Journal of Systematic Palaeontology, 2013 Vol. 11, Issue 7, 743–787, http://dx.doi.org/10.1080/14772019.2012.732723 Morphology and systematics of the anomalocaridid arthropod Hurdia from the Middle Cambrian of British Columbia and Utah Allison C. Daleya,b,c∗, Graham E. Buddc and Jean-Bernard Carond,e Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 a Department of Earth Sciences, Natural History Museum, Cromwell Road, London SW7 5BD, UK; bDepartment of Earth Sciences, University of Bristol, Wills Memorial Building, Queen’s Road, Bristol BS8 1RJ, UK; cDepartment of Earth Sciences, Palaeobiology, Uppsala University, Villavägen 16, Uppsala SE-752 36, Sweden; dDepartment of Natural History, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario M5S 2C6, Canada; eDepartment of Ecology & Evolutionary Biology, University of Toronto, 25 Willcocks Street, Toronto, Ontario M5S 3B2, Canada (Received 30 August 2011; accepted 1 April 2012; first published online 22 March 2013) In Cambrian fossil Lagerstätten like the Burgess Shale, exceptionally preserved arthropods constitute a large part of the taxonomic diversity, providing opportunities to study the early evolution of this phylum in detail. The anomalocaridids, large presumed pelagic predators, are particularly relevant owing to their unique combination of morphological characters and basal position in the arthropod stem lineage. Although isolated elements and fragmented specimens were first discovered over 100 years ago, subsequent findings of more complete bodies of Anomalocaris and Peytoia, especially in the 1980s, allowed for a better understanding of these enigmatic forms. Their evolutionary significance as stem group arthropods was further clarified by the recent discovery of a third anomalocaridid taxon, Hurdia. Here, examination of hundreds of Hurdia specimens from different stratigraphical layers within the Burgess Shale and Stephen Formation, combined with statistical analyses, provides a detailed description of the taphonomy, morphology and diversity of the genus and further elucidates anomalocaridid systematics. Hurdia is distinguished from other anomalocaridids in having mouthparts with extra rows of teeth, a large frontal carapace complex and diminutive swimming flaps with prominent setal structures. The two original species, H. victoria Walcott, 1912 and H. triangulata Walcott, 1912, are confirmed based on morphometric outline analyses of the frontal carapace components combined with stratigraphical evidence; a third species, Hurdia dentata Simonetta & Delle Cave, 1975, is synonymized with H. victoria. Morphology, preservation and stratigraphical distribution suggest that H. victoria and H. triangulata share the same type of frontal appendage; a second type of appendage, previously assigned to Hurdia (Morph A), belongs to Peytoia nathorsti. These and other morphological differences between the anomalocaridids may reflect different feeding strategies. Appendages and mouthparts of Hurdia indet. sp. are also identified from the Spence Shale Member of Utah, making Hurdia and Anomalocaris the most common and globally distributed anomalocaridid taxa. Keywords: Cambrian; Burgess Shale; Radiodonta; arthropods; multivariate statistics; taxonomy Introduction The anomalocaridids (= Radiodonta Collins, 1996) are a group of large Cambrian animals with a presumed pelagic predatory lifestyle, originally described from the Burgess Shale. Their complex history of description is the result of disarticulated body elements being initially studied in isolation, reinterpreted and eventually pieced together to form several taxa of similar overall morphology (see Collins 1996, and references therein; Daley et al. 2009). When the anomalocaridids were first recognized in their entirety, it had been over 100 years since the first body parts were described. Two species, Anomalocaris canadensis Whiteaves, 1892 and A. nathorsti Walcott, 1911b were described (Whittington & Briggs 1985), the latter becoming Laggania nathorsti (Collins 1996) and then Peytoia nathorsti (Daley & Bergström 2012). Substantial ∗ Corresponding author. Email: A.Daley@nhm.ac.uk  C 2013 Natural History Museum collecting efforts in the 1980s and 1990s by the Royal Ontario Museum revealed a third anomalocaridid genus from the Burgess Shale, previously known in isolation as a teardrop-shaped carapace named Hurdia Walcott, 1912. Its description from several full-body specimens increased the range of morphological diversity observed in this group and helped identify specimens previously misidentified as Anomalocaris and Peytoia (Daley et al. 2009). A detailed phylogenetic analysis placed the three taxa together as the clade Radiodonta Collins, 1996 in the stem lineage of Euarthropoda (Daley et al. 2009). All anomalocaridids have a general morphology consisting of a head region with circular mouthparts, stalked eyes and a pair of large, spiny frontal appendages, and a posterior body region with swimming flaps (lateral lobes, Hou & Bergström 2006) with associated setal structures. Additionally, Hurdia uniquely possesses a large frontal Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 744 A. C. Daley et al. carapace complex consisting of three sclerotized elements that easily disarticulate. These elements are often found in abundance at Burgess Shale localities from various stratigraphical intervals, and provide an opportunity to elucidate the systematics of the genus using quantitative methods. The frontal carapace, if considered as homologous to the cephalic carapaces of other arthropod taxa, extends the presence of head coverings deep into the arthropod stem lineage (Daley et al. 2009). Hurdia also uniquely preserves lanceolate blade structures of the setae in considerable detail. When first described in the anomalocaridids these structures were often considered to be gills (e.g. Whittington & Briggs 1985), although they have also been interpreted as non-respiratory structural elements used for swimming or passing water through the body (Suzuki & Bergström 2008; Suzuki et al. 2008). The details visible in these structures in Hurdia specimens supports a homology with the exites on the biramous limbs of upper stem group arthropods, with implications for the evolution of the arthropod biramous limb (Budd 1996; Daley et al. 2009). Initial work describing Hurdia focused on these evolutionary implications, and all specimens were described under one species, H. victoria (Daley et al. 2009). Variations in the shape of the frontal carapace elements and characteristics of the frontal appendages prompted a detailed evaluation of the systematics and morphological diversity of this animal, in particular using a quantitative approach. The morphology of Hurdia is here described based on whole-body specimens, disarticulated assemblages and isolated body parts such as the frontal carapace elements, mouthparts and frontal appendages. Geometric morphometric analyses and stratigraphical distribution of specimens from the Burgess Shale and nearby localities are used to describe variation within the genus and clarify the systematics. Data from the stratigraphical location of specimens (i.e. number of specimens per locality) are utilized not only in this respect, but also to identify temporal variation in the abundance of each species. Most specimens originate from the Burgess Shale, but appendages and mouthparts from the Spence Shale Member in Utah are also described for the first time. Temporal trends, combined with comparisons with other anomalocaridids, allow for a greater understanding of the diversity and ecology of the anomalocaridid clade. History of research Like Anomalocaris and Peytoia (see Collins 1996), parts of Hurdia were identified separately several decades ago but only recognized as belonging to the same species after years of further collecting from the Burgess Shale (Collins 1999; Daley et al. 2009). The Hurdia animal has anomalocaridid-like mouthparts (a ‘peytoia’), a pair of frontal appendages (of ‘appendage F’ type) and a body with serially repeated units of swimming flaps and setal structures, but is unique in having a prominent frontal carapace structure. The name Hurdia was originally applied by Walcott (1912) to isolated carapace fossils found at Burgess Shale locality 35K (= ‘Phyllopod Bed’), which he interpreted as carapaces of an unknown arthropod. These carapaces have a teardrop or triangular outline, with an anterior margin ending in a sharp point and a posterior margin with two corner notches (Fig. 1A–E). Walcott (1912) originally designated two species, with H. triangulata (Fig. 1B) differing from H. victoria (Fig. 1A) in having a “valve proportionately shorter and deeper” (Walcott 1912, p. 186). A third species, Hurdia dentata, was introduced by Simonetta & Delle Cave (1975) based on a single specimen with a denticulate lower margin (Fig. 1D, E). Another part of the Hurdia frontal carapace was assigned its own genus, Proboscicaris Rolfe, 1962, described as an elongated carapace with a spatulate protrusion or beak at one end and a prominent indentation at the other. Rolfe (1962) interpreted Proboscicaris to be part of a bivalved arthropod carapace structure that consisted of two valves attached along their straight dorsal margins. Two species were described, with P. agnosta Rolfe, 1962 (Fig. 1F) having a more prominent beak structure than P. ingens Rolfe, 1962 (Fig. 1G). A third species, P. obtusa Simonetta & Delle Cave, 1975 (Fig. 1H) was erected for a single specimen that has a wider beak and shorter dorsal margin than P. agnosta. Both P. obtusa and P. ingens were synonymized with P. agnosta by Robison & Richards (1981), who considered them to be different growth stages of a single species because no rigorous characters could be found to distinguish the different species. A fourth species, P. hospes Chlupáč & Kordule, 2002, consisted of a single carapace specimen from the Middle Cambrian (Series 3, Stage 5) Jince Formation of the Czech Republic that has a similar outline to P. agnosta but with a less pronounced protrusion. It was later recognized by Collins (1999) that two Proboscicaris carapaces and one Hurdia carapace make up the three-part frontal carapace complex of the Hurdia animal. Following the terminology of Daley et al. (2009), Hurdia and Proboscicaris carapaces are now referred to as the H- and P-elements, respectively. Ironically, it was pointed out by Rolfe (1962) that Walcott was likely discussing P-elements when he stated that “there are also fragments of the carapace of a very large form that possibly may be related to Hurdia victoria” (Walcott 1912, p. 183), meaning a relationship between H- and P-elements had been implied, but not implicitly stated, when they were first described. Collins (1999) showed that the mouthparts of Hurdia differ from those of Anomalocaris nathorsti in having additional rows of teeth within the central opening. One specimen with this morphology had been assigned to A. nathorsti by Whittington & Briggs (1985), and when Collins examined this specimen (Fig. 1C) he realized that some of the Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 745 Figure 1. Original type and figured specimens of the H- and P-elements of Hurdia from the Burgess Shale. All photographs taken under high-angle polarized lighting directed from above, except where indicated. A, Hurdia victoria, lectotype, USNM 57718. B, Hurdia triangulata, lectotype, USNM 57721. C, Disarticulated assemblage of an H-element and mouthparts, originally described as the mouthparts of Anomalocaris nathorsti (Whittington & Briggs 1985) and figured without the H-element; photographed with low-angle incident lighting from bottom left; USNM 368583. D, E, Hurdia dentata, lectotype, under D, low-angle polarized light from top left and E, high-angle incident lighting from above to accentuate the denticulated margin, USNM 189152. F, Holotype of Proboscicaris agnosta, the type species of the obsolete genus Proboscicaris, which is now synonymized as the P-element of the Hurdia animal, USNM 139871. G, Holotype of Proboscicaris ingens, USNM 139865. H, Holotype and only specimen of Proboscicaris obtusa, USNM 189209. Locality: 35K = “Phyllopod Bed” = WQ. Scale bars 10 mm. For abbreviations see Appendix. mouthparts had been covered by an H-element carapace that had been excavated