Biostratigraphy and taxonomy of Drumian (middle Cambrian) agnostid trilobites of the Manuels River Formation, Avalonian Newfoundland, Canada

New, accurately located and well‐preserved agnostid trilobite material has been collected from the type locality of the Drumian (middle Cambrian, Miaolingian) Manuels River Formation, Newfoundland, Canada. The well‐exposed grey to black shales containing the fauna were deposited on the former microcontinent Avalonia. Four interval zones for the succession are proposed, namely, the Tomagnostus fissus, Hypagnostus parvifrons, Ptychagnostus atavus and Ptychagnostus punctuosus zones. The Tomagnostus fissus and Ptychagnostus atavus zones are distinct from each other, which leads to a significantly shorter and stratigraphically higher Ptychagnostus atavus Zone compared with other sections (Scandinavia, South China, Utah and Nevada, Greenland). Specimens of Ptychagnostus atavus, however, occur abundantly, with 107 specimens identified in a short interval of the succession. The Ptychagnostus punctuosus Zone can be correlated globally. The faunal assemblage is comparable to that of Scandinavia, Greenland and England. Ptychagnostus affinis is recorded for the first time from Avalonia. The 1408 collected specimens were assigned to the suborders Agnostina and Eodiscina and to the families Peronopsidae and Ptychagnostidae, and Condylopygidae and Eodiscidae, respectively, with the following species: Peronopsis fallax (Linnarsson), Peronopsis scutalis (Hicks), Hypagnostus parvifrons (Linnarsson), Ptychagnostus punctuosus (Angelin), Ptychagnostus affinis (Brøgger), Ptychagnostus atavus (Tullberg), Tomagnostus fissus (Lundgren), Tomagnostus perrugatus (Grönwall), Pleuroctenium granulatum (Barrande) and Eodiscus punctatus (Salter).

T R I L O B I T E S of the Order Agnostida are common in Cambrian rocks and are important biostratigraphic tools, given their rapid evolution and wide distribution. They are generally abundant in open-shelf facies from the middle Wuliuan Stage to the end of the Drumian Stage (Geyer & Shergold 2000;Peng & Robison 2000). Because of their abundance and widespread distribution, they also provide data for palaeobiogeographic reconstructions (Høyberget & Bruton 2008;Weidner & Nielsen 2014).
The material for this study was collected from the type locality of the Drumian (middle Cambrian, Miaolingian) Manuels River Formation at Manuels River, Conception Bay South, Avalon Peninsula, Newfoundland (Fig. 1). The formation is also exposed at Trinity Bay, Placentia Bay and the western part of St Mary's Bay, and on the southern Burin Peninsula, Random Island, the French islands of St Pierre and Miquelon, and Cape Breton Island (Nova Scotia). Detailed palaeontological studies of the formation at Manuels River were carried out by Howell (1925) and Hutchinson (1962). Howell (1925) studied the succession in detail, and Hutchinson (1962) provided a lithostratigraphic overview of the Avalon region and systematic descriptions of the trilobite fauna. Subsequent studies on trilobites from the Manuels River Formation have largely focused on polymerids of the orders Ptychopariida and Redlichiida (e.g. Poulsen & Anderson 1975;Bergstr€ om & Levi-Setti 1978). Other palaeontological studies deal with microfossils, such as acritarchs (Martin & Dean 1988).
In order to obtain material for a revision and refinement of the systematics and biostratigraphy of agnostid trilobites, specimens were collected bed by bed from the highly fossiliferous shales exposed at the type locality of the formation.

GEOLOGICAL SETTING
In the early Palaeozoic, eastern Newfoundland was part of the microcontinent Avalonia, which is considered to be the largest terrane of the peri-Gondwanan realm (Pollock et al. 2012). In the middle Cambrian, Avalonia was connected to Baltica and Gondwana and was located in southern latitudes (Torsvik & Cocks 2017). The general geology of the Avalon Peninsula of Newfoundland has been described and discussed by several authors, for example Br€ uckner (1978), Anderson (1987), Poulsen & Anderson (1975), King (1990), Boyce (2001) and Fletcher (2006). The type locality of the Manuels River Formation is situated in the valley of the Manuels River, Conception Bay South, Newfoundland ( Fig. 1), at 47.525214, À52.951254 (WGS 84). The base of the formation is exposed c. 4 m above mean sealevel. The valley, including the outcrop and its fossils, is under government protection by Newfoundland and Labrador law, Regulation 67/ 11 of the Palaeontological Resource Regulations under the Historic Resources Act.
The Cambrian sedimentary succession exposed along Manuels River (Fig. 1) rests on Neoproterozoic magmatic rocks (e.g. rhyolites, granites and dacites) of the Holyrood Horst ( King 1988). The Holyrood Horst represents the basement of the overlying sedimentary rocks and is bordered by two main faults (Topsail and Brigus faults) on the western and eastern margins (Rose 1952;King 1988King , 1990 (Fig. 1). On the entire Avalon Peninsula, the Cambrian successions consist of shallow to deep marine sedimentary rocks with only minor faults (Hutchinson 1952(Hutchinson , 1962Fletcher 1972Fletcher , 2006Landing 2004).
At Manuels River, the lowermost Cambrian rocks belong to the Brigus Formation (Anderson 1987; Hutchinson 1962), which consists of a basal conglomerate overlain by mudstones with thin intercalations of limestone (Anderson 1987;Landing & Westrop 1998). The top of the formation is truncated by an erosional surface. The base of the disconformably overlying Miaolingian Chamberlain's Brook Formation consists of a manganese bed overlain by greenish mudstones. The boundary between the Chamberlain's Brook Formation and the conformably overlying Manuels River Formation is marked by a thin volcanic ash layer. Howell (1925) established the Chamberlin's Brook Formation (his beds 1-35), Long Pond Formation (his beds 36-92), and Kelligrew Brook Formation (his beds 93-125) for the strata along Manuels River. Subsequently, Hutchinson (1962) renamed and defined the Chamberlin's Brook Formation as Chamberlain's Brook Formation and merged Howell's (1925) Long Pond and Kelligrew Brook formations (beds 36-125) into his Manuels River Formation, composed mainly of black shales with interbedded thin volcanic ash layers and carbonate concretions (Austermann 2016). The contact with the overlying, not yet formally defined Elliot Cove formation, is unconformable and is marked by a coarse sandstone/conglomerate (Howell 1925, bed 125;Hutchinson 1962;Poulsen & Anderson 1975;Austermann 2016).

MATERIAL AND METHOD
A total of 1408 specimens have been collected from the type locality. All specimens are stored at The Rooms Corporation of Newfoundland and Labrador, Provincial Museum Division, Natural History Unit, St John's, Newfoundland, Canada. Currently, the agnostid collection of Manuels River is on loan to Heidelberg University, Heidelberg, Germany. The herein studied collection has been supplemented with type specimens from the collections of Hutchinson (1962) and Martin & Dean (1988), with their collections housed in the Geological Survey of Canada, Ottawa, Canada. Agnostid classification and terminology follow that of Shergold et al. (1990), Whittington et al. (1997), Peng & Robison (2000) and Robison (1982). Matthews (1973), Bengtson (1988) and Becker (2001) are followed for abbreviations and symbols used for detailed nomenclatural decisions and statements. The specimens were photographed and measured with a Keyence VHX-6000 digital microscope with integrated LED light from the upper left. The illustrated photographs were calibrated with Adobe Illustrator CC.

Local biostratigraphy
The first biostratigraphic zonation of the Manuels River Formation was established by Howell (1925). Before Howell's work, authors focused on the subdivision of the entire Conception Bay area, and the North American continent in general, rather than on detailed biostratigraphy (Murray 1869;Murray & Howley 1881;Matthew 1886Matthew , 1896Matthew , 1899Walcott 1888aWalcott , b, 1889Walcott , 1891Walcott , 1900Marcou 1890). Howell (1925) described the black shale successions along the Manuels River in detail and established three trilobite-based biozones, the Paradoxides bennetti Zone (Howell's beds 1-35), the Paradoxides hicksi Zone (beds 36-92), and the Paradoxides davidis Zone (beds 93-125), the latter two zones corresponding to the present Manuels River Formation. The distribution of agnostid trilobites and species of the orders Ptychopariida and Redlichiida were documented. Howell's (1925) fossil collections were deposited in different museums, the last identifiable location being the Smithsonian National Museum of Natural History, Washington DC, USA, where some of the type specimens are housed. The major part of Howell's material has not been found there or anywhere else and must be considered lost. Hutchinson (1962) confirmed the local biostratigraphic zones of Howell's (1925) beds 36-125 and introduced the name Manuels River Formation for this succession. He revised and supplemented the work of Howell, especially the systematics of the agnostid trilobites. His detailed work includes numerous descriptions and illustrations. Poulsen & Anderson (1975) subdivided the succession corresponding to the upper Manuels River Formation and the lower Elliot Cove formation biostratigraphically using Scandinavian trilobite zones, which are based on trilobites of different orders, including agnostids. They identified Ptychagnostus punctuosus (Angelin, 1851) in the Ptychagnostus punctuosus Zone from the Highland Cove assemblage, south-eastern Newfoundland (corresponding to the upper part of the Paradoxides davidis Zone of Hutchinson 1962). For global correlation of this upper middle Cambrian zone, see the section 'Ptychagnostus punctuosus '. Bergstr€ om & Levi-Setti (1978) studied the upper part of the Manuels River Formation bed by bed on the north-eastern banks of the river. They described two subspecies of Paradoxides davidis Salter, 1863 (Pa. davidis trapezopyge and Pa. davidis intermedius) from different intervals of the formation, but no agnostids were reported. The authors followed the concept of the Pa. davidis Zone of Howell (1925) and Hutchinson (1962). Martin & Dean (1981 studied for the first time acritarchs and trilobites and integrated the biozonations of these groups for the Manuels River succession. They redefined the existing biozones and established three trilobite biozones, namely, the Tomagnostus fissus-Ptychagnostus atavus Zone, the Hypagnostus parvifrons Zone (although with uncertain lower and upper boundaries) and the Ptychagnostus punctuosus Zone. The conglomerate of Howell's (1925) bed 125 was taken as the basal bed of the Elliot Cove formation, a concept followed here.

