pp. 227–28, pl. 1, figs 6–8). The details of the occipital ring of Šnajdr’s species are not known and the tubercles on L2 and L3 are more boss-like than those of the Scottish species. As is noted above, one of the two cranidia of B. (B.) roxana illustrated by Šnajdr (Reference Šnajdr1982, pl. 1, fig. 6) has the frontal lobe strongly tapered. The frontal lobe of B. (B.) xesta is similar to this but is more evenly tapered overall and the S3 furrow is distally more transversely directed. The hypostome of B. (B.) xesta has a coarsely granular sculpture on the middle lobe similar to that of B. (B.) roxana but it has a narrower lateral border and has a much less prominent tubercle posterior to the posterolateral part of the middle lobe.

B. (B.) xesta is similar to B. (B.) tridens Rushton & Hughes (Reference Rushton and Hughes1981, pp. 638–40, pl. 5, figs 10–16) in having a tongue-like anterior part of the frontal lobe but this is shorter (sag., exsag.), more tapered and less blunt-ended than in the English species. The smooth (cf. granular) surface, more prominent mesial tubercles on L2 and L3, the continuation of S3 across the glabella, narrower (tr.) palpebral lobes and a much less prominent tubercle posterior to the posterolateral part of the middle body of the hypostome also serve to distinguish the Scottish species from B. (B.) tridens, as do the furrows on the occipital ring and the large number of tubercles on the posterior margin of the occipital ring compared to the three spines in B. (B.) tridens.

Stubblefield (Reference Stubblefield1939, p. 61) noted that Bohemilla occurs in what is now termed the Ballyvorgal Group of County Clare, Ireland. The deep-water shelly fauna in that group is considered to be from close to the Caradoc–Ashgill boundary (Harper & Parkes in Fortey et al. Reference Fortey, Harper, Ingham, Owen, Parkes, Rushton and Woodcock2000, p. 55) and thus of broadly comparable age to the Girvan species described herein. Stubblefield cited a glabella illustrated by Baily (Reference Foot, Kinahan and Baily1862 in Foot et al., pp. 10, 20, fig. 1b) who ‘provisionally’ identified it as ‘Olenus?’. Stubblefield’s re-identification was repeated by Whittard (Reference Whittard1952, p. 313) and others (see Weir Reference Weir1959, p. 381) and is supported here. In describing the trilobite fauna of the Ballyvorgal Group, Weir (Reference Weir1959, pp, 369, 381) noted that Baily’s specimen was no longer available for study and that he had not found any further examples of Bohemilla in his own samples from the group. Baily’s illustration shows the shape of the glabella and its lobes typical of Bohemilla along with paired spines or tubercles on L2 and L3. Paired spines/tubercles are also shown on the occipital ring and L1, which is not typical of Bohemilla and probably reflect the rather schematic nature of the drawing. The relatively narrow frontal lobe compared to the width of the occipital ring on Baily’s drawing suggests that the specimen belongs in Bohemilla (Bohemilla) rather than B. (Fenniops) but the drawing is insufficient for meaningful species comparisons to be made with the chronostratigraphically and palaeogeographically comparable Girvan species. A second, very different, specimen also illustrated by Baily (Reference Foot, Kinahan and Baily1862 in Foot et al., fig. 1a) as ‘Olenus?’ was, probably rightly, assigned to Corrugatagnostus by Weir (Reference Weir1959, p. 370).