away to reveal the whole structure (Collins 1999). The frontal appendages of Hurdia were identified as ‘appendage F’ type (Collins 1999), a term used to identify segmented appendages with elongated ventral spines that were initially assigned to Sidneyia by Walcott (1911a), removed from that genus by Bruton (1981), described as the appendages of an unknown arthropod by Briggs (1979), and identified as belonging to Peytoia by Whittington & Briggs (1985). Three distinct ‘appendage F’ morphs are present in the Burgess Shale. Two were associated with Hurdia (Daley et al. 2009), but one of these (Morph A) is likely from Peytoia (this study), and a third morph from the S7 Burgess Shale locality has been tentatively assigned to ?Laggania (Daley & Budd 2010), which should now be considered as ?Peytoia (Daley & Bergström 2012). The ‘appendage F’ as reconstructed by Briggs (1979) is actually a combination of morphologies from several taxa, thus the term should be avoided, replaced herein with ‘frontal appendage’. A frontal appendage designated as belonging to Hurdia from The Monarch, British Columbia, has also been mentioned (Johnston et al. 2009) but was not figured and has not been examined in this study. Anomalocaridid material from the USA has typically been confined to the frontal appendages of Anomalocaris canadensis from the Latham Shale of California Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 746 A. C. Daley et al. (Briggs & Mount 1982), and of A. pennsylvanica Resser, 1929 from the Kinzers Formation of Pennsylvania (Resser 1929; Briggs 1979; Briggs & Mount 1982). Possible partially preserved anomalocaridid body specimens have been described from the Pioche Formation of Nevada (Lieberman 2003), the Marjum Formation (Briggs & Robison 1984), the Spence Shale member of the Langston Formation, and the Wheeler Formation (Briggs et al. 2008) in Utah, the last of which has also yielded anomalocaridid mouthparts (Conway Morris & Robison 1982, 1988; Briggs et al. 2008) and Proboscicaris (Robison & Richards 1981, figs 4.1, 9.3). A partial frontal appendage described from the Wheeler Formation as an indeterminate Anomalocarididae (Briggs et al. 2008, fig. 2.2) has elongated ventral spines, and so is similar to the Peytoia or Hurdia appendage. Additional frontal appendages described herein from the older Spence Shale Member extend the occurrence of Hurdia into the lower part of Cambrian Series 3 (Garson et al. 2012). Anomalocaridid material described from the Early Ordovician Fezouata Biota of south-eastern Morocco includes both H- and P-elements, as well as a possible Hurdia frontal appendage (Van Roy & Briggs 2011). Geological setting Burgess Shale The Burgess Shale is located within the Park-Main Ranges of southern Canadian Rocky Mountains (Rigby & Collins 2004, text-fig. 2). Most of this area was located along an offshelf slope and an open basin representing the deeper part of a laterally shifting environment, the Outer Detrital Belt, during the Middle Cambrian (Aitken 1997). Hurdia occurs in several stratigraphical levels within the Burgess Shale Formation (Fletcher & Collins 1998) on Fossil Ridge and on Mount Stephen near Field, and in several other smaller localities discovered by the ROM in the 1980s (Collins et al. 1983). All these sites are located along the basinal edge of a palaeotopographical feature called the Cathedral escarpment. This escarpment represents a submarine cliff at the edge of a vast carbonate platform (Aitken & McIlreath 1984) and has traditionally been interpreted to be important for the preservation and occurrence of the fossils (e.g. Conway Morris 1986). However, Burgess Shale-type fossils also occur within the lateral equivalent of the Burgess Shale Formation in the ‘thin’ Stephen Formation near Stanley Glacier in Kootenay National Park, about 40 km south-east of the type areas. This environment has recently been reinterpreted to represent a ramp setting with no evidence of an escarpment (Caron et al. 2010; Gaines 2011). Spence Shale Member The Spence Shale Member of the Langston Formation is located in northern Utah in the Wellsville Mountains and in southern Idaho in the Bear River Range (Liddell et al. 1997; Garson et al. 2012). In contrast to the classic Burgess Shale localities, deposition is not clustered near the edge of a submarine escarpment, but took place at the distal margin of a carbonate ramp (Robison 1991). Mixed carbonate-siliciclastic sediments were deposited on the muddy slope and basinal environments on the present-day western margin of Laurentia, which was a passive margin at the time of deposition (Robison 1976; Garson et al. 2012). The Spence Shale is dominated by limestones and represents deposition in a series of parasequences consisting of shallowing-upward cycles where fossil-bearing shales are replaced by lime mudstones, grainstones or nodular limestones (Garson et al. 2012). Soft-bodied preservation occurs predominantly in fine-scale laminated intervals within the shales, and to a lesser degree in bioturbated intervals (Garson et al. 2012). Material and methods Material A total of 981 Burgess Shale Hurdia specimens (Table 1) held by the Royal Ontario Museum (ROM), Geological Survey of Canada (GSC), National Museum for Natural History, Smithsonian Institution (USNM) and the Harvard University Museum of Comparative Zoology (MCZ) were identified and studied (see Online Supplementary Material). Four ROM specimens from the Spence Shale Member were also examined. Specimens were photographed digitally, often with polarizing filters placed at the camera and at the light source in order to enhance contrast between reflective and non-reflective areas of the fossils, particularly when specimens were immersed in water (Bengtson 2000). Some specimens were coated with ammonium chloride and photographed under low-angle lighting to accentuate low relief structures. Camera lucida drawings were made of selected specimens using a Nikon SMZ 1500 stereomicroscope. Measurements of specimens, and in particular the reticulate patterns of the H- and P-elements, were made in Photoshop CS4 (Adobe Systems, Inc.). For the reticulate pattern of Hurdia, a ratio (Ri) between the surface area of the largest polygon measured and the valve length was calculated (Vannier et al. 2007) and compared to the Ri of reticulate patterns in recent ostracods and Tuzoia Walcott, 1912. Many Burgess Shale specimens in the ROM collection have detailed stratigraphical information (see Online Supplementary Material), so morphological variation between specimens from different localities on Fossil Ridge and Mount Stephen of different relative ages could be examined. Hurdia occurs in 15 different localities in Yoho and Kootenay National Parks, but only the six major localities (S7 – Tulip Beds, WQ, RQ, EZ, UE, and StanG, see Appendix) were chosen for in-depth analyses because their Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Table 1. Tally of Hurdia specimens from the four largest collections of Burgess Shale material. Abbreviations as in the Appendix. ROM WQ RQ EZ UE StanG Other Total 76 68 8 0 32 1 21 10 3 16 3 3 0 0 48 20 16 12 0 0 0 0 15 197 179 18 0 110 45 8 57 14 26 14 6 4 4 216 94 58 64 0 0 0 0 24 316 233 78 5 171 111 6 54 65 105 51 6 38 7 137 44 73 20 5 0 4 1 60 23 13 8 2 16 1 11 4 6 11 6 1 4 1 8 0 5 3 2 0 1 1 17 59 44 13 2 51 0 32 19 9 21 10 0 5 5 15 0 8 7 2 0 1 1 34 39 21 16 2 27 3 13 11 11 8 14 0 10 4 6 0 5 1 1 0 0 1 0 36 32 4 0 18 6 3 9 2 4 4 1 2 1 25 6 9 10 0 0 0 0 8 746 590 145 11 425 167 94 164 110 191 102 18 62 22 455 164 174 117 10 0 5 5 158 GSC MCZ Total 197 190 6 1 113 82 9 22 3 54 3 2 0 1 133 38 42 53 1 0 0 1 7 28 23 5 0 16 10 0 6 1 7 7 1 6 0 18 6 5 7 0 0 0 0 1 10 10 0 0 8 8 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 981 813 156 12 562 267 103 192 114 254 108 21 64 23 606 208 221 177 11 0 5 6 166 Morphology and systematics of Hurdia Total number of Hurdia slabs Total isolated Hurdia elements Disarticulated assemblages Articulated assemblages H-elements total H. victoria type H. triangulata Unknown type Mouthparts P-elements Appendages in assemblage Morph A (Peytoia) Morph B (Hurdia) Unknown type Appendages in isolation Morph A (Peytoia) Morph B (Hurdia) Unknown type Appendages with carcasses Morph A (Peytoia) Morph B (Hurdia) Unknown type Setal structures S7 USNM 747 Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 748 A. C. Daley et al. stratigraphical positions are well known (Rigby & Collins 2004, text-figs 1, 2) and sample size is sufficient for quantitative comparisons. Data from other smaller localities on Fossil Ridge (FW, PZ, TZ), Mount Stephen (WS, ESG1, ESG2, ESG3, ESB) and Mount Odaray (ORF, ORU) do not materially affect or alter the stratigraphical results and conclusions presented herein. The oldest occurrence is in the Kicking Horse Shale Member (WS locality) and the youngest from Stanley Glacier (StanG) in the Waputik Member of the Stephen Formation. Hurdia occurs in all members of the Burgess Shale Formation except perhaps the Paradox and Marpole Limestone members in the type area on Fossil Ridge. Locality and stratigraphical information for all sites can be found in Collins & Stewart (1991), Fletcher & Collins (2003), Rigby & Collins (2004) and Fletcher (2011), and in O’Brien & Caron (2012) for the Tulip Beds. Specimens from the Spence Shale of Utah come from the ‘Miners Hollow’ locality (see Briggs et al. 2008 for locality information). Geometric morphometric analyses The morphological differences used to distinguish between the species of Hurdia (the H-element) and Proboscicaris (the P-element) were originally described qualitatively, with descriptions being based on differences in overall lengths and widths, or in the relative size and placement of distinctive features, such as the beak in the P-element (Walcott 1912; Rolfe 1962; Simonetta & Delle Cave 1975). Although the H- and P-elements are now known to be parts of a larger animal, their morphological variations can still provide critical clues regarding the systematics of the taxon and potential differences within or between populations, especially if they can be rigorously defined using quantitative methods. Large numbers of these body elements are preserved, allowing statistical techniques to be used to describe their morphology, if the effects of taphonomy can be taken into account. In general, the outlines of fossils with Burgess Shale-type preservation are highly variable and depend on the orientation of the fossils relative to the bedding plane when they were buried, because the animals are compressed perpendicularly to the bedding plane (Walton 1936; Whittington 1974). This aspect of Burgess Shale-type preservation is crucial for the reconstruction and interpretation of complex, three-dimensional structures, but makes the application of geometric morphometric techniques to such specimens nearly impossible. The highly variable positioning and orientation of the fossils precludes the ability to consistently trace outlines or enumerate landmarks, except when nearly planar fossils were buried parallel to bedding. Since the Hurdia H- and P-elements are essentially two-dimensional structures, morphometric techniques, such as outline analysis, are more feasible. This geometric morphometric method reduces complex shapes into a two-dimensional ordination plot that allows for quantita- tive analysis of shape variation. Taphonomy must still be considered and highly deformed specimens were excluded from the outline analyses. Oblique compression is readily identifiable in the H-element when its normally symmetrical outline is deformed and asymmetrical, but a similar check for deformation by oblique preservation could not be done for the P-element because its outline has no lines of symmetry. When geometric morphometric analyses were attempted on P-elements, the level of taphonomic noise was so great that any true morphological differences between P-elements were obscured. Thus, in-depth analysis of the morphometrics results was restricted to the H-elements only. The geometric morphometric outline analyses conducted on the H- and P-elements of the frontal carapace of Hurdia were based on digital photographs of Burgess Shale specimens. An outline drawn from the published photograph of the type specimen of Proboscicaris hospes (Chlupáč & Kordule 2002) was also