Results and correlation
The mostly cosmopolitan distribution of agnostid trilobites in open-marine sediments, their relatively short stratigraphic ranges and abundant occurrence make them appropriate index fossils (Robison 1984(Robison , 1994Peng & Robison 2000). The interval zone concept is a common tool for chronostratigraphic correlation, in which each agnostid biozone is defined as the interval between the lowest stratigraphic occurrence ('first occurrence', FO) of its eponymous agnostid species and the lowest occurrence of the next selected agnostid species (e.g. Robison 1984;Peng & Robison 2000;Høyberget & Bruton 2008;Weidner & Nielsen 2014). The defining species are selected on the basis of their cosmopolitan distribution, abundance, and their relatively short stratigraphic range (Robison 1984;Peng & Robison 2000). However, the use of FOs for global chronocorrelation suffers from the inherent diachroneity of the species, the magnitude of which varies between species. Precise correlation is possible only in combination with studies of sedimentology and calibration with other zonations (Landing et al. 2013). Several studies (e.g . Howell 1925;Westerg ard 1946;Westrop et al. 1996) based on agnostid and/or polymerid trilobites lack precise definitions of zonal boundaries or used different, non-reproducible descriptive methods, for dividing biozones, as also remarked by Peng & Robison (2000) and Høyberget & Bruton (2008). Illing (1915) did not define his biozones but described Paradoxides hicksi from his Paradoxides aurora Zone, which is stratigraphically lower than the Pa. hicksi Zone. Paradoxides hicksi is rare in both zones and becomes abundant only in the upper part of the Pa. hicksi Zone and does not occur at the base of its Zone (divided into a lower and an upper part), but in the uppermost part of the Pa. davidis Zone. Given that the biozones are not reproducible, the given range chart allows only a rough correlation. Rushton (1979) adopted Illing's (1915) biozones with minor supplements, although without precise definition. Howell (1925) and Hutchinson (1952Hutchinson ( , 1962 described their biozones in combination with lithostratigraphy. Given that Howell (1925) described the strata in more detail, Hutchinson (1962) often used Howell's work as a base for his biozones, although without definitions of the zones. The range charts of the described taxa by Howell (1925) and Hutchinson (1962) allow for correlation. According to Westerg ard (1946), Hypagnostus parvifrons occurs in his lower T. fissus and Pt. atavus Zone (Berg-Madsen 1984;Peng & Robison 2000). Here, we follow the concept of Peng & Robison (2000) to combine Westerg ard's (1946) Tomagnostus fissus-Ptychagnostus atavus and Hypagnostus parvifrons zones. As Illing (1915), Hutchinson (1952Hutchinson ( , 1962, Fletcher (1972,2006), Rushton (1979), andLanding &Westrop (1998) used polymerid trilobites for their biozones, the integration of Howell's (1925) high-resolution range chart and the range chart of the present study allows for correlation of our biozonation with those of these authors.
Only a few studies provide clear and exhaustive definitions of local biozones (Fletcher 1972(Fletcher , 2006Robison 1984Robison , 1994Peng & Robison 2000;Weidner & Nielsen 2014). Here we revise and complement the faunal assemblage and the local biozones of previous studies (e. g. Howell 1925;Hutchinson 1962;Martin & Dean 1988) and correlate the biozones of the Manuels River Formation at Conception Bay globally.
The four local interval zones proposed here are primarily correlated with East and West Avalonia (Newfoundland, Nova Scotia, England), supplemented with Baltica (Scandinavia), Siberia, Laurentia (Utah, Nevada, Greenland) and South China (Figs 2, 3). In the work by Howell (1925) on the type locality of the Manuels River Formation, the Paradoxides hicksi and Pa. davidis biozones were defined on the basis of the trilobites of the Order Redlichiida, whereas biozones proposed here are based on the Agnostida. We propose the following local biozones for the type locality of the Manuels River Formation.
F I G . 2 . Stratigraphic distribution and biozonation of agnostid species of the middle Cambrian Manuels River Formation at its type locality along Manuels River, Newfoundland.
Tomagnostus fissus Zone. The base of the zone is defined by the FO of T. fissus at 3.89 m above the base of the Manuels River Formation (Fig. 2). The species ranges up to 7.70 m in section. Co-occurring species at the base of this zone are Peronopsis scutalis, Eodiscus punctatus, Pleuroctenium granulatum and Tomagnostus perrugatus. The top of the zone is defined by the FO of Hypagnostus parvifrons at the level of 10.96 m.
Hypagnostus parvifrons Zone. The base of the zone is defined by the FO of H. parvifrons at 10.96 m above the base of the formation (Fig. 2). Hypagnostus parvifrons F I G . 3 . Global correlation of the middle Cambrian of Newfoundland, Avalonia. Data sources: Deep Cove: Fletcher 1972Fletcher , 2006Highland Cove: Hutchinson 1962;Landing & Westrop 1998;Nova Scotia: Hutchinson 1952;England: Illing 1915;Rushton 1979;Scandinavia: left column, Westerg ard 1946; right column, (1) Høyberget & Bruton 2008, (2) Weidner & Nielsen 2014Siberia: Egorova et al. 1982;Pegel 2000;South China: Peng & Robison 2000;Utah, Nevada, USA andGreenland: Robison 1984, 1994. ranges well into the overlying Pt. atavus Zone. The associated assemblage consists of Peronopsis fallax, T. perrugatus and Pl. granulatum. Eodiscus punctatus has not been found in the zone, although it occurs in the underlying and overlying zones. The top of the zone is defined by the FO of Ptychagnostus atavus at 11.90 m. Ptychagnostus atavus Zone. The base of the zone is defined by the FO of Pt. atavus at 11.90 m above the base of the formation (Fig. 2). The species occurs abundantly up to 12.08 m. Additional species at the base of the zone are T. perrugatus, Pe. fallax, Pl. granulatum

Tomagnostus fissus Zone
The T. fissus Zone is here established as a local biozone. Howell (1925) proposed the Paradoxides hicksi Zone for his beds 36-92, and Hutchinson (1962) followed this suggestion. Nevertheless, our T. fissus Zone can be correlated with Howell's Pa. hicksi Zone thanks to the precise work of Howell (1925), despite the lack of systematic work on and illustrations of the species. Peronopsis scutalis and Tomagnostus perrugatus are missing from the faunal assemblage of Howell (1925) . Howell reported and Hutchinson (1962) described Hypagnostus parvifrons from the Pa. hicksi Zone, a species we have found only at a stratigraphically higher level. Hutchinson (1962) described T. fissus as the most common agnostid in the Pa. hicksi Zone of the Manuels River section. Ptychagnostus atavus was described by Hutchinson (1962) from the upper part of the Pa. hicksi Zone and the basal Paradoxides davidis Zone. Martin & Dean (1988) combined the agnostid occurrences to a T. fissus-Pt. atavus Zone, approximately corresponding to the Pa. hicksi Zone of Howell (1925). However, they did not describe or report T. fissus or Pt. atavus from the Manuels River section.
Interregional correlation is possible with the T. fissus-Pt. atavus Zone of Baltica (Westerg ard 1946) and T. fissus Zone of Siberia (Egorova et al. 1982). Tomagnostus fissus is common in the middle Cambrian successions of Avalonia, Baltica and eastern Laurentia. The species has not been reported from Gondwana, which makes the establishment of a globally working biozone difficult. Howell (1925) proposed the Pa. davidis Zone for his beds 93-125, which was followed by Hutchinson (1962). Based on a comparison with contemporaneous Scandinavian strata, Martin & Dean (1988) proposed the interval as belonging to the H. parvifrons Zone, although with uncertain lower boundary. The dashed line of the boundary is placed in the interval of beds 94-99 of Howell (1925) (10.65-11.30 m), which is close to the FO of H. parvifrons at the level of 10.96 m.

Hypagnostus parvifrons Zone
The zone correlates with the H. parvifrons Zone of Sweden, as defined by Westerg ard (1946). The associated agnostid assemblage of Peronopsis fallax, T. perrugatus and Peuroctenium granulatum is similar to that in Sweden. Westerg ard (1946) also described Pt. affinis from the H. parvifrons Zone, a species that in the Manuels River succession appears only in the overlying Pt. atavus Zone.
Although H. parvifrons is a widely distributed species, its stratigraphic range varies significantly between regions. Also, the FOs of H. parvifrons and Pt. atavus occur at close stratigraphic levels, for example in China, Siberia and USA (Peng & Robison 2000). Therefore, the H. parvifrons Zone described here should be taken as a local biozone.

Ptychagnostus atavus Zone
The Pt. atavus Zone is part of the Pa. davidis Zone proposed by Howell (1925) for his beds 93-125, and later adopted by Hutchinson (1962). For beds 99-114, Howell (1925) described the same faunal assemblage as here except for the species Pt. atavus and Pt. affinis. Detailed sampling for the present study has shown that Pt. atavus occurs only in the narrow interval of 11.90-12.08 m, here with 107 specimens. Hutchinson (1962) and Fletcher (2006) reported Pt. atavus from the upper Hydrocephalus hicksi Zone to the lower Pa. davidis Zone from Highland Cove and Cape St Mary's, respectively. Westrop et al. (1996) reported Pt. atavus from western (Laurentian) Newfoundland. It is remarkable that Pt. atavus has not been reported from other areas of Avalonia, for example England, Wales, New Brunswick or Nova Scotia (Matthew 1886;Illing 1915;Hutchinson 1952Hutchinson , 1962Fletcher 2006;Rees et al. 2014). Ptychagnostus affinis is reported for the first time from the succession at the Manuels River, and hence also for the entire microcontinent Avalonia.
Ptychagnostus atavus and Pt. affinis occur in a very short stratigraphic interval at the type locality, shorter than in other areas in eastern Newfoundland and than in most of the other global occurrences. However, Fletcher (1972,2006) has shown that the Manuels River Formation is condensed at its type locality compared with other areas in eastern Newfoundland.
The GSSP for the Drumian Stage, Cambrian Series 3, was defined in the Wheeler Formation, Drum Mountains, Utah, USA (Babcock et al. 2007). The level was selected at the FO of Pt. atavus in the section on the basis of its cosmopolitan occurrence compared with other middle Cambrian agnostid trilobites, such as H. parvifrons or T. fissus (Geyer & Shergold 2000;Peng & Robison 2000;Babcock et al. 2007).

Ptychagnostus punctuosus Zone
Ptychagnostus punctuosus is widespread globally and its zone is easy to correlate. Howell (1925)  Remarks. Peronopsis includes the geologically oldest agnostids of the middle Cambrian (Laurie 1990; Naimark 2012). More than 100 species have been assigned to this genus (Rushton 1979;Robison 1994Robison , 1995Weidner & Nielsen 2014). The morphological characters vary during ontogeny and within population, which led to a complex taxonomic subdivision of Peronopsis (Robison 1982;Naimark 2012). This is why there are frequent discussions about synonyms. Acadagnostus Kobayashi, 1939, is a frequently discussed synonym (e.g. Rushton 1979;Laurie 1990;Robison 1994Robison , 1995Shergold & Laurie 1997). Originally the genus was described by Kobayashi (1939) as having a pygidial lanceolate axis and a pygidial median postaxial furrow and lacking a pair of pygidial posterolateral spines. In contrast, Peronopsis is very variable in the pygidial median postaxial furrow as well as in the occurrence of spines. Shergold & Laurie (1997) presented a diagnosis of Acadagnostus in which the pygidial spines and the pygidial axis never reach the border furrow.
In some smaller cephala the border furrow is narrower. The cephalic axial node is small but visible. There is a larger variation in the pygidia of the specimens due to the axis varying in broadness and in the form of the posterior end. Some specimens show a tapered end reaching the border furrow, or the axis is long without a contact to the furrow. A median postaxial furrow is visible in some specimens. Some pygidia show a weakly transverse depression at the midlength of the axis. All pygidia show a pair of posterolateral spines.