included. Outlines were traced in tpsDIG version 1.40 (Rohlf 2004) and the two-dimensional Cartesian coordinates were imported into Morpheus et al. (Slice 1998) where an elliptical Fourier analysis (EFA) was conducted to reduce coordinate data to a series of sinusoidal components representing the sum of harmonically related ellipses by Fourier orthogonal decomposition (Kuhl & Giardina 1982; Ferson et al. 1985). Four sets of coefficients are derived per harmonic, and eight harmonics were sufficient to capture adequately the shape of the outlines. The four coefficients associated with the first harmonic were discarded, since they are related to the program’s routine for normalizing size and orientation, leaving a total of 28 coefficients. Perimeter length and surface area of each carapace were also calculated in Morpheus et al. (Slice 1998). The EFA coefficients were subjected to a series of ordination techniques to analyse the variation in shape using the program Canoco for Windows version 4.5 (ter Braak & Šmilauer 2002). Direct (or constrained) ordination techniques were utilized to determine the effects of size (represented by perimeter and area) and location (divided into S7, WQ, RQ, EZ, UE and StanG) on carapace outline shape. Size and location were examined because variation in outline shape could be related to ontogeny and/or evolutionary change. Initially, a detrended correspondence analysis was conducted to determine the gradient lengths of the data, in order to determine if the data show a unimodal or linear response (Hill & Gauch 1980). Since the gradient length was relatively short (1.328 SD for H-elements and 1.416 SD for P-elements), the data had a linear response (Hill & Gauch 1980) and redundancy analysis (RDA) was the most appropriate analysis for this study, as opposed to canonical correspondence analysis (CCA). RDA allows for the structure of a dataset, in this case the EFA coefficients for outline, to be analysed in the light of known and measured variables, such as size and location. RDA analysis was first conducted with all variables for size and location included to visualize Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia the structure of the outline data and how they relate to the size and locality. Then RDA was rerun to examine size and location independently using a Monte Carlo global permutation test (GPT) to determine the significance that each factor, separately, has on outline shape. A partial RDA was also run with a GPT to test if the effect of size on outline shape is still significant when the effects of location are removed, and vice versa. If in either case the factor is found to be insignificant, it means that this factor has a suboptimal importance in accounting for variability in shape outline. Finally, forward selection was applied to test the importance of size and location variables separately to decide which to include in the analysis, and RDA was conducted again using only those variables that were determined to be statistically valid. Taphonomy In all localities, specimens are preserved in homogeneous mudstone facies. Detailed sedimentological studies have shown that, at least in the Phyllopod Bed (Walcott Quarry), fossils were buried by density currents (Gabbott et al. 2008). Like other Burgess Shale fossils, Hurdia is preserved as two-dimensional kerogenized films that would have been partially replaced by aluminosilicates during late-stage diagenesis (e.g. Butterfield et al. 2007). Like other anomalocaridids, the body of Hurdia tends to disarticulate, such that approximately 83% of all known specimens consist only of isolated body components (Table 1). Specimens with multiple Hurdia components are either disarticulated assemblages with elements spread apart (16%), disarticulated assemblages with some elements close together, or articulated assemblages showing a relatively complete body with all elements intact and in their original positions. Only nine articulated assemblages and four nearly complete disarticulated assemblages are known in the four collections, representing about 1% of the specimens. No specimens show evidence of internal organs preservation, thus these articulated and disarticulated assemblages could represent either moults or decayed carcasses. Despite the general high levels of disarticulation, the relative positioning of all body parts reconstructed herein is consistent within specimens. For example, H- and Pelements are found together in close association in many disarticulated assemblages, as are frontal appendages and mouthparts (Fig. 2H). The most complete specimens also show consistent associations and provide a better approximation of what the animal would have looked like in life. The only specimen with eyes preserved (Fig. 3) shows them protruding upwards through the posterior notches of the frontal carapace complex, a position that discounts the latter having slid forward along the length of the body or flipping forward into its anterior position (see Daley et al. 2009, Online Supplementary Material). The mouthparts of Hurdia (Fig. 2) are unique in possessing up to five inner rows of teeth (Fig. 2A–C, I, J), not only because all 749 mouthpart specimens in articulated or disarticulated Hurdia assemblages with visible central openings have these extra teeth, but also because such rows of teeth have not been observed in Anomalocaris or Peytoia. The outer plates of the Hurdia mouthparts are curved into a dome shape with high relief, presumably to accommodate the extra rows of teeth, as evidenced by specimens preserved in lateral aspect (Fig. 2D, E) or with outer plates curving downwards into the sediment (Fig. 2F), and the presence of wrinkles along their outer boundaries indicating previous relief (Fig. 2G). Other thin wrinkles on the surface of typically smooth carapaces are present along structures that have a gentle relief. In H-elements, the wrinkles are preserved most often running parallel and immediately adjacent to the lateral sides of the carapaces (Fig. 4A), sometimes curving down next to the notches in the posterolateral corners (Fig. 4B–D) and eventually running parallel to the posterior margin. These wrinkles suggest that the lateral sides of the H-element, and possibly the posterior margin, were gently curved in a concavo-convex fashion. P-elements preserve wrinkles most often along their ventral margin and on the posterior protrusions and notch (Fig. 5A), as well as occasionally along the dorsal surface, indicating that this carapace also had a gentle concavo-convex curve. The wrinkles, as well as the presence of specimens with obvious bends (Fig. 5C) and rips (Fig. 1A, B, G), indicate these carapace elements were not made of a rigid or brittle material. In some specimens, carapace elements are distorted almost unrecognizably due to angle of burial, such as in ROM 53572, where one P-element has a normal outline, but the second is elongated and thin, appearing only as a thin strip of carapace extending to the frontal appendages (Fig. 5E). ROM 60030 also has highly deformed P-elements that must have been oriented obliquely to the sediment at the time of burial, such that their beaks are thin and the rest of the carapace is shortened and covered in many wrinkles indicating previous relief (Fig. 4B). Some carapaces also preserve diminutive traces (Fig. 5D, F), are draped over underlying structures such as trilobites or relief in the sediment (Figs 1H, 4D), or have a mottled texture (Fig. 4C, D), suggesting variable amounts of decay prior to burial, and in all cases an evidently quick burial. The different body elements of Hurdia have very different relative abundances at all localities examined (Fig. 6). Despite the expectation that P-elements should be twice as abundant as H-elements, the latter are always the most abundant element at every locality, P-elements and frontal appendages are highly variable, and mouthparts are the least abundant. The preservation potential of the Hand P-elements differ presumably because of differences in the thickness or structure of the cuticular carapaces. However, both H- and P-elements evidently had the same composition, and when they occur together in assemblages, they show similar preservation. It is also possible that differences in relative abundances could indicate variations Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 750 A. C. Daley et al. Figure 2. Hurdia mouthparts from the Burgess Shale. A–C, typical Hurdia mouthpart with extra rows of teeth under high-angle polarized light from above, ROM 59257; A, photograph of whole specimen; B, camera lucida drawing of mouthparts; inner teeth on outer plates are shaded in grey; C, detail of the central opening showing at least three rows of extra teeth under low-angle polarized light from bottom. D, Laterally preserved mouthpart showing relief under low-angle incident lighting from top, ROM 60039. E, Laterally preserved mouthpart under high-angle polarized lighting from top, ROM 60060. F, Tilted mouthpart showing three-dimensional structure of outer plates under low-angle incident lighting from top, ROM 60027. G, Mouthpart with wrinkles (arrows) indicating previous relief at the outer margins of outer plates under high-angle polarized lighting from above, ROM 60040. H, I, Disarticulated assemblage with H-element, two frontal appendages and the mouthparts under low-angle incident lighting from top left (H), with close-up of mouthpart showing layering of outer plates under low-angle lighting from top right (I), ROM 60019. J, Complete mouthpart with at least three rows of extra teeth under high-angle polarized lighting from top right, ROM 59260. Localities: RQ (A–F, H–J) and WQ (G). Scale bars equal 5 mm. For abbreviations see Appendix. in the tempo of moulting between the different body parts of Hurdia, with the H-elements being shed more often than the P-elements, although articulated assemblages that could represent moult assemblages usually have both types of carapace preserved together. Hydrodynamic conditions may have played a limited role in discrimination against some elements, but no evidence of size sorting or preferred directions has been detected, suggesting low currents at least at the time of deposition. This, however, does not preclude the possibility that some parts were shed within the water column and sank to the seafloor at different rates. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 751 Figure 3. Hurdia, USNM 274155 and 274158 and the counterpart 274159 from the Burgess Shale. All photographs taken underwater and high-angle polarized lighting, except where indicated. A, counterpart showing whole body with frontal carapace complex and eyes; B, camera lucida drawing of counterpart; C, part of specimen showing whole body except frontal carapace complex and eyes; D, mouthparts in lateral aspect in counterpart; E, frontal carapace complex of counterpart showing reticulate pattern and wrinkles indicating previous relief, dry and under low-angle lighting from top left; F, paired eyes on stalks in counterpart; G, lanceolate blades of the setal structures. Scale bars equal 10 mm in A–D, and 5 mm in E–G. Locality: WQ. For abbreviations see Appendix. Morphometric analysis results H-elements Of the 562 H-elements examined, 282 H-element specimens were complete enough for geometric morphometric outline analysis (Fig. 7A). RDA of the EFA coefficients segregated the H-elements into two distinct groupings, which were further accentuated when size (area and perimeter) and locality were included as variables. One grouping is characterized by a long and slender teardrop outline and contains the holotype specimen of Hurdia victoria (Fig. 1A). The second group of specimens contains the holotype of Hurdia triangulata and is characterized by a short and wide teardrop outline. The type and only specimen of Hurdia dentata Simonetta & Delle Cave, 1975 (Fig. 1D, E) falls within the H. victoria group of H-elements in the Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 752 A. C. Daley et al. Figure 4. Taphonomy of Hurdia H-elements from the Burgess Shale. All photographs taken under low-angle incident lighting from top right, unless otherwise indicated. A, Pair of frontal appendages and the H. victoria H-element with wrinkles indicating previous relief (arrows) along lateral sides, ROM 60033. B, Disarticulated assemblage with H. triangulata H-element showing wrinkles indicating previous relief (arrows) near posterior notches, and two highly deformed P-elements preserved obliquely and almost perpendicular to bedding, ROM 60030. C, H. triangulata H-element with mottled texture and wrinkles indicating previous relief (arrows) along posterior notches under high-angle polarized lighting from above, ROM 60014. D, H. triangulata H-element partially decomposed and with wrinkles indicating previous relief (arrow) near posterior notches, ROM 60054. Localities: RQ (A) and UE (B–D). Scale bars 10 mm. For abbreviations see Appendix. morphometric analyses (Fig. 7A). RDA revealed that size (area and perimeter) and locality together account for 64.1% of the variation in shape, the rest is due to unknown factors and natural variability. Of this, 35.6% of the variation in shape size is due to the interaction of these factors (size: GPT, f = 101.28, p = 0.001; location: GPT, f = 21.75, p = 0.001), with 17.5% of the variation due to size alone (GPT, f = 41.87, p = 0.001), and 11.0% due to location alone (GPT, f = 27.50, p = 0.001). When forward selection (Fig. 7A) was used to determine which factors were the most important in determining shape, perimeter, area, UE, S7, RQ and WQ (in descending order) were found to have a statistically significant influence (Table 2). Since so much of the variation in shape is due to a combination of both size and location together, it does not suggest that shape variation in H-elements is due to an ontogenetic sequence. Rather, at certain locations, the H-elements are either long and slender (H. victoria type), such as at WQ and RQ, or short and wide (H. triangulata type), such as at UE and S7, with the former generally having a longer perimeter and larger area than the latter. H. triangulata is present in low numbers at all sites, but is most abundant at S7, UE and StanG (Fig. 8). H. victoria is absent at S7, EZ and UE, but is abundant at WQ and RQ. These opposite patterns do not suggest that the two H-element morphs represent sexual variants, because if they did represent sexual Table 2. Results of forward selection analysis conducted on size and locality variables in RDA of H- and P-elements, with those with p-values less than 0.01 considered as significant. H-elements P-elements Variable Rank f -value p-value Variable Rank f -value p-value Perimeter Area UE S7 RQ WQ EZ 1 2 3 4 5 6 7 83.35 81.80 17.32 13.30 6.73 8.03 1.19 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.31 Perimeter Area EZ S7 WQ VK RQ 1 2 3 4 5 6 7 10.83 8.03 2.54 1.89 0.88 1.03 0.40 <0.001 <0.001 0.05 0.10 0.49 0.36 0.91 Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 753 Figure 5. Taphonomy of Hurdia frontal carapace complex from the Burgess Shale, showing evidence that each individual carapace was not brittle, but was subject to wrinkling, bending, distortion and colonization by trace fossil makers. All photographs taken under lowangle incident lighting from top right, unless otherwise indicated. A, P-element with wrinkles along posterior notches (with arrow) under low-angle polarized lighting from top left, ROM 60015. B, P-element preserved obliquely with wrinkles indicating previous relief (black arrow) along dorsal margin under high-angle polarized lighting from above, ROM 60022. C, H. victoria H-element with bent pointed tip suggesting flexibility of the carapace, ROM 60025. D, Pair of P-elements with diminutive trace fossils in negative and mostly positive epirelief, ROM 60013. E, Disarticulated assemblage showing a laterally preserved P-element and a second P-element that is preserved vertically and appears elongated and very thin (black arrow); adjacent is a pair of frontal appendages, under high-angle polarized lighting from top left; ROM 53572. F, H. triangulata H-element with diminutive trace fossils and reticulate pattern (white and black arrows), ROM 60045. Localities: RQ (B, C), StanG (F), UE (A, D) and WQ (E). Scale bars 10 mm. For abbreviations see Appendix. variants, one would expect to find both H-element morphs at the same localities in roughly equal numbers. It is more likely that the difference in H-element shape represents different species. Hurdia victoria is most common at RQ and WQ, and Hurdia triangulata is most common at UE, S7, EZ and StanG. P-elements Of the 254 P-elements available, 71 were complete enough for geometric morphometric outline analysis. When the EFA coefficients were subjected to RDA (Fig. 7B), the resulting scatterplots showed no segregation of the vari- ous P-elements into distinct groups. Holotype specimens of all three Proboscicaris species from Walcott Quarry (Fig. 1F–H) plot relatively close together in the centre of the scatterplot, together with specimens from S7 and WQ. The type specimen of P. hospes from the Czech Republic plots at the left edge of the cloud of points, while specimens from RQ and EZ make up the majority of specimens on the right half of the scatterplot. When forward selection (Fig. 7B) was used to identify those variables that have the greatest influence on size, only perimeter and area were selected (Table 2). Thus, location is of less importance in accounting for variability 754 A. C. Daley et al. of shape outline of P-elements than size; however, the influence of both size and location is generally small. The species of P-elements are not well segregated, and there is no strong trend in outline relating to size or location. As was initially suggested by Robison & Richards (1981), the four ‘morphospecies’ of Proboscicaris represent natural variation of a single morphological feature. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Systematic palaeontology Figure 6. Relative abundance of Hurdia body elements from the Burgess Shale, including H- and P-elements, frontal appendages, mouthparts and setal blades. Values to the right represent the total number of specimens found at each locality, including those from isolated specimens, disarticulated assemblages and carcasses or moults. For abbreviations see Appendix. Figure 7. Scatterplot of RDA results with forward selection for significant size and location variables, conducted on the EFA coefficients describing the outline of Hurdia frontal carapace elements. A, H-elements showing two main groups of outlines emphasized by dotted lines. B, P-elements with no clear groups of specimens; type specimens are indicated with stars, and localities with a statistically significant (p <0.01) effect on shape are indicated with triangles; biplots showing area and perimeter trends are indicated with arrows, and the general shape outline of each carapace element for each region of the scatterplot is shown in grey. C, Outline drawing of P-element with prominent posterior notch and anterior beak. D, Outline drawing of P-element with non-existent posterior notch and small anterior beak. E, Outline drawing of P-element with extended anterior beak and wrinkles indicating previous relief along dorsal margin. Paneuarthropoda (= Euarthropoda) Lankester, 1904 Order Radiodonta Collins, 1996 Genus Hurdia Walcott, 1912 ?1911a Amiella Walcott: 27. 1962 Proboscicaris Rolfe: 2. ?1990 Liantuoia Cui & Huo: 329. ?1990 Huangshandongia Cui & Huo: 329. Published and illustrated specimens considered as belonging to Hurdia. In addition to the genera synonymized above, several individual specimens of mouthparts, appendages and whole bodies which were previously described as different taxa are now considered as belonging to Hurdia. Hurdia appendages were initially described as the frontal appendage of Sidneyia (Walcott 1911a, pp. 25–26, pl. 4, fig. 3; Walcott 1911b, p. 517, fig. 3; Simonetta & Della Cave 1975, p. 20, pl. 7, pl. 14, fig. 3) and then as ‘Appendage F’ of an unknown anomalocaridid (Briggs 1979, pp. 641, 644–648, text-figs 23, 30, pl. 80, fig. 5, pl. 81, fig. 4). The four original species of Proboscicaris were not distinguished by the morphometric techniques used herein, and it has not been possible to assign any Proboscicaris species to either of the two retained Hurdia species. All Proboscicaris species are synonymized with P. agnosta, which is herein synonymized with Hurdia. The whole-body Hurdia specimen USNM 274155 and 274158 with counterpart 274159 (Fig. 3) was first described as Emeraldella (Simonetta & Della Cave 1975, pl. 27, fig. 5) and then as Anomalocaris (Whittington & Briggs 1985, pp. 586–588, pl. 16, figs 72–76, pl. 17, figs 77, 78, fig. 99). Hurdia mouthparts with extra teeth (Fig. 1C) (Whittington & Briggs 1985, p. 583, fig. 68, pl. 15, figs 69–71) were described as part of Anomalocaris. A pair of undetermined carapaces from Robison & Richards (1981, fig. 9.3) is herein considered to be a pair of P-elements of Hurdia. Amiella ornata may represent an incomplete body specimen of Hurdia. This taxon was first described from a single specimen (Walcott 1911a, pp. 27–28, pl. 5, fig. 4) and later placed in Sidneyia ornata based on a second specimen that eventually was redescribed as Sidneyia inexpectans (Simonetta 1963, pp. 97, 104; Simonetta & Della Cave 1975, pl. 13, fig. 7). The type was then returned to Amiella ornata (Whittington & Briggs 1985, pp. 604–606, pl. 29, figs 90–92, 94). USNM 274154 (Whittington & Briggs Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 1985, pp. 588–590, pl. 18, figs 81–86, pl. 19, figs 87–89, 100) was initially described as Anomalocaris but is likely a partial Hurdia body. Possible whole body specimens from the Pioche Formation in Nevada (Anomalocarididae gen. and sp. indet, Lieberman 2003, pp. 683–684, figs 6.3–6.5), and individual frontal appendages from the Wheeler Formation in Utah (Briggs et al. 2008, pp. 241–242, fig. 2.2) and the Monarch in British Columbia (Johnston et al. 2009, p. 98) may belong to Hurdia. Early Ordovician anomalocaridid material from the Fezouata Biota in south-eastern Morocco includes large fragments of bodies showing setal structures, possible P- (Van Roy & Briggs 2011, p. 512, figs 1d, S4a) and H-elements (Van Roy & Briggs 2011, p. 512, figs 1e-i, S4b, c), and a possible Hurdia (Morph B) frontal appendage (Van Roy & Briggs 2011, p. 512, fig. S3c, d). These specimens seem to belong to a Hurdia-like animal but additional material, in particular specimens with articulated anterior regions with carapaces, is needed to confirm this placement. Type species. Hurdia victoria Walcott, 1912. Revised diagnosis. Modified from Daley et al. (2009). Anomalocaridid with body divided into two components of subequal length; anterior with a non-mineralized reticulated frontal carapace and posterior consisting of a trunk with seven to nine lightly cuticularized segments. The frontal carapace includes a triangular H-element attached dorsally and a pair of lateral P-elements. Posterior to the frontal carapace is a pair of dorsolateral oval eyes on short annulated stalks. The anteroventral mouthparts consist of an outer radial arrangement of 32 broadly elliptical plates bearing teeth (similar to Peytoia and Anomalocaris) forming a domed structure within which is found a maximum of five inner rows of teeth (lacking in Peytoia and Anomalocaris). A pair of appendages is located on either side of the mouthparts, consisting of usually nine podomeres each, with single dorsal spines on all podomeres but the first, elongated ventral spines with single auxiliary spines and anteriorly curved tips on podomeres 2 to 6, and short, smooth ventral spines on podomeres 7 and 8. The posterior half of the body consists of seven to nine reversely imbricated lateral flaps bearing a series of wide lanceolate blades. The body lacks a posterior tapering outline and tail fan (in contrast to Anomalocaris), and the terminal body segment has two small lobe-shaped outgrowths. Occurrence. Cambrian Series 3, Stage 5, Burgess Shale Formation (Fossil Ridge, Mount Field, Mount Odaray, Mount Stephen, The Monarch); Yoho and Kootenay National Parks; and Cambrian Series 3, Stage 5, Stephen Shale Formation (Stanley Glacier), Kootenay National Park, British Columbia, Canada. Cambrian Series 3, Drumian, Jince Formation, Czech Republic. Cambrian Series 3, Stage 5, Spence Shale Member; Cambrian Series 3, Drumian, Wheeler Formation, House Range, UT, USA. 755 Figure 8. Abundance of Hurdia victoria and Hurdia triangulata H-elements, and Hurdia and Peytoia frontal appendages at different Burgess Shale localities, arranged in stratigraphical order with the oldest locality (S7-Tulip Beds) at the bottom and the youngest (StanG) at the top. H. triangulata is common in the stratigraphically higher localities, while H. victoria and Peytoia are more common in the stratigraphically lower localities. Number of specimens corresponds to all isolated elements as well as those found in disarticulated assemblages and carcass specimens. For abbreviations see Appendix. Cambrian Series 2, Stage 3, Shuijingtuo Formation, West Hubei, China. Early Ordovician (Tremadocian and Floian), Fezouata Biota, Morocco. Hurdia victoria Walcott, 1912 (Figs 1A, 4A, 5C, 9D, 10A, B, 11, 12) v∗ 1912 Hurdia victoria Walcott: 186, pl. 32, fig. 9. v.1975 Hurdia victoria (Walcott); Simonetta & Della Cave: 9, pl. 6, fig. 8; pl. 43, fig. 15; pl. 45, figs 1–5; pl. 46, fig. 1. v.1975 Hurdia dentata Simonetta & Delle Cave: 9, pl. 6, fig. 4; pl. 44, fig. 6. v.2009 Hurdia victoria (Walcott); Daley et al.: figs 2A, G. Diagnosis. Hurdia with an elongated H-element that has a maximum length twice as long as the width. Types. USNM 57718 (holotype); ROM 59254, ROM 49930 and ROM 60017 (paratypes). Material. A total of 267 specimens, of which 10 are held by the GSC, 82 by the USNM and eight by the MCZ. From the ROM collection, locality S7 is represented by one isolated specimen, WQ by 45, RQ by 111, EZ by one specimen, and StanG, WS and FW each by three specimens. Occurrence. Cambrian Series 3, Stage 5, Burgess Shale Formation (Fossil Ridge, Mount Field and Mount Stephen); Yoho and Kootenay National Parks; and Cambrian Series 3, Stage 5, Stephen Shale Formation (Stanley Glacier), Kootenay National Park, British Columbia, Canada. Cambrian Series 3, Drumian, Jince Formation, Czech Republic. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 756 A. C. Daley et al. Figure 9. P-element preservation and reticulate pattern on the surface of H- and P-elements of the Hurdia frontal carapace complex from the Burgess Shale. All photographs taken under high-angle incident lighting from directly above, unless otherwise indicated. A, disarticulated assemblage with H-element and two P-elements showing surface reticulate pattern, mouthparts and a partial appendage underwater and high-angle polarized lighting, ROM 60028. B, Close-up of two P-elements showing reticulate pattern, ROM 60028. C, Paired P-elements joined at the beak and preserved flat, with reticulate pattern visible, ROM 60047. D, Disarticulated assemblage with pointed end of H. victoria H-element with well-preserved reticulate pattern and mouthpart with extra teeth, ROM 60034. E, H. victoria H-element with reticulate pattern, ROM 60059. F, H. triangulata H-element with reticulate pattern under polarized lighting, ROM 60046. G, Two P-elements joined at their beaks and folded over on one another under low-angle polarized lighting from top left, ROM 60037. H, Paired P-elements joined at beaks and flat with low-angle incident lighting from top, ROM 59262. I, close-up of reticulate pattern on an H-element, ROM 60011. Localities: S7 (I), RQ (A–E, G), UE (H) and StanG (F). Scale bars 10 mm. For abbreviations see Appendix. Hurdia triangulata Walcott, 1912 (Figs 1B, 4B–D, 5F, 9F) Diagnosis. Hurdia with a short, wide H-element that has a maximum length 1.5 times longer than the width. v.1912 Hurdia triangulata Walcott: 186, pl. 34, fig. 1. v.1975 Hurdia triangulata (Walcott); Simonetta & Della Cave: 9, pl. 6, fig. 7; pl. 44, figs 2–4. v.2009 Hurdia victoria (Walcott); Daley et al.: fig. 1C, D. Types. USNM 57721 (holotype); ROM 59252 and ROM 59255 (paratypes). Material. A total of 103 specimens, of which nine specimens are held by the USNM. From the ROM collection, S7 Morphology and systematics of Hurdia 757 is represented by 21 specimens, WQ by eight specimens, RQ by six specimens, EZ by 11 specimens, UE by 32 specimens, StanG by 13 specimens, ESG1 by two specimens, and WS by one specimen. Occurrence. Cambrian Series 3, Stage 5, Burgess Shale Formation (Fossil Ridge, Mount Field and Mount Stephen); Yoho and Kootenay National Parks; and Cambrian Series 3, Stage 5, Stephen Shale Formation (Stanley Glacier), Kootenay National Park, British Columbia, Canada. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Hurdia indet. sp. Walcott, 1912 Material. A total of 192 specimens possess an H-element of unknown type but definitively belonging to the Hurdia genus. This includes 22 H-elements held by the USNM and six held at the GSC. From the ROM collection, S7 is represented by 10 specimens, WQ by 57 specimens, RQ by 54 specimens, EZ by four specimens, UE by 19 specimens, StanG by 11 specimens, WS by one specimen, ORF by one specimen, FW by five specimens, C100 by one specimen and ESA by one specimen. Additionally, 421 specimens consisting of Hurdia appendages, mouthparts, P-elements and carcasses are found without associated H-elements and so cannot be assigned to a Hurdia species. Occurrence. Cambrian Series 3, Stage 5, Burgess Shale Formation (Fossil Ridge, Mount Field, Mount Odaray, Mount Stephen, The Monarch), Yoho and Kootenay National Parks; and Cambrian Series 3, Stage 5, Stephen Shale Formation (Stanley Glacier), Kootenay National Park, British Columbia, Canada. Cambrian Series 3, Drumian, Jince Formation, Czech Republic. Cambrian Series 3, Stage 5, Spence Shale Member; Cambrian Series 3, Drumian, Wheeler Formation, House Range, Utah. Cambrian Series 2, Stage 3, Shuijingtuo Formation, West Hubei, China. Early Ordovician (Tremadocian and Floian), Fezouata Biota, Morocco. Anatomical description of Burgess Shale specimens Figure 10. Structure of H-elements of the Hurdia frontal carapace from the Burgess Shale, with anterior tips showing two layers of cuticle separated by sediment (white arrows). Photographs taken under low-angle incident lighting from the top right, unless otherwise indicated. A, USNM 270983. B, ROM 60023. C, ROM 60053, under low-angle polarized lighting from top left. Localities: WQ (A, C) and RQ (B). Scale bars 5 mm. For abbreviations see Appendix. Since Hurdia specimens are characterized by a high degree of disarticulation (see Taphonomy section), a complete anatomical description of the animal must necessarily take into account information from isolated body elements, disarticulated assemblages and relatively complete wholebody specimens. Detailed information about the morphology of each individual body part is best obtained from isolated specimens, while their relative placement on the body can be reconstructed by examining the articulated assemblages. Species designations have been made based on the abundant isolated body parts, in this case through morphometric analysis of the frontal carapace elements, which identified two species, H. victoria and H. triangulata. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 758 A. C. Daley et al. Figure 11. Hurdia victoria P-elements from the Burgess Shale, showing morphological variation. A, P-element with prominent posterior notch under high-angle polarized lighting from top left, ROM 61492. B, P-element with prominent posterior notch under low-angle incident lighting from top left, USNM 139868. C, P-element with non-existent posterior notch and relatively short anterior beak, under low-angle polarized lighting from bottom right, USNM 271630. D, P-element with nearly non-existent posterior notch under high-angle polarized lighting directed from above, ROM 61493. E, assemblage showing a P-element with long, thin anterior beak attached to a highly distorted P- or H-element, under low-angle incident lighting directed from bottom right, ROM 61494. F, P-element with relatively shallow height, under low-angle polarized lighting from top left, ROM 60020. G, Pair of P-elements attached at their elongated and thin anterior beaks, under high-angle polarized lighting directed from above, USNM 199058. Localities: WQ (B, C, E, G), RQ (A, F) and S7 (D). Scale bars 10 mm. For abbreviations see Appendix. The articulated assemblages and whole-body specimens provide valuable information about the morphology of the genus, but are rarely identifiable to species level because the individual body components used to designate species are typically distorted, partially covered by other body parts or absent. As such, the following anatomical description of Hurdia considers the morphology and stratigraphical distribution of each individual body element first, and then pieces together a reconstruction for the genus using information from the articulated assemblages and most complete disarticulated assemblages. Individual body elements H- and P-elements. A total of 562 H-elements and 254 Pelements were examined (Table 1). H-elements are the most common Hurdia body component at all localities, while P- elements are only relatively abundant at RQ and are rare at other localities (Table 1). A polygonal pattern is preserved on the surface of 139 H-elements and 114 P-elements (Fig. 9). This reticulation is highly reflective but either has no relief (Fig. 9A–F), or is preserved as either low, narrow ridges or valleys on the carapace surface (Fig. 9I). The polygons have between four and seven sides, with six-sided polygons being at least twice as common as any other type. Corners are sharp to rounded, and in some specimens it is difficult to determine the number of sides of each polygon due to the rounding, distortion or incompleteness of the polygons. In the best-preserved H-element specimens, the average area of the polygons on each carapace is significantly correlated with the size of that carapace (measured as length) (Pearson correlation, R = 0.72, p < 0.001, N = 22), but the same correlation was not found for P-elements (Pearson Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 759 Figure 12. Hurdia victoria frontal appendages from the Burgess Shale. All photographs taken under high-angle polarized lighting from above, unless otherwise indicated. A, Complete appendage under low-angle incident lighting from left, ROM 60026. B, Camera lucida drawing of ROM 60026. C, Complete appendage with well-preserved auxiliary spines underwater, ROM 60048. D, Camera lucida drawing of ROM 60048. E, Complete appendage with well-preserved dorsal and auxiliary spines, ROM 60020. F, Terminal spines and anterior two ventral spines of frontal appendage, ROM 60021. G, Complete appendage with well-preserved auxiliary spines, USNM 240928. Scale bars 10 mm in A–E, G; 5 mm in F. Localities: WQ (G) and RQ (A–F). For abbreviations see Appendix. correlation, R = 0.22, p = 0.39, N = 18). The area of the polygons found on H-elements (range = 0.90–13.78 mm2, mean = 3.74 mm2, SD = 2.23 mm, N = 235) is generally higher than those found in P-elements (range = 0.36–7.40 mm2, mean = 2.62 mm2, SD = 1.48 mm, N = 207). The Ri for H-elements ranges between 0.03 and 0.11, and the Ri for P-elements between 0.01 and 0.09. Few specimens preserve the reticulation over their entire surface, but comparisons of multiple specimens with partial preservation indicate that the polygonal pattern did cover the entire surface of H- and P-elements. This partial preservation also suggests that the total absence of reticulation patterns on many H- and P-elements is due to taphonomic removal. Since one single specimen can have areas both with and without a visible polygonal pattern, we suggest that either the reticulation pattern is prone to removal through partial decay of the outer cuticle layers, or that partial decay of these layers is necessary to reveal the underlying reticulate pattern. If the reticulation pattern is an internal structure, it may only be visible when the plane of cleavage passes through the level where the reticulation is located. A few H-element specimens show evidence of two separate layers of carapace separated by a thin layer of sediment (Fig. 10). Most specimens exhibit this particularly in the pointed anterior tip, but several also show a double layer of cuticle in the central region of the carapace. The Helements did not consist of a single solid piece of cuticle, but had separate dorsal and ventral cuticle layers that were joined along the lateral (Fig. 4D) and posterior margins and over most of the pointed anterior tip (Fig. 10C). P-elements appear to have been made of a single layer of carapace material, with no evidence of the double-wall