Remarks. Peronopsis fallax is a long-ranging species with a high variability in morphology (Robison 1982;Høyberget & Bruton 2008). The cephala are subquadrate to subcircular. The relative width of the cephala and pygidia increases during ontogeny. This is, for example, visible in characters such as border furrows, which become wider from meraspis to holaspis stage (Robison 1995). The length of the pygidial axis varies, from ending in the postaxial area near the border furrow, to contacting the median postaxial furrow, or reaching the border furrow. Høyberget & Bruton (2008) explained this variation in association with the stratigraphic occurrence of the species and with intraspecies variations. In stratigraphically older specimens, from the early middle Cambrian, the pygidial axis reaches the border furrow, while in later occurring specimens (Pt. atavus Zone) the pygidial axis contacts the median postaxial furrow. In addition, the median postaxial furrow weakens or vanishes during advanced ontogeny (Robison 1982). Due to these variations the systematic position of Pe. fallax is still controversial (e.g. Hutchinson 1952;Gil Cid 1981;Laurie 1990; Robison 1994Robison , 1995Høyberget & Bruton 2008;Weidner & Nielsen 2009. Robison (1995) and Weidner & Nielsen (2014) considered Pe. fallax as a junior subjective synonym of Agnostus acadicus Hartt in Dawson, 1868). Robison (1995) noted that the type material must be viewed critically, because Hartt (in Dawson 1868) was uncertain about the true association of the cephalon and pygidium. € Opik (1979) and Robison (1995) suggested that the questionable type pygidium of Hartt be assigned to H. parvifrons (Linnarsson, 1869) (see the H. parvifrons section). The illustrated specimen in Dawson (1868) shows a spineless pygidium, which is not characteristic for Pe. fallax. In addition, Matthew (1896) and Robison (1995) collected both at the same horizon in the St John's region and found many specimens of Ag. acadicus with pygidial spines instead of non-spinose pygidia. The cephalon sampled by Hartt (in Dawson 1868) is in Robison's (1995) view still the holotype of Ag. acadicus and therefore a senior subjective synonym of Pe. fallax. It is now the type species of Acadagnostus Kobayashi, 1939, as described in Shergold & Laurie (1997) (see the Peronopsis section). The present study follows the suggestion of Høyberget & Bruton (2008), that Hartt's species should be regarded as a nomen dubium because the type material is to be seen as doubtful and is not reviewable.
Peronopsis fallax is distinguished from the closely related species Peronopsis ferox (Tullberg, 1880) mainly by characteristics of the pygidia. Peronopsis ferox has a broader and shorter pygidial axis than Pe. fallax, thus the postaxial area is wider and a median postaxial furrow is absent. In addition, Pe. ferox often has a crescent-shaped border between the pair of pygidial posterolateral spines, whereas Pe. fallax developed a narrow border of constant width between the spines (Høyberget & Bruton 2008). Peronopsis scutalis (Hicks, 1872) differs from Pe. fallax by having narrower cephalic and pygidial border furrows. The pygidium shows more distinguishing characteristics such as a subcircular to semiovate form. The axis is lanceolate, with a smaller axial node and a secondary median node on the midpoint of the posteroaxis. A median postaxial furrow is present and the pygidium is spineless.
The figured pygidium (Brøgger 1879, fig. 1a) has a third spine at the pygidial margin, in the middle of the pair of posterolateral spines, and is here excluded from Pe. fallax. Linnarsson (1879) figured a cephalon without the characteristic wide border furrow, making an assignment to Pe. fallax doubtful. Tullberg (1880) divided Pe. fallax in two subspecies: Agnostus fallax typica and Ag. fallax ferox. The illustrated cephalon for the subspecies typica agrees well with Pe. Fallax, but the axis of the pygidia is ogival and does not show the characteristic broad form, making their assignment questionable, although an intraspecies variation is possible. Matthew (1886) illustrated Ag. acadicus. The cephalon matches that of Pe. fallax, but the pygidium in fig. 5b of Matthew (1886) is here assigned to H. parvifrons. As discussed above, the illustrated spineless pygidium also has the characteristic lanceolate axis and the distinct forward projection of the posterior border. Matthew (1886) described the new species Agnostus vir and the subspecies Ag. vir concinnus. Both illustrations show the cephalic and pygidial wide border furrows, and broad pygidial axis with a large axial node and a pair of posterolateral spines. Matthew (1886) distinguished Ag. vir from Pe. fallax by the form of the anterior glabellar lobe and the trisection segmentation of the posterior thorax segment. Agnostus vir concinnus has even more narrow lateral furrows of the glabella. These described characters by Matthew (1886) are here considered to be intraspecies variations of Pe. fallax. Because of the above described variability within Pe. fallax it is here suggested to avoid the usage of subspecies, as also discussed by Robison (1982).
The pygidium of Matthew (1896) shows a wide border furrow, but the pygidial axis is very broad and too short, thus it is here assigned to Pe. ferox. Matthew (1896) illustrated Ag. fallax concinnus with the characteristic wide border furrow of the pygidium, but the cephalon has only a narrow border furrow. The cephalon may represent a meraspis stage, making an assignment to Pe. fallax questionable.
Lorenz (1906) described the new subspecies Ag. fallax laiwuensis, with the wide border furrows on cephalon and pygidium, a broad pygidial axis with a large axial node that reaches the border furrow, and a pair of posterolateral spines. The subspecies is distinguished from Pe. fallax by the faint furrows at the posterior end of the posteroaxis. The character described by Lorenz (1906) is here suggested to be an intraspecies variation of Pe. fallax.
Grabau & Shimer (1910) illustrated a cephalon and pygidium of Ag. acadicus. The cephalon is here assigned to Pe. fallax because of the typical wide border furrow and the subquadrate, subcircular shape (cf. Robison 1995). Howell (1925) described the new species Agnostus clarae given its pygidia with a short axis, and mentioned a connection to Pe. fallax and Pe. ferox. The illustrated pygidium is subquadrate, and has a wide border furrow, a long axis and a pair of posterolateral spines. Because of the typical characteristics Ag. clarae is here assigned to Pe. fallax.
In his figure 19, Strand (1929) named the illustrated taxa Agnostus fallax and Agnostus Ag. parvifrons mammillatus. However, only cephala and pygidia of H. parvifrons are illustrated, with an effaced anterior glabellar lobe that is half as long as the cephalon, and non-spinose pygidia that show the distinct forward projection of the posterior border (see the Hypagnostus parvifrons section). It is here suggested that Strand (1929) illustrated the wrong species, therefore it is not assigned to Pe. fallax.
The illustrated specimens of Hutchinson (1952) and Henningsmoen (1952) are poorly preserved, and no characteristics are visible, therefore any assignment is questionable. Dean (1982) illustrated four cephala and two pygidia of Pe. fallax minor, which all have the typical characteristics of Pe. fallax such as a wide border furrow, an elongate pygidial axial node and a pair of posterolateral spines. Egorova et al. (1982) illustrated several cephala and pygidia of Pe. fallax. A cephalon and two pygidia are poorly preserved, so any assignment is here suggested to be questionable (cf. Robison 1995). Samson et al. (1990) assigned a cephalon and a pygidium to Pe. fallax, but their fig. 5B is here excluded, because of the absence of the pair of pygidial posterolateral spines. Weidner & Nielsen (2009) illustrated a poorly preserved cephalon and pygidium, and their assignment is here considered to be doubtful.  Robison 1982;Samson et al. 1990), Norway (Høyberget & Bruton 2008;Maletz & Steiner 2015) and Denmark (Weidner & Nielsen 2014). Peronopsis fallax has also been reported from the upper part of the Pt. atavus Zone of New Brunswick and Nova Scotia (Hutchinson 1952), Western Newfoundland (Young & Ludvigsen 1989), Australia ( € Opik 1979), England (Rushton 1979), Spain (Gil Cid 1981), Russia (Siberia) (Egorova et al. 1982), Eastern Turkey (Dean 1982), Antarctica (Palmer & Gatehouse 1972), Greenland (Robison 1994) and Germany, in erratic boulders (Rudolph 1994). In the present study Pe. fallax ranges from the T. fissus Zone to the Pt. atavus Zone of Eastern Newfoundland.
Description. The specimens are mainly well-preserved. The cephala vary in size from 1.3 to 3.4 mm in width and from 1.2 to 3.4 mm in length, and the pygidia vary in size from 2.5 to 4.6 mm in width and from 2.4 to 4.8 mm in length. At the 2.55 m level two of the collected pygidia show a yellow surface from the pyrite contained in the shale. The cephala and pygidia are characteristically subcircular to subquadrate in shape. Anteriorly the specimens have a broadly rounded anterior glabellar lobe and posteriorly a straight F3 furrow. The pygidial border is narrow in some specimens and wider in others. In some specimens the small, secondary median node is visible, situated at the midpoint of the lanceolate posteroaxis. The pygidial pleural fields are rounded anteriorly and separated by a deep, median, postaxial furrow. The length of the postaxial furrow varies in some specimens.
Remarks. Peronopsis scutalis shows morphological variation within populations (Robison 1994;Høyberget & Bruton 2008;Weidner & Nielsen 2014). The cephalon changes less than the pygidia throughout ontogeny and within populations. The cephala have a small axial node on the posterior glabellar lobe. Due to preservation differences, the node is not visible in some specimens. Westerg ard (1946) described cephala and pygidia from Sweden with a length of 2-3 mm. In contrast, 5-6-mmlong cephala and pygidia from Greenland are common (Robison 1994). In the present study, the length of the cephala and pygidia matches well with the Swedish specimens described by Westerg ard (1946).
There is a large variation in length and width in the pygidia (Westerg ard 1946;Weidner & Nielsen 2014). The length of the axis extends to two-thirds of the pygidial length with, in some specimens, a posteriorly tapered end. Usually, the pygidial pleural fields have nearly the same width as the axis and the ends are rounded to straight, the axis becomes longer, and axial and border furrows become deeper throughout ontogeny and in larger specimens (Robison 1994). The pleural fields are separated by a deep, median postaxial furrow. The small secondary median node at the midpoint of the posteroaxis varies from weak to absent, which is probably a matter of preservation (Weidner & Nielsen 2014). Peronopsis scutalis is distinguished from other species of Peronopsis by the long anterior glabellar lobe, the small basal lobes, the very narrow border furrows and the absence of spines (Høyberget & Bruton 2008).
Hicks (1872) used the name Pe. scutalis Salter (1866) in his description, given that Salter (1866) first mentioned the species name in his report. Therefore, especially in older studies, Salter (1866) is listed in several synonymy lists (e.g. in Lake 1906). Salter (1866) mentioned the name Pe. scutalis without a description of the species, thus he compiled a nomen nudum. According to the ICZN the name and date with the first description of a species are valid, thus Hicks (1872) is listed here. Tullberg (1880) illustrated a pygidium of Hypagnostus parvifrons. Typical characteristics of Pe. scutalis are a pygidium with a long pygidial axis, tapered posteriorly, and F1 and F2 furrows absent.
Gr€ onwall (1902) described the species Agnostus exaratus, figuring a single pygidium with a broad and short pygidial axis, rounded pleural fields and a median postaxial furrow. The broader pygidial axis is here considered an intraspecies variation, and the other characters match those of Pe. scutalis. Therefore, the species Ag. exaratus is here considered to be a synonym of Pe. scutalis (cf. Høyberget & Bruton 2008). Illing (1915) described the subspecies Ag. exaratus tenuis, distinguished from Pe. scutalis by its larger size, the subquadrate shape of the anterior glabellar lobe and the narrower and more tapering pygidial axis. The illustrated specimens of Ag. exaratus tenuis have a relatively short pygidial axis, which agrees well with Pe. scutalis, and hence Ag. exaratus tenuis is here suggested to be a synonym of Pe. scutalis (cf. Høyberget & Bruton 2008).