structure seen in H-elements. This may partially explain why H-elements are found much more commonly than P-elements, and could be related to the moulting process of these carapaces. No similar morphology to Hurdia dentata Simonetta & Delle Cave, 1975 could be identified in any H-elements from this study. It is unclear if the denticulated margin diagnostic of this species represents a line of breakage, or if the carapace has been affected by folding or other taphonomic alteration. We herein synonymize H. dentata with H. victoria, based on the geometric morphometric analysis of H-element outlines, which places H. dentata in the H. victoria group. The morphology of the P-elements shows a great deal of variation, although taphonomic artefacts prevent Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 760 A. C. Daley et al. this morphological variation from being subdivided into discrete shape groups in the morphometric analyses (Fig. 7B). P-elements are typically roughly rectangular, with a notch at one end and an elongated ‘beak’ or anterior protrusion at the other. The posterior notch can be well defined (Figs 7C, 9A, C, G, H, 11A, B) or almost non-existent (Figs 1F, G, 5A, 7D, E, 11C). The P-element can be relatively deep or tall (Figs 1F, G, 9C, G) or shallow (Figs 5B, 7E, 11F). Beaks can also range in shape from being very truncated (Figs 1F, G, 7D, 11C) to elongated and thin (Figs 4B, 5B, E, 7E, 11E, G), though the majority have moderately sized beaks that are no larger than 25% of the total length of the dorsal margin (Figs 1H, 5A, 7C, 9A, C, G, H). Many of the P-element specimens have a beak that appears much longer than that seen in any of the original Burgess Shale Proboscicaris types (Fig. 1F–H). Some of this variation could be accounted for by angle of burial. If the P-elements were buried nearly perpendicularly along the dorsal line, the beak portion would appear similar in width to the rest of the specimen, creating an illusion where the beak looks elongated. In rare specimens, the beak region is greatly extended and relatively thin (Figs 5B, E, 11G), even exceeding the length of the rest of the carapace (Figs 4B, 11E). Variation in length of the beak is continuous between specimens, such that there are no discrete size classes of beak lengths. Only 15 assemblages have both complete Hand P-elements, with seven of these having H. triangulata H-elements, and eight with H. victoria H-elements. There is no statistically significant correlation between the ratio of P-element beak length/total length and the ratio of the H-element length/width (Pearson correlation, R = 0.17, p = 0.56, N = 15) and there is no significant difference between the means of the ratios of beak length/total length of P-elements found with H. victoria H-elements (0.25) as compared to those found with H. triangulata H-elements (0.26) (t-test, t = –0.35, p = 0.73, N = 15). The P-elements found with Hurdia victoria type H-elements do not show any consistent morphological differences when compared to those found with Hurdia triangulata H-elements, even though it might be expected that the beak length of the former would exceed that of the latter, based on the longer length of the H. victoria H-elements. Frontal appendages. In the four Burgess Shale collections examined, 725 frontal appendage specimens were identified (Table 1), 119 of which were found with other Hurdia elements (P- or H- elements, or mouthparts with extra teeth). The frontal appendages fall into two distinct morphologies (described below), both of which were originally assigned to Hurdia (Daley et al. 2009); however, reexamination of the assemblages suggest instead that only one of these types actually belongs to Hurdia (‘Morph B’ of Daley et al. 2009). The other type of frontal appendage (‘Morph A’) is assigned to Peytoia and included in Table 1. Hurdia frontal appendages (referred to as ‘Morph B’ frontal appendages by Daley et al. (2009)) usually consist of nine roughly rectangular podomeres, or more rarely 10 or 11 podomeres, which decrease in length and width distally (Fig. 12). Inconsistencies in podomere numbers are likely the result of ambiguity in identifying the boundaries of the most distal podomeres, which are small and closely packed. The dorsal margin of the appendage is convexly curved, and the most proximal podomere is rectangular and elongated in the dorsal–ventral direction, with convex margins. Approximately 10% of specimens have single, anteriorly directed spines 1–2 mm in length protruding from the dorsal surface of their podomeres (DS in Fig. 12D, E). Podomeres 2 to 6 (numbered sequentially from attachment point to distal end) bear elongated ventral spines with tips that are usually strongly curved towards the anterior (VS in Fig. 12B, D). Ventral spines range in length from 4 mm to 54 mm, with an average length of 20.5 mm (mean = 20.54 mm, SD = 6.68 mm, N = 262), and are significantly positively correlated with both length (Pearson correlation, R = 0.34, p < 0.001, N = 230) and width (Pearson correlation, R = 0.51, p < 0.001, N = 229) of podomere 2, taken as a proxy for overall size of appendage. Ventral spines are often long enough that their curved distal ends extend beyond the end of the appendage (Fig. 12A–E). A maximum of nine evenly spaced auxiliary spines are arranged along the distal margin of each ventral spine (AS in Fig. 12D, E). Auxiliary spines are robust and straight, oriented at an angle of 60–90◦ , and have an average length of 3 mm, but can reach a maximum length of 6 mm and may alternate with smaller spines approximately 1–2 mm long (Fig. 12C–E, G). Podomeres 7 and 8 often each bear one single, short, smooth ventral spine that curves distally (VS6 and VS7 in Fig. 12B, D). The most distal podomere of the Hurdia frontal appendage is generally very small and tapers sharply before terminating in one robust spine, although rare specimens exhibit two or even three terminal spines (Fig. 12F). Terminal spines are typically 2–3 mm in length and may have a slight dorsal curvature. Hurdia appendages are relatively common at all major sites, in particular RQ and StanG (Fig. 8). Of the 290 known Hurdia frontal appendages, 63 are found with definite Hurdia body elements in disarticulated assemblages (e.g. Fig. 2H), and six with articulated assemblages and nearly complete disarticulated assemblages (ROM 59320, 60012, 60017, 42985, 60038, 60029). Many of the disarticulated assemblages have setal structures and/or body segments present on the same slab, or several Hurdia body components in close proximity on the same sedimentary level. H-elements with identifiable outlines are found together with this morphology of frontal appendage in 29 disarticulated assemblages, 22 of which are the H. victoria type, and seven of which are H. triangulata. Two articulated assemblages preserve this type of frontal appendage with an H. victoria H-element (ROM 60017, 59320). Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia Peytoia frontal appendages (‘Morph A’ frontal appendages of Daley et al. (2009)) are larger and more robust than Hurdia frontal appendages, and have 11 rectangular or triangular podomeres (Po in Fig. 13B, D) with ventral, lateral and dorsal spines. The most proximal podomere is square or slightly rectangular, often with the dorsal margin wider than the ventral margin and a slight convex curve to the proximal and distal margins that imparts a distinct spindle shape. The length and width of podomeres decrease distally and this, combined with the triangular shape of some podomeres, gives the dorsal margin of the appendage a distinct arching curve (Fig. 13A–D). All podomeres except the first and last bear both dorsal and lateral spines. Single dorsal spines (DS in Fig. 13B, D, E) attach to a rounded basal protrusion located at the anterior 761 edge of each podomere (Fig. 13F). Spines are straight and directed anteriorly, commonly 3–5 mm in length but sometimes reaching a maximum length of 16 mm, and decreasing in size distally. Lateral spines (LS in Fig. 13E) are preserved uncommonly and consist either of a single spine, or one longer spine flanked by a shorter spine on either side. Lateral spines are 1–7 mm in length and usually attach at the proximal margin or in the middle of the podomere, and point towards the distal end. Long, straight ventral spines (VS in Fig. 13B, D) protrude from the ventral margin of the appendage on podomeres 2 to 6, and very rarely 7. In some specimens, the rounded distal ends of the ventral spines terminate in a long, straight auxiliary spine 5–10 mm in length (Fig. 13G). The distal margins of the ventral spines are lined with as many as eight auxiliary spines, which Figure 13. Probable Peytoia nathorsti appendages from the Burgess Shale. All specimens photographed under polarized lighting. A, Appendage under low-angle lighting directed from bottom right, ROM 60052. B, Camera lucida drawing, ROM 60052. C, Appendage under high-angle lighting directed from above, ROM 60036. D, Camera lucida drawing, ROM 60036. E, Appendage with clear lateral and dorsal spines underwater with high-angle lighting from above, ROM 60043. F, Dorsal spines of appendage under low-angle lighting directed from the top right, ROM 60051. G, Terminal end of appendage showing terminal spines and tips of ventral spines under high-angle lighting from above, ROM 60044. Localities: WQ (A–B, E–G) and RQ (C–D). Scale bars 10 mm. For abbreviations see Appendix. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 762 A. C. Daley et al. are slender and tiny, 1–3 mm in length, and oriented at an angle of 60–90◦ . The most distal podomere is wide and gently rounded, bearing as many as three terminal spines (Fig. 13G). Terminal spines are typically 4–6 cm in length, and when all three are present the middle spine is longer than the two flanking it. This second frontal appendage type was previously assigned to the Hurdia animal (Daley et al. 2009); however, we instead suggest that it belongs to Peytoia due to its rarity and poor preservation in Hurdia assemblages. Of the 229 frontal appendages with this type of morphology, only 21 are found with other Hurdia body components, and it is not present in any of the articulated Hurdia assemblages. In most of the 21 disarticulated assemblages, this type of frontal appendage is located relatively far from the Hurdia body elements, which are clustered closely together and, in some cases, on a slightly different level 2–3 mm above or below the Hurdia body components. Four of these disarticulated assemblages have a Hurdia victoria H-element, and one has a Hurdia triangulata H-element. While none of the articulated Hurdia assemblages have this type of frontal appendage associated with them, there is one wholebody Peytoia specimen (Fig. 14A–C) that has a partially preserved frontal appendage with very similar morphology to this type (Whittington & Briggs 1985, RF and LF in fig. 30). Whole-body specimens of the anomalocaridid Peytioa are found only in the S7 and WQ localities, a stratigraphical distribution similar to that of this morphology of frontal appendage, which is found in S7, WQ and RQ (Fig. 8), and both are particularly abundant in WQ. Thus, this morphology of appendage is more likely to be associated with Peytoia than Hurdia. Another appendage was described as a putative Peytoia? appendage by Daley & Budd (2010), which may represent a second Peytoia species, perhaps endemic to the S7 locality. Mouthparts. The mouthparts of Hurdia consist of outer and inner circles of plates bearing spines (Figs 1C, 2). Of the 340 mouthpart specimens examined, 114 were definitively identified as belonging to Hurdia based on morphology and close relationships with other body components. As in Peytoia, the domed outer row consists of four large tapering, subrectangular plates (Pl in Fig. 2B) that are arranged perpendicularly and separated by seven smaller plates, creating a radial structure of 32 plates (Fig. 2). There is a central opening contained within the plates that is square or slightly rectangular (X in Fig. 2B). The four large plates have three triangular spines, a large one in the centre with a smaller one on either side, extending into the central opening, and the smaller plates have only two equally sized spines. The outer row of plates is a three-dimensional structure, with