The illustrated pygidia of Peronopsis scutalis by Westerg ard (1946, figs 9-11) show a broad axis, posteriorly rounded pleural fields and a forward projection of the broad pygidial border furrow, and are more typical of Hypagnostus parvifrons. Hutchinson (1962) followed the concept of the subgenus Pe. (Acadagnostus) scutalis; his illustrated specimens match with Pe. scutalis (cf. Høyberget & Bruton 2008). € Opik (1979) and later Rudolph (1994) and Weidner & Nielsen (2014) assigned scutalis to the genus Acadagnostus Kobayashi, 1939. As discussed under the genus Peronopsis (see the Peronopsis section) Pe. scutalis is non-spinose; thus it is here included into Peronopsis. Rushton (1979) divided Pe. scutalis into the subspecies Pe. scutalis scutalis and Pe. scutalis tenuis. He illustrated a poorly preserved pygidium of questionable Pe. scutalis scutalis, which seems to be compressed at the posterior part. His illustrated complete specimen of Pe. scutalis tenuis shows the typical cephalon with the big anterior glabellar lobe and the cephalic small axial node. The pygidium has a short but well-tapered end, characteristic of Pe. scutalis, and the subspecies of Rushton (1979) is here considered to be a synonym of Pe. scutalis. Egorova et al. (1982) illustrated two specimens as Pe. scutalis with a semiovate cephalon, an anterior small and subcircular glabellar lobe and a long posterior glabellar lobe. The pygidium shows a long and broad axis. Both illustrated specimens have broad border furrows and axial furrows, uncharacteristic of Pe. scutalis. Martin & Dean (1988) described the subspecies Pe. scutalis exarata with all specific characters of Pe. scutalis and so it is here included into this species (cf. Høyberget & Bruton 2008). Westrop et al. (1996) illustrated a cephalon and three poorly preserved pygidia with fragmentary axes. These authors remarked that their few specimens are not sufficient for a confident identification, a view followed here. Høyberget & Bruton (2008) erroneously listed ' Dean 1982' for Martin & Dean (1988 in their synonymy list. As discussed above, Pe. scutalis is a highly variable species, and the attempts by many authors to split this species into subspecies were common in the past and are here considered unjustified. Thus, the broad species concept of Westerg ard (1946), Rushton (1979), Robison (1994), Westrop et al. (1996), Høyberget & Bruton (2008) and Weidner & Nielsen (2014), is followed here.
Occurrence. Peronopsis scutalis is widespread and has been reported from the middle Cambrian Ptychagnostus gibbus Zone to the Lejopyge laevigata Zone (Robison 1994;Høyberget & Bruton 2008;Weidner & Nielsen 2014). It has been noted from the Pt. gibbus Zone of Australia (Northern Territory) and Germany, in erratic boulders ( € Opik 1979;Rudolph 1994), the lower to upper Pt. atavus Zone of Sweden and Greenland (Westerg ard 1946;Robison 1994;Rudolph 1994), the Pt. atavus Zone of Denmark and Germany, in erratic boulders (Rudolph 1994;Weidner & Nielsen 2014), the Pt. atavus Zone to the lower part of the Pt. punctuosus Zone of Denmark, England, Wales and Norway (Gr€ onwall 1902; Rushton 1979;Høyberget & Bruton 2008) and the Lejopyge laevigata Zone of Western Newfoundland (Westrop et al. 1996). In the present study Pe. scutalis occurs in the T. fissus Zone of Eastern Newfoundland.
Diagnosis. Cephalon with effaced anterior glabellar lobe; F3 furrow rounded or truncated; median preglabellar furrow absent; posterior glabellar lobe short with glabellar node; pygidial border broader than the cephalic border; pygidial axis long (based on Robison 1964;Westrop et al. 1996;Shergold & Laurie 1997, with modifications). Synonyms. Spinagnostus Howell, 1935;Cyclopagnostus Howell, 1937. Remarks. The most conspicuous characteristics of Hypagnostus are the effacement of the anterior glabellar lobe, the short posterior glabellar lobe and the absence of a median preglabellar furrow (Westerg ard 1946;Peng & Robison 2000). Intraspecies variations in large populations are specific to the pygidium and include characters such as faint to absent F1 and F2 furrows on the axis, various combinations of shape and length of the axis, a weak axial node, a median postaxial furrow and a pygidial pair of posterolateral spines (Peng & Robison 2000). Because of these variable characters, the systematic position and the assignment of synonyms of Hypagnostus are still controversial. Shergold et al. (1990) and Shergold & Laurie (1997) assigned Hypagnostus to the Subfamily Spinagnostinae Howell, 1935. Westrop et al. (1996 followed the concept of Robison (1994) to leave Hypagnostus in the Family Peronopsidae because of an uncertainty about the monophyly of Spinagnostinae. Both Spinagnostinae and Peronopsidae are primarily determined by the effacement of the anterior glabellar lobe, and this may be a polyphyletically developed characteristic. Westrop et al. (1996) considered Cotalagnostus to be a synonym of Hypagnostus. Peng & Robison (2000) consider Cotalagnostus a separate genus, a view followed here. Cotalagnostus is distinguished from Hypagnostus by an absent glabellar F3 furrow and faint to absent pygidial furrows surrounding the posteroaxis (Shergold & Laurie 1997). Tomagnostella was considered a synonym of Hypagnostus (e.g. Westerg ard 1946;Robison 1964). Westrop et al. (1996) left both as separate genera on the basis of the pygidial axis character of distinct F1 and F2 furrows, a view followed here.    fig. 5b (non fig. 5a). 1886 Agnostus acadicus, var. declivis, n. var. Matthew,, pl. 7, fig. 6b (non fig. 6a (pygidium), by subsequent designation of Westerg ard (1946, pl. 4, figs 27-28), refigured by Shergold & Laurie (1997, fig. 226.5a-b). The cephalon and pygidium were illustrated originally by Linnarsson (1869, pl. 2, figs 56-57).
Diagnosis. Short oval posterior glabellar lobe not exceeding half the length of the cephalon; cephalon and pygidium non-spinose; pygidial axis with effaced F1 and F2 furrows; pygidial axial node weak, located at anterior part of axis; pleural fields divided by median postaxial furrow; pygidial border showing forward projection (based on Robison 1964;Høyberget & Bruton 2008, with modifications).
Description. The specimens are poorly preserved. Two complete specimens are 4.9 mm and 6.6 mm long, respectively. The cephala vary in size from 1.0 to 5.1 mm in width and from 1.1 to 4.8 mm in length, and the pygidia vary in size from 1.5 to 2.9 mm in width and from 1.7 to 2.8 mm in length. At the level of 12.62 m, the specimens show a yellow surface, typical of the pyrite in the shales. All cephala show the characteristic oval posterior glabellar lobe. In some cases, the F3 furrow is truncated and vaguely defined. The small cephalic axial node is visible in several specimens; when absent, it is most likely a matter of preservation. In addition, some of the cephala show a weak scrobiculation on the cephalic genae. The cephalic border is narrow and the pygidial border broader. All the pygidia have a median postaxial furrow with posteriorly rounded pleural fields. The forward projection of the pygidial border is visible. The pygidial axis is in most of the specimens moderately broad and the axial node is always visible.
Remarks. Hypagnostus parvifrons shows several variable characters within populations (Westerg ard 1946;Robison 1964;Westrop et al. 1996;Peng & Robison 2000;Høyberget & Bruton 2008). The cephala are variable in the length of the posterior glabellar lobe, which does not exceed half the length of the cephalon. Usually, the cephalic genae are smooth, but in some cases a weak scrobiculation is developed (Westerg ard 1946;Peng & Robison 2000;Høyberget & Bruton 2008). The cephalic and pygidial width of the border is another intraspecies variation (Robison 1964;Westrop et al. 1996). The pygidial axis is long but variable in length and width (Peng & Robison 2000). The pleural fields are always longer than the pygidial axis and in most cases rounded posteriorly. In addition, the length of the axis affects also the length of the median postaxial furrow.
Specimens of an ontogenetic series were illustrated by Robison (1964 (Weidner & Nielsen 2014). Hypagnostus truncatus differs from H. parvifrons in having a wider cephalic border and a longer pygidial axis (Høyberget & Bruton 2008). As also discussed under Pe. fallax (see above), Hartt (in Dawson 1868) described the species Agnostus acadicus. The sampled cephalon and pygidium of Dawson (1868) were later revised by € Opik (1979) and Robison (1995), who proposed that the questionable type pygidium of Hartt be assigned to Hypagnostus parvifrons (Linnarsson, 1869). In the present study the suggestion of Høyberget & Bruton (2008) is followed, as shown under Pe. fallax.
Tullberg (1880) figured two complete specimens and a single pygidium with a long, posteriorly tapered pygidial axis, in which F1 and F2 furrows are absent. The pygidial pleural fields are not rounded as in H. parvifrons and there is no forward projection of the pygidial border visible. These characters agree with Pe. scutalis. Matthew (1886) illustrated a cephalon and pygidium of Ag. acadicus with the specific characters of H. parvifrons. However, the figured cephalon is here excluded from H. parvifrons because of its subquadrate shape, wider border furrow and the presence of an anterior glabellar lobe. The specimen in figure 5a of Matthew (1886) is here assigned to Peronopsis fallax (see the Pe. fallax section). Matthew (1886) also described the new subspecies Ag. acadicus declivis. The pygidium figured by Matthew (1886) shows the characters of H. parvifrons (cf. Peng & Robison 2000). The figured cephalon in Tullberg (1880) shows a shorter glabella than in Matthew (1886), but an anterior glabellar lobe is also visible. Therefore, it is here not assigned to H. parvifrons. Furthermore, Matthew (1886) described the new species Agnostus umbo. Its cephalon shows an oval posterior glabellar lobe, an effaced anterior glabellar lobe and a narrow border furrow. The figured pygidium (Tullberg 1880) shows the same character as Matthew's (1886) figures 5b and 6b; Ag. umbo is here considered to be a synonym of H. parvifrons. Matthew (1896) also illustrated the species Ag. acadicus and the subspecies Ag. acadicus declivis. The figured pygidia of Ag. Acadicus (Matthew 1896, fig. 10b) and the subspecies ( fig. 11b) are here included in H. parvifrons, because they show the typical characters as described above. The figured cephalon in Matthew (1896) matches well with the figured cephalon in Matthew (1886) and is therefore referred to Pe. fallax. The cephalon in Matthew (1896) shows an anterior glabellar lobe and is therefore not considered to be H. parvifrons. One of the pygidia of Ag. acadicus declivis figured by Matthew (1896) shows faint horizontal furrows across the pygidial pleural fields and the anteroaxis, and another pygidium has five pits on the axis near the axial node. Because of these characters, the specimens shown in figs 11c-d of Matthew (1896) are excluded from H. parvifrons. Lorenz (1906) described the new subspecies Agnostus parvifrons latelimbatus, which he distinguished from H. parvifrons by a broader pygidial posterior border, although not as broad as in H. mammillatus. Its cephalon agrees closely with that of H. parvifrons. The pygidia figured by Lorenz (1906) have a relatively broad border, typical of H. parvifrons, although Lorenz (1906, fig. 9b) shows only a side view of a flat specimen, as shown by Høyberget & Bruton (2008, pl. 4, fig. L). In contrast, on side view of H. mammillatus the pygidium is raised. The characters described by Lorenz (1906) are here interpreted as intraspecies variations and therefore the separations into subspecies are unjustified.
Walcott (1913) figured a cephalon and a pygidium of the subspecies Ag. parvifrons latelimbatus with the characteristics of H. parvifrons. The specimens figured by Illing (1915) are poorly preserved, and a species assignment is here considered doubtful. The subspecies Ag. parvifrons punctifer described and figured by Howell (1925), shows a pygidium with a very broad border and a prominent axial node situated at the midpoint of the axis, and a side view of the pygidium with typical features of H. mammillatus. Strand (1929) described Pe. fallax, but figured several species of H. parvifrons. Therefore, the figured specimens of Pe. fallax are here assigned to this species. Cobbold & Pocock (1934) figured pygidia with a broad border and a prominent axial node, typical of H. mammillatus. Whitehouse (1939) described the new species H. clipeus, distinguished from H. parvifrons by the longer posterior glabellar lobe and the narrower border. The figured specimens fall within the morphology of H. parvifrons, and therefore H. clipeus is considered a synonym. Rasetti (1948) described the new species H. m etisensis to be very close to H. parvifrons, from which it differs only by a narrower cephalic border. However, the figured specimens have all of the characters of H. parvifrons. In addition, it is here suggested that Rasetti's (1948) described variation of the cephalic border, is an intraspecies variation and therefore H. m etisensis is here assigned as a synonym of H. parvifrons, as also discussed by Robison (1964). Rasetti (1967) figured three cephala of H. parvifrons, which show the oval posterior glabellar lobe half as long as the cephalon. Without a figured pygidium an assignment to H. parvifrons is questionable (cf. Peng & Robison 2000). € Opik (1979) figured one complete specimen and a pygidium of H. clipeus Whitehouse, 1939. The cephalon has the same characters as H. parvifrons and the pygidium has a shorter and more slender pygidial axis, with a median postaxial furrow and posteriorly rounded pleural fields. The second pygidium shows a broad and longer axis with the posterior end near the border furrow, both here suggested to be intraspecies variations of H. parvifrons (cf. Peng & Robison 2000). All illustrated specimens of Egorova et al. (1982) and Samson et al. (1990) are poorly preserved and their assignment to H. parvifrons is here considered doubtful (cf. Peng & Robison 2000). Fatka & Kordule (1992) figured a complete specimen of H. cf. parvifrons, which is here included into H. mammillatus, with its broader pygidial border. For the same reason, Rudolph's (1994) figured specimens of H. parvifrons parvifrons are here included into H. mammillatus. As discussed above, Fatka et al. (2009) figured specimens that agree with those of Fatka & Kordule (1992) and are here included into H. mammillatus. Hypagnostus parvifrons was in the past an index fossil of the H. parvifrons Zone, which is not practicable because of the long stratigraphic range of the species (from the global lower Pt. atavus Zone to the Pt. punctuosus Zone) (Høyberget & Bruton 2008 (Rasetti 1948;Robison 1964Robison , 1994€ Opik 1979;Palmer et al. 1984;Weidner & Nielsen 2014), the Pt. atavus Zone to the Pt. punctuosus Zone of Russia (Siberia) and Norway (Egorova et al. 1982;Høyberget & Bruton 2008;Wolvers & Maletz 2016) and the Pt. atavus Zone to the Lejopyge laevigata Zone of Japan, Argentina and China (Kobayashi 1939;Tortello & Bordonaro 1997;Peng & Robison 2000). The species was also reported from East Asia (Lorenz 1906 Remarks. Ptychagnostus punctuosus, the type species of Ptychagnostus (Jaekel, 1909), was originally described from an anthraconite boulder from Andrarum, Scania, Sweden. The genus characteristics of Ptychagnostus, such as spines, genal scrobiculation and surface granulation, are highly variable within populations (Robison 1984;Peng & Robison 2000;Ahlberg et al. 2007). Jaekel (1909) and Westerg ard (1946) included Ptychagnostus in the subfamily Agnostinae. Westerg ard (1946) divided the genus into two subgenera, Ptychagnostus and Triplagnostus. The family Ptychagnostidae fulfils the requirements for good index fossils in the Cambrian as discussed by Peng & Robison (2000), such as short stratigraphic ranges, wide palaeogeographic distribution in open-marine facies and abundant occurrence.  Lectotype. AR 9539, by subsequent designation of Westerg ard (1946, p. 78, pl. 12, figs 3a-b).
Description. The specimens are well-preserved. Complete specimens are 9.0 mm and 11.1 mm long, respectively. The cephala vary in size from 3.5 to 4.3 mm in width and from 3.6 to 4.4 mm in length, and the pygidia vary in size from 3.7 to 4.2 mm in width and from 3.5 to 4.3 in length. All cephala and pygidia show a characteristic granulation with visible median preglabellar furrow and scrobiculate genae. The anterior glabellar lobe is semiovate to triangular. Some specimens show a sulcus, which extends to the median preglabellar furrow. This sulcus is either a matter of preservation or an intraspecies morphological variation.
Remarks. Ptychagnostus punctuosus is a variable species, in different ontogenetic stages and also within populations (Illing 1915;Westerg ard 1946;€ Opik 1979;Robison 1984). Holaspids differ from meraspids mainly in having a coarse genal surface granulation, prominent genal scrobicules and a longer and more defined pygidial axis. The median postaxial furrow is clearly developed in the meraspis stage and becomes fainter throughout ontogeny, as observed by Lake (1906) and Illing (1915). Axheimer & Ahlberg (2003) proposed five different ontogenetic stages for the pygidia of Pt. punctuosus based on data from nine pygidia and four cephala. In this model they defined the change in the stages from meraspis to holaspis based on the previously mentioned characteristics. Although the definition is detailed, the low number of observed specimens is problematical for statistical analysis. In addition, variation of the diagnostic characters depends on the size of the specimens. Some larger specimens show a faint frontal sulcus on the anterior glabellar lobe, which extends the median preglabellar furrow. The basal lobes vary from broad and triangular to elongate and entire or divided. The scrobiculate surface of the genae is very faint to strongly accentuated and also the granulation of the genae and the pygidium varies from weak to strong. The median postaxial furrow is more developed in larger specimens, as noted by Westerg ard (1946).
Subgenera of Ptychagnostus and subspecies of Pt. punctuosus were commonly distinguished in the past (e.g. Westerg ard 1946; € Opik 1979;Berg-Madsen 1984). The most frequently described subspecies are Pt. punctuosus punctuosus and Pt. punctuosus affinis. € Opik (1979) introduced the subspecies Pt. punctuosus fermexilis. The typical characters are the slenderness of the glabella, the prominent node on M2 of the pygidial axis, the short median postaxial furrow and the absence of pygidial granulation on the pleural fields. These described characters may be variations within populations and ontogenetic stages of Pt. punctuosus. Also the illustrations of the species match with Pt. punctuosus, therefore Pt. punctuosus fermexilis is here assigned as a synonym of Pt. punctuosus (cf. Peng & Robison 2000).
The wide morphological variability between ontogenetic stages and within populations challenges the subdivision into subspecies, and it is here suggested to avoid the usage of subspecies, as also suggested by Robison (1984). The closely related species Pt. affinis differs from Pt. punctuosus by the absence of genal granules and by the weakly developed pygidial granules on the pleural fields. Ptychagnostus atavus differs in its absence of granules on the genae and on the pygidial pleural fields.
The description of the species Agnostus scutalis and the two pygidia of Salter (in Hicks 1872, pl. 5, figs 9-10) match with Pt. punctuosus. The other figures of Salter (in Hicks 1872, figs 11-14) show complete specimens of Agnostus scutalis. Howell (1925) reported the species from shales in the upper part of the type locality of the Manuels River Formation (beds 95-120) and agreed with Howell's record, but without proper description or illustrations. The species was also mentioned by Jaekel (1909), also without descriptions or figures.
Specimens of Hutchinson (1962), GSC 13048, and Martin & Dean (1988), GSC 83300-83302, originating from Manuels River were examined. Their host lithology differs significantly from that of the specimens collected for the present study. Hutchinson's (1962) material appears to originate from loose samples from the east bank of the Manuels River, as stated in his fieldbook, instead of the west bank, as for ours. The material of Martin & Dean (1988) appears to have a similar lithology to that of Hutchinson, and its origin is therefore also uncertain. Both studies (Hutchinson 1962;Martin & Dean 1988) lack lithologs or maps, they only postulate that their specimens are from the west bank of the river. Compared with the material studied here, and going by the notes in the fieldbook from Hutchinson (1962), the stated origin is doubtful. The precise level of origin of these specimens remains uncertain, as already mentioned by Austermann (2016).
Description. The specimens are mainly well-preserved. The complete specimen is 11.4 mm long. The cephala vary in size from 2.7 to 5.2 mm in width and from 2.6 to 5.4 mm in length. The pygidia vary in size from 3.1 to 4.7 mm in width and from 3.3 to 4.8 mm in length. All cephala are scrobiculate with a median preglabellar furrow. Some specimens show a cephalic axial node situated at the midpoint of the posterior glabellar lobe, on other specimens the node is missing, which possibly is a matter of preservation. The pygidia show a long, lanceolate to moderately ogival axis with the characteristic hexagonal M2 in outline. The pygidial median axial node is elongate and the secondary median node, sometimes visible, is situated at the midpoint of the posteroaxis. The median postaxial furrow becomes fainter in larger specimens and the pygidial pleural fields are weakly granulated.
Ptychagnostus affinis shows a lanceolate posteroaxis with 2-5 pairs of lateral impressions, in some Swedish specimens (Weidner & Nielsen 2009). Some specimens of Pt. atavus have a crescentic pair of furrows, located opposite the anterior glabellar lobe, which is an intraspecies variation. These furrows have not been described or figured for Pt. affinis. Ptychagnostus punctuosus is distinguished from Pt. affinis by strong granulation on cephala and pygidia (Westerg ard 1946; Robison 1984;Høyberget & Bruton 2008). The axial node on the cephalon is more prominent and the pygidial posteroaxis is more rounded than in Pt. affinis (Høyberget & Bruton 2008). This indicates that Pt. affinis is an intermediate form between Pt. atavus and Pt. punctuosus, as discussed by several authors (Westerg ard 1946;Robison 1984;Peng & Robison 2000;Axheimer & Ahlberg 2003;Høyberget & Bruton 2008). Ptychagnostus intermedius differs from Pt. affinis mainly in having a crescentic pair of furrows, located opposite the anterior glabellar lobe, a very faint and less elongate median axial node and a nearly straight F2 on the pygidium, which produce the typical pentagonal M2 in outline (Robison 1984). Brøgger (1879) described the subspecies Agnostus punctuosus affinis on the basis of a few cephala. He postulated that these cephala have the same characters as in Pt. punctuosus, but the pygidia in Pt. punctuosus have a stronger granulation compared with Ag. punctuosus affinis, together with a fainter granulation on the pygidial pleural fields. The drawings of the subspecies agree closely with Pt. affinis. Between the description of Pt. affinis by Brøgger (1879) and that by Westerg ard (1946) is a gap of 67 years, with no other known studies of Pt. affinis. Subgenera of Ptychagnostus and subspecies of Pt. punctuosus were commonly used in the past (e. g. Westerg ard 1946;Palmer 1968;€ Opik 1979).  Palmer's (1968, figs 26-27) specimens are here excluded from Pt. affinis because the pygidial posteroaxis has an ogival form and there are no faint granules, uncharacteristic of Pt. affinis. The cephalon of Palmer (1968) has a crescentic pair of furrows on the opposite side of the anterior glabellar lobe and a pair of pygidial posterolateral spines as in Pt. affinis. € Opik (1979) erected the species Ptychagnostus mesostatus on the basis of one complete exoskeleton and one pygidium. The species was defined by a posterior glabellar lobe with parallel flanks, characters that are not visible in the illustrations, making the species questionable (cf. Peng & Robison 2000). Robison (1984) elevated Pt. punctuosus affinis to species rank. Rudolph (1994) figured two cephala, here considered questionable, because characteristics are not visible (cf. Peng & Robison 2000). Hammer & Svensen (2017) figured a cephalon with characteristic trapezoid shape of the posterior glabellar lobe and smooth scrobiculate cephalic genae, and a poorly preserved pygidium without weak granulation of the pleural fields and without hexagonal M2. Although the cephalon is convincing, the assignment of the pygidium is here considered doubtful.

Diagnosis.
Cephalon convex; genae smooth to moderately scrobiculate; posterior glabellar lobe trapezoid; cephalon and pygidium non-spinose; pygidial axis with hexagonal M2 in outline and prominent median node next to rear margin of M2; small secondary median node on posteroaxis faint to moderate (based on Robison 1984;Peng & Robison 2000, with modifications).
Description. The specimens are mainly well-preserved, with all complete specimens preserved as moulds. The complete specimens are 3.7, 5. 8  Remarks. Ptychagnostus atavus is a common ptychagnostid species. It shows wide variations in morphology during different ontogenetic stages and within populations (e.g. Westerg ard 1946;Hutchinson 1962;Robison 1982;Peng & Robison 2000;Ahlberg et al. 2007;Weidner & Nielsen 2014). The cephala vary less than the pygidia throughout ontogeny. The meraspis stage is characterized by smooth to weak scrobiculate genae, stronger in the holaspis stage. Also, the median preglabellar furrow (as is the median postaxial furrow) is clearly developed in the meraspis stage and becomes increasingly fainter throughout ontogeny (Westerg ard 1946; Robison 1982). The pygidial axis grows from a short and lanceolate to a longer and more ogival shape (Fig. 10G-K). The F2 furrow is not yet developed in the meraspis stage and the axial node is often fainter than in the holaspis stage. Some cephala have a crescentic pair of furrows, which may vary in size and shape. These furrows are located opposite the anterior glabellar lobe (Robison 1982), for example, from Sweden (Westerg ard 1946), western North America (Robison 1982), western Newfoundland (Westrop et al. 1996), Hunan in China (Peng & Robison 2000) and Bornholm, Denmark (Weidner & Nielsen 2014). There is also considerable variation in the position of the axial node on the cephala: it can be found on the middle of the posterior glabellar lobe, and the vertical level on M2 varies from the F1 to the F2 furrows (e.g. Westrop et al. 1996;Ahlberg et al. 2007;Weidner & Nielsen 2009). Other intraspecies variations involve the two nodes on the pygidia and the shape of the axis. The median axial node is commonly prominent on the hexagonal M2 near the rear margin. The reason for this shape with the sometimes strongly backward-bent F2 furrow is the varying position of the node and also its size. When the node is situated in the middle of M2, the F2 furrow is less rounded than when the node is located near the margin of M2. The secondary median node is normally small and its position varies on the middle axis of the posteroaxis. It is often a matter of preservation as to whether the node is strong, faint or absent. The shape of the posteroaxis varies from moderately ogival with a tapering axis to ogival and wellrounded without a tip. On account of these variations, the systematic position of Pt. atavus is controversial, as discussed by many authors (Westerg ard 1946;€ Opik 1979;Robison 1982Robison , 1984Laurie 1988;Ahlberg et al. 2007;Weidner & Nielsen 2014). Jaekel (1909) erected Ptychagnostus and assigned Pt. atavus to this genus. Later, Westerg ard (1946) erected the subgenus Ptychagnostus (Ptychagnostus) and € Opik (1979) the subgenus Ptychagnostus (Acidusus), with the main distinguishing character being the presence of a pygidial terminal node. Laurie (1988) elevated Acidusus to genus rank. The main distinguishing characters stated by him are the position of the cephalic axial node next to F1 and a large pygidial axial node next to the F2 furrow. Laurie (1988) did not mention the pygidial terminal node, which, in contrast, had been noted by € Opik (1979). Laurie (1988) assigned the species atavus to Acidusus, but mentioned that atavus is not a typical representative of Acidusus because some specimens have a crescentic pair of furrows on the cephalic genae. Rudolph (1994), Fletcher (2006) and Weidner & Nielsen (2009 also assigned atavus to Acidusus. In contrast, it is here suggested that atavus be assigned to Ptychagnostus. Acidusus is here considered to be a synonym of Ptychagnostus (cf. Peng & Robison 2000) because of the variable morphological characters described above. Westerg ard (1946) and Robison (1994) postulate that the species Pt. atavus, Pt. affinis and Pt. punctuosus form a continuous evolutionary series.
The closely related species Pt. affinis differs from Pt. atavus by its weakly developed pygidial granules on the pleural fields. Isolated cephala of these species, therefore, may be difficult to distinguish. Ptychagnostus punctuosus differs by the strong granulation on the cephalic genae and the pygidial pleural fields. Ptychagnostus intermedius differs in having a weak pygidial median axial node, a pentagonal M2 in outline, faint granules on the pygidial pleural fields and an axial node on the cephalon generally located at the posterior glabellar midpoint, as discussed by Robison (1982). The cephalon figured by Strand (1929) is here excluded from Pt. atavus because the posterior glabellar lobe is atypically rounded. Westerg ard's (1946, pl. 11, figs 19-21) specimens are assigned to Pt. intermedius on account of the characteristic pygidial pentagonal M2 (Westerg ard 1946, fig. 21). All figured specimens are syntypes, collected by Tullberg (1880). € Opik (1979) described the new species Ptychagnostus (Acidusus) navus on the basis of a single complete specimen, here considered questionable, because the distinguishing character of Pt. atavus, the absence of the median postaxial furrow, is instead related to ontogeny or to variations within population. Ptychagnostus atavus coartatus was erected by € Opik (1979) on the basis of two cephala with the characteristics of Pt. atavus. Therefore, both species of € Opik (1979) are here considered synonyms of Pt. atavus. Furthermore, € Opik (1979) assigned a complete specimen to Pt. intermedius (Tullberg, 1880), although the pygidium shows a hexagonal M2 in outline with a prominent median node; hence, the specimen is better referred to Pt. atavus. Egorova et al. (1982) figured a subquadrate and not a rounded cephalon, and the characteristic scrobiculation of Pt. atavus is not seen in the figure. Rudolph's (1994) single, figured pygidium of Acidusus atavus is poorly preserved and its assignment is questionable. He also described the new species Ac. sterleyi, which he distinguished from Pt. atavus by the lanceolate shape of the pygidial axis and the shaped pygidial F2 furrow by the median pygidial node. These characters are above described as intraspecies variations and Ac. sterleyi is thus a synonym of Pt. atavus.
Occurrence. Ptychagnostus atavus is widespread and was reported from the middle Cambrian lower part of the Pt. atavus Zone of Sweden ( € Oland) and Denmark (Weidner & Nielsen 2009, the Pt. atavus Zone of Utah and Nevada, USA; Mexico, Norway, Sweden, Kazakhstan, Russia (Siberia), Korea and Australia ( € Opik 1979;Ergaliev 1980;Egorova et al. 1982;Robison 1982;Høyberget & Bruton 2008;Hong & Choi 2015;Ahlberg et al. 2016;Wolvers & Maletz 2016;Beresi et al. 2017), the upper part of the Pt. atavus Zone of Greenland (Robison 1994) and the Pt. atavus Zone to the lower part of the Pt. punctuosus Zone of China (Peng & Robison 2000). The species is also reported from Newfoundland and Germany, in erratic boulders (Hutchinson 1962;Rudolph 1994;Westrop et al. 1996). In the present study Pt. atavus occurs in the lower part of the Pt. atavus Zone of Eastern Newfoundland.
Description. The specimens are poorly to well-preserved. The complete specimens are 4 mm and 7.9 mm long, respectively. The cephala vary in size from 0.9 to 4.1 mm in width and from 1.1 to 3.9 mm in length, and the pygidia vary in size from 2.0 to 3.1 mm in width and from 2.0 to 2.9 mm in length. Some of the specimens have a yellow surface, typical and natural from the pyrite in the shales of the formation. The smaller specimens, meraspis stage, have smooth to weak scrobiculate cephalic genae. Larger specimens have moderately developed scrobicules, some with a well-developed frontal sulcus, which may extend into the preglabellar field. The typical elongate cephalic axial node is visible in all specimens, with the F3 furrow strongly curved anteriorly. The pygidial secondary median node and the depressed pygidial posteroaxis are visible in some specimens. This is possibly a matter of preservation.
Remarks. This species is easily distinguished from closely related species by its subelliptical to subquadrate cephalon and pygidium (Westerg ard 1946;Weidner & Nielsen 2014). The anterior glabellar lobe has a frontal sulcus developed, which often extends into the preglabellar field. Another characteristic feature is the strongly anteriorly curved glabellar F3 furrow (Høyberget & Bruton 2008). The closely related species Tomagnostus perrugatus (Gr€ onwall, 1902) differs by its strong scrobiculate cephalic genae, with often a crescentic pair of furrows situated opposite the anterior glabellar lobe (Weidner & Nielsen 2014). Both species show a depressed pygidial posteroaxis, but T. perrugatus also has a pair of pygidial posterolateral spines.
Linnarsson (1879) erected Tomagnostus fissus and described the characteristic scrobiculation on the cephalic genae. In his view, the scrobiculation on his specimens at hand was too faint, therefore he figured the cephalon as smooth. Matthew's (1896) subspecies Agnostus fissus trifissus is distinguished from T. fissus by two additional furrows at the front of the anterior glabellar lobe, parallel to the frontal sulcus. Because of the considerable variability of the development of the scrobicules on the cephalon, the subspecies trifissus is here considered merely an intraspecies variation. An example of the different cephalic scrobicules is shown by Robison (1994). In the synonymy list of Rudolph (1994), Agnostus fissus of Matthew (1896, pl. 16, fig. 10) was listed. This was actually a figure of the subspecies Agnostus fissus var. trifissus. Samson et al. (1990) figured a poorly preserved cephalon and two pygidia, which are here included into T. fissus with doubts. Rudolph (1994) listed Triplagnostus fissus of Shergold et al. (1990) in his synonymy list, although that description refers to Tomagnostus fissus. Westrop et al. (1996) assigned the species to the subgenus Ptychagnostus (Ptychagnostus) in a very wide sense, which means he accepted a wide morphological variability within T. fissus. The figured cephalon of Laurie (2008) does not show characteristics of T. fissus, that is, the cephalon has no scrobicules, the frontal sulcus is small and faint and the glabellar F3 furrow is not curved anteriorly; the assignment to T. fissus is here considered questionable.
Occurrence. Tomagnostus fissus was reported globally from the middle Cambrian Pt. gibbus Zone to the upper part of the Pt. atavus Zone (Høyberget & Bruton 2008;Weidner & Nielsen 2014). It is also reported from the Pt. gibbus Zone of Germany, in erratic boulders (Rudolph 1994), the Triplagnostus gibbus Zone to the Ac. atavus Zone of Sweden (Weidner & Nielsen 2015), the upper part of the Pt. gibbus Zone to the lower part of the Pt. atavus Zone of Sweden and Greenland (Westerg ard 1946;Robison 1994), the lower part of the Pt. atavus Zone of Denmark and Siberia (Pegel 2000;Weidner & Nielsen 2014), the Pt. atavus Zone of North and South Carolina, USA, Norway, England, and Western Newfoundland (Kindle 1982;Illing 1915;Samson et al. 1990;Høyberget & Bruton 2008) and the T. fissus Zone of Wales and Siberia (Rees et al. 2014;Korovnikov & Shabanov 2016). The species was also reported from New Brunswick, Japan and Australia (Howell 1935;Kobayashi 1939;Shergold et al. 1990). In the present study T. fissus occurs at the lower part of the T. fissus Zone of Eastern Newfoundland. Tomagnostus perrugatus (Gr€ onwall, 1902) Figures 12, 13 Manuels River Formation, type locality, Conception Bay South, Newfoundland, Canada.
Diagnosis. Strong scrobiculate genae and small pits next to the border; anterior glabellar lobe with frontal sulcus; crescentic pair of furrows located opposite the anterior glabellar lobe; pygidium rounded to pentagonal; lanceolate pygidial axis; pygidial pleural fields with small pits; pair of pygidial posterolateral spines (based on Robison 1994, with modifications).
Description. The specimens are well-preserved. The complete specimens vary from 4.3 to 11.0 mm in length. The cephala vary in size from 2.1 to 4.9 mm in width and from 2.1 to 5.4 mm in length, and the pygidia vary in size from 2.0 to 6.2 mm in width and from 2.2 to 6.3 mm in length. All cephala have strong scrobiculate genae and a crescentic pair of furrows located opposite the anterior glabellar lobe. The frontal sulcus at the anterior glabellar lobe may extend into the preglabellar field as a median preglabellar furrow. Larger specimens show characteristic small pits on the genae along the border furrow. Pygidia have a characteristic rounded to pentagonal outline with a curved to tapered pygidial margin in between the posterolateral spines. The transverse depression near midlength of the posteroaxis is well-developed. The nodes on M2 are prominently exposed. All pygidia show the posterolateral spines.
Remarks. This species is easily distinguished from the closely related T. fissus by rounded cephala and pygidia, strong scrobiculate cephalic genae, a crescentic pair of furrows opposite the anterior glabellar lobe, small pits on the pygidial pleural fields and small pygidial posterolateral spines (Weidner & Nielsen 2014). Westerg ard (1946) and Rushton (1979) described differences in pygidia shape depending on the stratigraphic horizon. According to Westerg ard (1946), specimens from earlier zones have a pygidial collar on the posterior border between the pair of posterolateral spines. Specimens from stratigraphically younger horizons have flat and broad borders and a curved to tapered margin in between the posterolateral spines. Rushton (1979) figured and described all pygidia with the curved posterior border, and some specimens with the pygidial collar. In an emended diagnosis, Robison (1994) suggested that the specimens are without or with a pygidial collar. Robison (1994, fig. 31.2-3) showed pygidia with a curved to tapered margin in between the posterolateral spines. Specimens from Newfoundland show the curved to tapered margin without pygidial collars. The pygidial collar seems to be an intraspecies variation of Tomagnostus perrugatus.
Gr€ onwall (1902) described and figured a characteristic cephalon with strong scrobiculate genae and the crescentic pair of furrows situated opposite the anterior glabellar lobe. In his view these characters were not adequate to describe a new species. Illing (1915) figured a poorly preserved cephalon, here tentatively assigned to T. perrugatus. The figure is not clear enough to show if a crescentic pair of furrows is developed. Illing (1915) erected the species Agnostus sulcatus. His four complete specimens show typical characters of T. perrugatus such as cephalic scrobiculate genae, the frontal sulcus and the crescentic pair of furrows next to the anterior glabellar lobe. Their pygidia have the characteristic rounded to pentagonal margin with a broad border, pits on the pleural fields and a pair of posterolateral spines. The poorly preserved specimen of Illing (1915) might show a pygidial collar. Agnostus sulcatus is thus here suggested to be a synonym of T. perrugatus. Fatka et al. (1981) assigned their three figured cephala to the species Tomagnostus renata. The figures and the description agree well with T. perrugatus. According to Fatka et al. (1981), the main distinguishing characters are different scrobicules on the cephalon, which are not visible on the figures. These specimens are here assigned to T. perrugatus. Egorova et al. (1982) described Tomagnostus deformis and figured three specimens, among them a complete specimen (Egorova et al. 1982, pl. 5, fig. 5) that has all the characteristics of T. perrugatus. The two other specimens are too poorly preserved for determination. Naimark & Pegel (2017) illustrated a complete specimen with specific characters of T. perrugatus.
Occurrence. Tomagnostus perrugatus is widespread and was observed from the middle Cambrian Pt. gibbus to the Pt. punctuosus Zone (Rushton 1979;Robison 1994). It has been reported in the Pt. gibbus Zone of Germany, in erratic boulders (Rudolph 1994), the Pt. gibbus to the Pt. atavus Zone of England, Sweden and the Czech Republic (Westerg ard 1946; Rushton 1979;Fatka & Kordule 1992), the Triplagnostus gibbus Zone to the Ac. atavus Zone of Sweden (Weidner & Nielsen 2015, the T. fissus Zone to the H. parvifrons Zone of Wales (Rees et al. 2014), the T. fissus Zone to the Anopolenus henrici Zone of Siberia (Naimark & Pegel 2017) and the Pt. atavus Zone of Newfoundland, Canada, Greenland, Denmark and Russia (Siberia) (Hutchinson 1962;Egorova et al. 1982;Robison 1994;Weidner & Nielsen 2014;Naimark & Pegel 2017 Diagnosis. Large anterior glabellar lobe crescentic, enclosing the posterior glabellar lobe; cephalic and pygidial borders and border furrows narrow; surface granular; spines developed; discrete spines from axial pygidial nodes (based on Shergold et al. 1990;Shergold & Laurie 1997, with modifications). Synonyms. Dichagnostus Jaekel, 1909. Remarks. The genera of the superfamily Condylopygoidea are the most morphologically differentiated members of the order Agnostida (Shergold et al. 1990; Naimark 2012). The main characters that differentiate Condylopygoidea from Agnostoidea are the expansion of the anterior glabellar lobe, the absence of basal lobes and the presence of three segments in the pygidial anteroaxis. In addition, the Condylopygoidea are characterized by their variable spines on cephala and pygidia, described by Rushton (1966Rushton ( , 1979. The superfamily contains a single family Condylopygidae and two genera, Condylopyge and Pleuroctenium (Shergold et al. 1990). Pleuroctenium is distinguished from Condylopyge by the cephalic and pygidial narrower border furrows, the granular surface and the large crescentic anterior glabellar lobe, which encloses the posterior glabellar lobe. In contrast, Condylopyge also has a large anterior glabellar lobe, but the shape is semicircular and therefore the lobe does not surround the posterior glabellar lobe. Pleuroctenium granulatum (Barrande, 1846)

Diagnosis.
Anterior glabellar lobe with variable frontal sulcus; axial node on posterior glabellar lobe on rearmost posterior part; pygidial axis broad and rounded posteriorly; lateral pygidial border and spines serrated; pygidial posterolateral spines variable in length.
Description. Most specimens are poorly preserved and fragmentary. The cephala vary in size from 0.6 to 2.6 mm in width and from 0.7 to 2.8 mm in length, and the pygidia vary in size from 1.4 to 2.9 mm in width and from 1.5 to 3.3 mm in length. Most specimens, especially the cephala, are very small. In contrast, the pygidia are larger in intersection, as described above. These variations might represent different ontogenetic stages. The anterior glabellar lobe carries a frontal sulcus and the axial node of the posterior glabellar lobe is situated at the rearmost posterior part of the lobe. The cephalic and pygidial borders are narrow. All pygidia have a broad and rounded axis with three pairs of lateral furrows and nodes. The lateral borders of the pygidia are serrated. One fragmentary pygidium has a long posterolateral spine, all other pygidia have short spines, which might be an intraspecies variation. In addition, all cephala and pygidia show a granular surface.
Remarks. Pleuroctenium granulatum shows strong morphological variations within populations (Rushton 1966(Rushton , 1979. The anterior glabellar lobe is in some specimens rounded anteriorly and truncated in other specimens. The rounded anterior lobe is interrupted by a frontal sulcus. Other specimens have a median lateral furrow through the lobe that is variable in depth ( Rushton 1966). Also, a small spine next to the axial node and posterolateral cephalic spines vary in development or are absent. A pair of pygidial posterolateral spines is always present and the spines vary in length. Fatka et al. (2004) described the serrated lateral margin of the pygidia and the spines. Between c. 22 and 30 small spines occur on each side of the margin. The closely related species Pl. bifurcatum differs from Pl. granulatum in having a rounded anterior glabellar lobe in front and a long median spine on the posterior thoracic segment, which extends backwards across the pygidial axis (Rushton 1979). In addition, the pygidial axis of Pl. bifurcatum is slender, in contrast to the broad axis of Pl. granulatum. Neither the lateral border, nor the spines are serrated in Pl. bifurcatum. Pleuroctenium tuberculatum is distinguished from Pl. granulatum by a very small anterior glabellar lobe, which is rounded anteriorly. The posterior glabellar lobe of Pl. tuberculatum is ovate and slender. The pygidial posteroaxis is shorter than that of Pl. granulatum (Rushton 1966). In addition, the lateral serration of the pygidial border and spines is absent in Pl. tuberculatum.
Hawle & Corda (1847) described the new genus Pleuroctenium to which they assigned the species Pl. granulatum. Jaekel (1909) described the new genus Dichagnostus on the basis of such characters as a large anterior glabellar lobe with furrow, a broad tripartite pygidial axis and pygidial spines. In addition, Jaekel's (1909) illustration of Pl. granulatum shows a serrated lateral margin of the pygidia and the spines, it is here assigned to Pl. granulatum. Westerg ard (1946) erected the species Pl. scanense on the basis of two cephala and four pygidia. He remarked that the specimens were too poorly preserved for a detailed description but concluded that this species is distinguished from Pl. granulatum by a deeper frontal sulcus of the anterior glabellar lobe, a slender pygidial posteroaxis and a shorter pair of pygidial posterolateral spines. In contrast, the figured specimens show the typical characters of Pl. granulatum, here suggested to be intraspecies variations. Pleuroctenium scanense is here considered a synonym of Pl. granulatum. The specimen of Snajdr (1958) consists only of fragments in poor preservation, so an assignment to this species is suggested to be doubtful. A complete specimen of Snajdr (1958) shows an anteriorly rounded anterior glabellar lobe, cephalic spines and a slender pygidial axis, characters that are typical of Pl. bifurcatum. Hutchinson (1962) misspelled the genus Pleurectinium and provided no description. The figured specimens of Smith & White (1963) are poorly preserved and the assignment to Pl. granulatum is considered questionable. Because of the wide intraspecies variations of Pl. granulatum the concept of Rushton (1966) to divide the species into three subspecies was followed by several authors. Rushton (1966) characterized Pl. granulatum granulatum by the truncate anterior glabellar lobe with a median lateral furrow and the serration at the pygidial lateral margin down to the long spines. Later, Rushton (1979), Shergold et al. (1990) and Fatka et al. (2004) followed the subspecies concept of Rushton (1966). The subspecies Pl. granulatum scanense of Rushton (1966), which was first described by Westerg ard (1946), was characterized by a frontal sulcus visible at the anterior glabellar lobe, short pygidial posterolateral spines, and serration at the lateral pygidial margin. Later, Rushton (1979), Shergold et al. (1990), Shergold & Laurie (1997) and Fletcher (2006) followed the subspecies concept of Rushton (1966). Here, it is suggested that the characters described by Rushton (1966) and the figured specimens from the studies listed above are merely intraspecies variations. The third subspecies of Rushton (1966), Pl. granulatum pileatum, was characterized by a rounded anterior glabellar lobe, long pygidial posterolateral spines and the absence of serration on the pygidial lateral margin. In addition, the figured specimens show also a slender pygidial axis. Therefore, this subspecies is here suggested to be a synonym of Pl. bifurcatum.
Occurrence. Pleuroctenium granulatum was reported from the middle Cambrian Pt. atavus Zone of Newfoundland, England and the Czech Republic ( Snajdr 1958;Hutchinson 1962;Smith & White 1963;Rushton 1966Rushton , 1979, and the upper part of the Pt. atavus Zone of Sweden (Westerg ard 1946). The species is also reported from Japan and Australia (Kobayashi 1939;Shergold et al. 1990). In the present study Pl. granulatum ranges from the T. fissus Zone to the Pt. atavus Zone of Eastern Newfoundland.
Synonyms. Microdiscus Salter, 1864;Spinodiscus Kobayashi, 1943;Deltadiscus Kobayashi, 1943. Remarks. Microdiscus scanicus, the type species of Eodiscus, was originally described from Andrarum, Scania, Sweden. Eodiscus is easily distinguished from other genera of Eodiscidae by the deep median preglabellar furrow, the narrow cephalic border, the absence of eyes, a strong occipital spine and the pygidial axis with several segmented rings. Serrodiscus differs from Eodiscus by ventral spines situated along the pygidial border, which are absent in Eodiscus. Dawsonia is distinguished from Eodiscus by a coarse crenulate cephalic border, the absence of a median preglabellar furrow, two thoracic segments and the shorter pygidial axis with only six segments (Rasetti 1952;Whittington et al. 1997). The systematic position of eodiscid trilobites is still under discussion. Babcock (1994) classified Eodiscidae as doubtful because of the phylogenetic status. Axheimer & Ahlberg (2003) classified the order as uncertain without any remarks. The order Agnostida is originally described by the matching outline of pygidium and cephalon, glabella widest at posterior end and two or three thoracic segments (Harrington et al. 1959;Whittington et al. 1997). The concept to assign all genera of Eodiscidae to the Agnostida, is here followed.
Description. The specimens are mainly well-preserved. The complete specimens, preserved as moulds, are 7.8 and 8.8 mm long, respectively. The cephala vary in size from 2.2 to 3.9 mm in width and from 1.6 to 3.6 mm in length, and the pygidia vary in size from 2.2 to 4.2 mm in width and from 1.9 to 3.9 mm in length. Some specimens have a yellow surface from the containing pyrite in the shales. The cephala and pygidia have a semiovate shape, rather broad than long. Cephala are mainly tapered anteriorly, in some cases they are more rounded. The cephalic border shows fine radiating furrows, often better preserved in moulds than on body fossils. All specimens have deep median preglabellar furrows and glabellar furrows. The cephalic occipital spine is in most specimens long, in others the spine is broken, probably a matter of preservation. The surface of the cephalic genae and the pygidial pleural fields varies from nearly smooth to strongly punctate, with the strongest punctation often on larger specimens. All specimens have a characteristic convex shape of genae and pleural fields. The pygidial axis has deep axial furrows and seven to, mainly, nine rings.
Remarks. Eodiscus punctatus shows a wide range of intraspecies morphological variations throughout ontogeny (Lake 1907; Westerg ard 1946; Rasetti 1952;Høyberget & Bruton 2008;Weidner & Nielsen 2014). The fine radiating furrows on the cephalic borders vary from very faint to well-developed. The surface ornamentation of the cephala and pygidia varies from smooth punctate to strongly punctate and is an inappropriate characteristic for species determination (Westerg ard 1946; Rasetti 1952). It also varies during different ontogenetic stages from fainter in the meraspis stage to stronger in the holaspis stage.
The occipital spine is in some specimens longer, in others shorter. In addition, the angle of the spine varies from 20°to 45°( Westerg ard 1946;Weidner & Nielsen 2014). Because of the length, the slenderness and the angle, the spine is preserved broken in some specimens. In the past only the length and the angle of the spine were used to distinguish E. punctatus from E. scanicus (Rasetti 1952;Hutchinson 1962;Poulsen 1969), but it is rarely preserved and thus is a challenging characteristic. Other characters can be used for distinguishing E. punctatus and E. scanicus, such as the wider cephalic border, the deep median preglabellar furrow, the deep and broader axial furrows and the longer occipital spine. Further, E. punctatus differs in the more convex shape of the cephalic genae and the pygidial pleural fields, and the number of pygidial axial rings, which range from seven to nine segments. In contrast, E. scanicus has 10 or more pygidial axial rings (Rasetti 1952;Høyberget & Bruton 2008;Weidner & Nielsen 2014). Linnarsson (1883) described the species Microdiscus eucentrus. He figured a cephalon and a pygidium, which both have the characteristic punctate surface of E. punctatus. The cephalon has a long occipital spine and deep axial furrows. The pygidium shows a long and narrow axis segmented by nine rings, both typical for E. punctatus, thus the species M. eucentrus is here considered a synonym of E. punctatus (cf. Høyberget & Bruton 2008).
Walcott (1884) figured three cephala and a single pygidium of Microdiscus punctatus. The pygidium shows the characteristics of E. punctatus such as a pygidial axis with nine rings and convex pygidial pleural fields. All cephala show a broad glabella with a short spine. In addition, two of the cephala have tubercles along the cephalic border. These described characters do not agree with E. punctatus, thus the three cephala are here not assigned to this species. Matthew (1886) described the subspecies M. punctatus precursor. He figured a single cephalon with an ogival-shaped cephalon, a broad glabella without an occipital spine and broad and deep median preglabellar furrow and axial furrows. In the absence of an illustrated pygidium for the subspecies, the assignment to E. punctatus is here considered questionable. Furthermore, Matthew (1886) described the subspecies M. punctatus pulchellus. He figured a cephalon, a side view of a cephalon and a pygidium. The cephalon shows a narrow median preglabellar furrow and narrow axial furrows, an occipital spine as long as the glabella, and a narrow pygidial axis segmented by 11 rings. Therefore, this subspecies is included into E. scanicus. Lake (1907) illustrated two complete specimens and four cephala assigned to M. punctatus, which all match well with E. punctatus. Further, he figured three pygidia with a long and narrow pygidial axis segmented by 10 rings. The pygidial pleural fields are not as convex as those of E. punctatus and one of the pygidia has a smooth surface. The figured pygidia of Lake (1907, figs 16-17a) are here assigned to E. scanicus. Cobbold (1911) collected scattered pygidia uncertain about the assignment to E. punctatus. The illustrated pygidia show the typical characteristics of E. punctatus such as a long narrow pygidial axis with seven ring segments. Raymond (1913) illustrated a complete specimen assigned E. punctatus with a cephalon with tubercles along the cephalic border, narrow axial furrows and a short occipital spine. These characters are not specific for this species, thus Raymond's (1913) specimen is here excluded from E. punctatus. Illing (1915) figured three complete specimens assigned to the subspecies M. punctatus scanicus. They all show a faint median preglabellar furrow and 11 rings on the pygidial axis, typical for E. scanicus. Kobayashi (1944) figured a complete specimen and erected the genus Spinodiscus. He assigned M. punctatus as the type species of the new genus. Spinodiscus is here considered a synonym of Eodiscus (see above). Hutchinson (1962) illustrated three cephala and pygidia under E. scanicus. These specimens have exactly the same characters as his figures of E. punctatus and he stated that the two species are very similar and they show intermediate forms. His specimens of E. scanicus are here included into E. punctatus. Poulsen (1969) figured a cephalon and a pygidium. The pygidium agrees well with E. punctatus but the cephalon shows a faint median preglabellar furrow and the genae are flat. Thus, the cephalon (Poulsen, 1969, fig. 2a) is here not assigned to this species. Egorova et al. (1982) figured several specimens. Two cephala have faint median preglabellar furrows and flat and large genae, which are not characteristic for E. punctatus. The specimens are here not assigned to this species. The figured cephalon and pygidium from Kindle (1982) show the typical characteristics of E. scanicus such as a faint median preglabellar furrow, a short occipital spine, 11 pygidial axial rings and a smooth surface. Martin & Dean (1988) figured two cephala and one pygidium, with typical characters of E. punctatus, therefore they are here assigned to this species (cf. Høyberget & Bruton 2008). Fletcher (2006) figured an overview of several specimens on a rock slab with a resolution that was not appropriate for the determination of any species. Therefore, an assignment to E. punctatus is here considered doubtful. Rees et al. (2014) illustrated one complete specimen and three cephala. All cephala have a faint to moderate median preglabellar furrow, which is typical for E. scanicus. The complete specimen shows eight pygidial axial rings and convex cephalic genae and pygidial pleural fields. These characters are typical for E. punctatus. Only the deep median preglabellar furrow is absent. This is a morphological intraspecies variation or a variation throughout ontogeny, hence an assignment to this species is here suggested to be questionable. Weidner & Nielsen (2014) figured six cephala and three pygidia. Almost all specimens have the typical characteristics of E. punctatus and therefore match well with the species. In contrast, all cephala have a faint median preglabellar furrow, typical of E. scanicus. Thus, the assignment to E. punctatus is here suggested to be questionable.
Occurrence. Eodiscus punctatus is widespread and has been reported from the middle Cambrian lower part of the Pt. atavus Zone of Siberia (Egorova et al. 1982), the Pt. atavus Zone of New Brunswick and Nova Scotia, Canada, and Greenland (Hutchinson 1952;Poulsen 1969), the upper part of the Pt. atavus Zone of Wales and England (Matthew 1886;Illing 1915), the upper H. parvifrons Zone of Wales (Rees et al. 2014), the upper part of the Pt. atavus Zone to the lower part of the Pt. punctuosus Zone of Newfoundland, Denmark and Sweden (Gr€ onwall 1902;Westerg ard 1946;Hutchinson 1962;Fletcher 2006;Weidner & Nielsen 2014), and the Pt. punctuosus Zone of Germany, in erratic boulders, and Norway (Rudolph 1994;Høyberget & Bruton 2008). The species is also reported from Japan (Kobayashi 1944). In the present study E. punctatus ranges from the T. fissus Zone to the Pt. atavus Zone of Eastern Newfoundland.