stepwise partial overlaps between the plates. The four large plates (Pl in Fig. 2B) each partially cover the adjacent smaller plates, and each of these smaller plates partially overlaps the plate next to it, as do the ones next to that. The middle four smallest plates (SP in Fig. 2B) are covered partially on both sides by adjacent plates and are the lowest compared to the first series of plates. Unlike Anomalocaris and Peytoia, the oral structures of Hurdia have inner rows of plates with spines within the central opening. These arcuate plates bear as many as 11 (but commonly 5–8) small triangular spines directed inwards (Fig. 2C). They are present at five different levels that are partially stacked on top of each other, decreasing in size towards the centre, with one set of plates situated parallel to each of the four sides of the central opening. No isolated specimens of these inner plates have been found in any of the collections studied, so the lack of inner plates in Anomalocaris and Peytoia is likely real and not due to taphonomic loss (also see Daley et al. 2009, Online Supplementary Material). Setal structures. Rows of setal structures are prominent features on the trunk of the Hurdia body. They are found at all localities except StanG (Table 1), and are preserved with disarticulated (Fig. 15C, F–H) and articulated assemblages (e.g. Fig. 3), and in isolation (Fig. 15A, B, D, E). The structure consists of up to 70 lanceolate blades arranged in parallel and stacked closely together, giving a total length up to 105 mm. Individual lanceolate blades are typically 1–2 mm in width, with lengths ranging from 10 to 60 mm. Each individual blade has a pronounced dark line running along its elongated dorsal margin (Fig. 15A; Daley et al 2009, fig. 2E). Lanceolate blades attach to each other at their anterior ends and hang freely towards the posterior (Fig. 15A–C). In articulated specimens, the setal structures are located in the posterior half of the animal, behind the frontal carapaces, and span almost the entire lateral side of the animal, between the dorsal and ventral surfaces. In some articulated specimens, they are splayed out such that individual lanceolate blades are clearly visible (Fig. 3A, B, G), but in other specimens they are preserved as a series of small ridges (Figs 16, 17C, D, 18A–C, 19A, B, 20D–F) or darkened lines (Fig. 21). In disarticulated assemblages, isolated setal structures typically have a rectangular shape with constant width, although some specimens have a distinct tapering in width at one end (Fig. 15A). They are often found as paired sets arranged in series (Fig. 15C, H), and in close proximity to other Hurdia body elements (Fig. 15C, F–H). Articulated assemblages USNM 274155 and 274158, and counterpart 274159. This specimen represents the most complete example of the Hurdia animal, and is the only specimen bearing eyes (Fig. 3). It was collected by Charles Walcott from interval 35k, which refers to the Phyllopod Bed (now part of the Walcott Quarry). The counterpart was previously described by Simonetta & Delle Cave (1975) as Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 763 Figure 14. Previously described anomalocaridids from the Burgess Shale. All specimens photographed under polarized lighting. A, Cephalic region of the only specimen of Peytoia nathorsti with clear appendages preserved, underwater with high-angle lighting from above, USNM 274164. B, Close-up of appendages of Peytoia, which have a similar morphology to the isolated appendages previously described as ‘Morph A’ type (Daley et al. 2009), taken dry under high-angle lighting from top left, USNM 274164. C, Camera lucida drawing of the appendages of USNM 274164. D, Anomalocaridid specimen of unknown identity due to lack of cephalic region, with clear swimming flaps and setae (darker structures), underwater and high-angle lighting from above, USNM 274154. E, P. nathorsti with possible frontal carapace (white arrows) preserved at the anterior and lateral margin of the head seen in ventral view, dry and under low-angle lighting from top right, USNM 274142. F, Amiella ornata part underwater and high-angle lighting from above, USNM 57499. G, Close-up of setal structures (black arrows) of A. ornata in counterpart, underwater and with high-angle lighting from above, USNM 57499. H, Possible frontal carapace (white arrows) of P. nathorsti preserved with the anterior and lateral margin of the head seen in ventral view, dry and under low-angle lighting from the top left, USNM 57555. Localities: WQ. Scale bars 10 mm in A, C–H; 5 mm in B. For abbreviations see Appendix. Emeraldella brocki Walcott, 1912, and Whittington & Briggs (1985) prepared and described both the part and counterpart in detail, attributing the specimen to Anomalocaris nathorsti. This specimen is in fact Hurdia (Daley et al. 2009). Previous authors did not figure or describe its frontal carapace complex, which was considered part of another organism (Simonetta & Delle Cave 1975; Whittington & Briggs 1985). The anterior portion of this specimen is preserved in oblique lateral view, while the posterior section of the body is rotated and is preserved in an oblique dorsal view. The part (Fig. 3C, D) does not preserve the frontal carapace complex and eyes, but shows the upper dorsolateral surface of the body and all layers beneath. The counterpart (Fig. 3A, B, E–G) preserves the frontal carapace complex and eyes. Despite being complete, there is some evidence of post-mortem decay and disarticulation. The frontal carapace complex and eyes are extended relatively far forward from the rest of the body and the mouthparts appear to have been moved from their original position. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 764 A. C. Daley et al. Figure 15. Anomalocaridid setal structures from the Burgess Shale. A, Well-preserved isolated specimens where individual lanceolate blades with one darkened margin are visible, photographed underwater with high-angle lighting from above, ROM 59261. B, Isolated specimen showing lanceolate blades with darkened margins underwater with high-angle lighting from top right, ROM 60050. C, Several setal structures paired and arranged in series, in close association with mouthparts lacking extra rows of teeth in the central opening, under high-angle lighting from above, ROM 60032. D, Isolated specimen showing blades accentuated by relief, under low-angle lighting from top left, ROM 60049. E, Several isolated specimens under low-angle lighting from top left, ROM 60042. F, Disarticulated assemblage with setal structures in close association with mouthparts under low-angle lighting from top right, ROM 60031. G, Disarticulated assemblage with H-element, mouthparts with extra teeth, frontal appendage and many setal structures associated with swimming flaps, under highangle lighting from top right, ROM 60041. H, Close-up of setae in disarticulated assemblage under low-angle lighting from top right, ROM 60041. Localities: WQ (A–B, D–E), RQ (C, F), RT (G, H: talus material probably originating from RQ). Scale bars 10 mm. For abbreviations see Appendix. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 765 Figure 16. ROM 59252 from the Burgess Shale showing Hurdia in dorsal view. A, specimen dry and under low angle lighting from top right. B, Camera lucida drawing. C, Posterior region of body showing detail of setae and lobes (anterior to right), coated with ammonium chloride and under low-angle lighting from the top left. D, Close-up of frontal appendages (black arrows) and mouthparts, coated with ammonium chloride and under low-angle lighting from top right. Locality: WT (talus material probably originating from WQ). Scale bars 10 mm. For abbreviations see Appendix. The frontal carapace is large, its length almost equalling that of the rest of the body. The carapace on the top surface (H in Fig. 3B) extends from the area of the eye stalks to the most distal end, and is crossed by a small vein of calcite (V in Fig. 3B). The bottom carapace (P in Fig. 3B) consists of a roughly rectangular surface located directly anterior of the frontal appendage, and a small, triangular region near the proximal end of the carapaces, bounded at the top by the small vein. A highly reflective reticulate pattern is present on both carapace surfaces (Fig. 3E, Re in Fig. 3B). The top carapace is interpreted as the H-element, because it overlies the other carapace and extends between the two eye stalks. Its outline is unclear and the anterior point is not preserved. The lower surface has many wrinkles (A in Fig. 3B) running parallel or oblique to its curved ventral margin, and is assumed to be a P-element, oriented with Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 766 A. C. Daley et al. Figure 17. Whole body specimens of Hurdia in lateral view from the Burgess Shale. A, ROM 59254, dry and under cross-polarized, low-angle lighting from top left. B, Camera lucida drawing of ROM 59254. C, ROM 49930, coated with ammonium chloride and under incident low-angle lighting from the bottom right. D, Camera lucida drawing of ROM 49930. Localities: RQ (A–B) and UE (C–D). Scale bars 10 mm. For abbreviations see Appendix. its beak directed anteriorly and represented by the small triangular region. The eyes (E in Fig. 3B) consist of two highly reflective oval structures, measuring 10 mm in length by 5 mm in width, that are attached directly posterior to the frontal carapace by relatively short annulated stalks (ES in Fig. 3B, F). They are featureless, except for concentric, semicircular striations or wrinkles along their outer margin. The stalks are slightly narrower than the length of the eye, and extend from the eye for at least 2–3 mm. The proximal part of the stalks and attachment sites are obscured by the frontal carapace. The eye stalk of the upper eye (more dorsal) shows striations 0.5 mm apart, interpreted as annulations, and the left eye stalk preserves some wrinkling but no annulations. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 767 Figure 18. Nearly complete articulated Hurdia assemblage from the Burgess Shale, showing unique arrangement of setal structures as ‘loops’ in the posterior region of the body, ROM 59320. Specimens were photographed dry and under low-angle incident lighting from the top left. A, Overview of entire specimen. B, Camera lucida drawing. C, Close-up of setal structures on body segments 5 to 8, showing unique ‘loops’ of setae. D, Partial frontal appendage. Locality: RQ. Scale bars 10 mm. For abbreviations see Appendix. Protruding from the ventral surface, immediately posterior to the frontal carapace complex, is a partial frontal appendage consisting of at least six podomeres and four elongated ventral spines (F in Fig. 3B). The mouthparts (Fig. 3D, M in Fig. 3B) are situated immediately posterior of the frontal appendage, and in lateral aspect are directed posteriorly, as opposed to lying flat on the ventral surface. The mouthparts consist of overlapping, curved plates with pointed ends directed posteriorly and inwards. There are three large plates visible, separated by seven smaller plates each, giving a total of 17 visible plates. Eight small anterior setal structures (An in Fig. 3B) surround the dorsal, anterior and ventral margins of the mouthparts, situated posterior to the frontal appendage. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 768 A. C. Daley et al. Figure 19. Articulated assemblages of Hurdia from the Burgess Shale. A, Specimen with all body elements present except for P-elements and anterior setal structures, ROM 60017, coated with ammonium chloride and under low-angle polarized lighting directed from the bottom right. B, Camera lucida drawing of ROM 60017. C, Mostly articulated specimen with all body elements except the mouthparts and anterior setal structures, dry and under low-angle polarized lighting from the top right, ROM 42986. D, Camera lucida drawing of ROM 42986. Localities: RQ (A–B) and AW (C–D). Scale bars 10 mm. For abbreviations see Appendix. The trunk of the body has prominent setal structures preserved, alternating with layers of smooth cuticle. No triangular swimming flaps are visible in this specimen. There are at least six clear setal structures (S1–S6 in Fig. 3B), with faint impressions of a seventh between S3 and S4 (S? in Fig. 3B). They appear to cover the entire lateral side of the animal (S1 and S3 in Fig. 3B). The most posterior two sets (S5 and S6 in Fig. 3B) are arranged in a chevron orientation, with the point directed anteriorly (Fig. 3C). The point of the chevron is interpreted as the dorsal midline of the animal, with a portion of the setal structures belonging to the other Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia 769 Figure 20. Disarticulated assemblages associated with body segments from the Burgess Shale. All photographs taken under low-angle polarized lighting directed from the top right, except where indicated. A, H-element and two P-elements joined at their beaks, associated with body segments, two frontal appendages and mouthparts, ROM 60038. B, Camera lucida drawing of ROM 60038. C, Close-up of frontal appendages and mouthpart under polarized lighting directed from bottom right, ROM 60038. D, Assemblage with frontal carapace fragments, body segments and frontal appendage, ROM 60029. E, Camera lucida drawing of ROM 60029. F, Close-up of body segments with swimming flaps and setal structures in ROM 60029. Localities: RQ (A–F). Scale bars 10 mm. For abbreviations see Appendix. side of the body apparent at the posterior extremity of the body. Several layers of smooth cuticle are also present in the posterior region of this specimen, often interlayered with the setal structures. Whittington & Briggs (1985) also identified dorsal and ventral external, marginal plates (Whittington & Briggs 1985, D and V in fig. 99); however, photographs of the specimen underwater and with cross-polarized light show that there is no distinction between these plates and the smooth material that transverses the entire width of the body in between S4 and S5 (Fig. 3B; Whittington & Briggs 1985, gx in figs 75, 76, 99). The body trunk terminates abruptly, with two small lobe-shaped protrusions that extend outward from the terminus of the body (T1 and T2 in Fig. 3B). These structures are made of smooth cuticle, and no details can be seen. A. C. Daley et al. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 770 Figure 21. Articulated assemblages of Hurdia from the Burgess Shale. A, Mostly articulated assemblage with posterior end of body and frontal carapace, under high-angle polarized lighting directed from above, ROM 60010. B, Camera lucida drawing of ROM 60010. C, Specimen with most body elements, underwater and polarized lighting directed from above, ROM 60012. D, Camera lucida drawing of ROM 60012. E, Partial articulated assemblage with posterior half of body, frontal appendage and mouthparts, but no frontal carapace complex, under low-angle incident lighting directed from the top left, ROM 42985. F, Camera lucida drawing of ROM 42985. G, Close up of appendages and mouthparts, ROM 42985. Localities: StanG (A, B), UE (C–D), AW (E–G). Scale bars 10 mm. For abbreviations see Appendix. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia ROM 59252. This is the best-preserved dorsal view specimen of Hurdia (Fig. 16). All body elements, including the H- and P-elements, mouthparts, frontal appendages and body segments with setae, are present. An elongated burrow crosses the body near its midway point and disappears into the sediment beyond the body (Bu in Fig. 16B). The frontal carapace complex is preserved with the Helement pointing anteriorly and the two P-elements on either side. The P-elements are joined at their beaks, immediately below the tip of the H-element, with their straight margins running parallel to the lateral margins of the Helement. The H-element grouped with Hurdia triangulata in the morphometric analysis. It has wrinkles indicating previous relief along the posterior margin (A in Fig. 16B, D) and a surficial polygonal pattern preserved near the tip and along the posterior margin (Re in Fig. 16B). The Pelements also have the reticulate pattern and wrinkles, indicating previous relief in some areas (Fig. 16B). The specimen is 81 mm long, with the frontal carapace comprising 38% of the total length. Immediately posterior of the H-element are two partially preserved frontal appendages (Fig. 16D, F in Fig. 16B). The more complete appendage consists of the five most proximal podomeres and ventral spines (F1 in Fig. 16B). Ventral spines are long and straight, with no evidence of auxiliary spines preserved. The second appendage consists of only three podomeres and parts of their straight, apparently spineless ventral spines (F2 in Fig. 16B). Several outer plates of the mouthparts are visible due to a gap in the overlying rock (Fig. 16D, M in Fig. 16B). They consist of one larger outer plate and nine smaller plates arranged radially, with remnants of spines facing into the central opening. The mouthparts are in planar orientation. Surrounding the mouthparts and immediately posterior to the frontal appendages are three pairs of setal structures that overlap or imbricate towards the posterior (An1–6 in Fig. 16B). Individual lanceolate blades are preserved as faint lines or ridges. Five paired sets of swimming flap structures with associated setal structures follow posteriorly. These are also reversely imbricated, and the setae alternate with layers of smooth cuticle. The tail region is small and consists of several overlapping small pieces of smooth cuticle, some of which are extended posteriorly into points (T in Fig. 16B). ROM 59254. The specimen is preserved in lateral view (Fig. 17A, B). All body elements are visible except for the mouthparts, and the frontal carapace complex is particularly well preserved. The posterior half of the body is poorly preserved and lacks details. The frontal carapace complex comprises 59% of the total body length of 209 mm. The H-element of the frontal carapace complex is dorsal and oriented with its point towards the anterior (H in Fig. 17B). It has wrinkles (A in Fig. 17B) along the dorsal and anteroventral margins, and is too deformed for the type to 771 be determined. The P-elements overlie one another and are folded backwards from their attached beaks, which are directly anteriorly (P in Fig. 17B). A roughly circular dark brown structure with irregular margin overlies the P-elements (D in Fig. 17B), but no details can be seen, and it is interpreted as an unidentifiable fossil unrelated to the Hurdia body. A partially preserved frontal appendage extends down from the ventral margin immediately posterior of the P-elements (F in Fig. 17B). The appendage consists of the distal end and at least six straight ventral spines, with the two most anterior spines being shorter and straighter than the other four. No podomere boundaries or auxiliary spines are preserved. Immediately posterior of the frontal carapace on the dorsal side is an area of setal structures and smooth cuticle (An? in Fig. 17B) that is separated from the rest of the body by a pronounced ridge of sediment. The area consists of at least three distinct structures of lanceolate blades, preserved as a series of closely spaced thin ridges, and three areas of smooth cuticle. The rest of the posterior region of the body is generally poorly preserved in low relief. The dorsal region consists of a series of thin layers that overlap each other towards the posterior. In the anterior region, there are four layers of setal structures (S1 to S4 in Fig. 17B), followed by a wide area of smooth cuticle (Sm in Fig. 17B), and then another layer of setae (S5 in Fig. 17B). These layers extend from the dorsal midline to approximately halfway down the side of the body, before merging into a single cuticle surface that covers most of the ventral half of the body. This ventral cuticle is extended in two places into triangular flaps (L1 and L2 in Fig. 17B). Few details are visible posterior of this region. The dorsal side of the body has four layers of smooth cuticle, some of which have faint ridges at their bases, perhaps representing either setae (S? in Fig. 17B) or strengthening rays (SR? in Fig. 17B). On the ventral side of the body, a single large triangular flap is made of smooth cuticle (L3 in Fig. 17B). The posterior end consists of one large rectangular flap, and a smaller round flap (T in Fig. 17B). ROM 49930. ROM 49930 is another specimen in lateral aspect (Fig. 17C, D), with almost all elements visible except for the mouthparts and one P-element. The posterior region of this specimen is particularly well preserved, and the frontal carapace complex shows the relative positioning of the H- and P-elements. The specimen is 85 mm long, with the frontal carapace complex comprising 41% of this length. The H-element (H in Fig. 17D) has its pointed tip oriented towards the anterior and tilted upwards dorsally. Its outline has been distorted and many wrinkles cover the surface (A in Fig. 17D). A single P-element is situated ventral of the H-element, with its beak pointed towards the anterior (P in Fig. 17D). Most of its straight dorsal margin is missing owing to breakage in the rock, and some wrinkles are Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 772 A. C. Daley et al. visible along the ventral margin (A in Fig. 17D). A polygonal pattern is prominent in the posterior regions of the Hand P-elements (Re in Fig. 17D). The region of the body posterior of the carapaces consists of many fragmented layers of setal structures, with a partial frontal appendage projecting down from the ventral surface (Fig. 17C; F in Fig. 17D). At least five layers of setae extend from the dorsal margin to the ventral margin of the body and are reversely imbricated, with individual lanceolate blades preserved as ridges. The partially preserved frontal appendage has at least six podomeres, but no spines are visible. Posterior of the anterior setal structures and frontal appendage is a series of six or seven relatively clear body segments (S1 to S7? in Fig. 17D). The segments are delineated by setal structures that extend from the dorsal margin and nearly reach the ventral margin of the body, and are interlayered with smooth cuticle material. Triangular areas of smooth cuticle extend from the ventral margin. The second, fourth and sixth of these triangular areas (L2, L4 and L6 in Fig. 17D) are much shorter than the first, third and fifth (L1, L3 and L5 in Fig. 17D), and in some cases are on a lower sedimentary level, suggesting that they may belong to the opposite side of the body. However, the second and fourth lobes at least are relatively closely associated with setal structures on the top surface of the fossil, so the swimming flaps may simply be alternating in length in this specimen. The tail region is poorly preserved and consists of at least two large lobes and possibly a third smaller lobe (T in Fig. 17D). ROM 59320. This specimen represents a nearly complete, oblique lateral articulated assemblage (Fig. 18). All body elements except the mouthparts are present, and the frontal carapace complex, which makes up 64% of the 95 mm total length, is particularly informative, as it displays the relative arrangement of the H- and P-elements (Fig. 18A, B). The P-elements overlap one another and are joined at their beaks (P in Fig. 18B), with their dorsal margins overlain by the lateral margin of the dorsal H-element (H in Fig. 18B). Wrinkles indicating previous relief are present along the lateral margin of the H-element (A in Fig. 18B), which has an outline similar to that of Hurdia victoria, and reticulate patterns are preserved on both P-elements (Re in Fig. 18B). There is also a relatively small, incomplete carapace with a pointed margin and smooth surface anterior of the Pelement beaks (C1 in Fig. 18B). It is not a fragment of the H-element because its anterior end is complete, but could be a twisted fragment of the beak of one of the P-elements, or alternatively another portion of the ventral cephalic region of Hurdia. A single frontal appendage is located just posterior of the more complete P-element (F1 in Fig. 18B). Six ventral spines (VS in Fig. 18B), two of which bear up to three auxiliary spines (AS in Fig. 18B), extend from five indistinct podomeres. The anterior end of the frontal appendage is missing. Dorsal to the frontal appendage there are small fragments of an indistinct and featureless cuticle (C2 in Fig. 18B). A scarp or ridge in the rock (Rd1 in Fig. 18B) is present posterior of the frontal appendage and cuticle pieces, with the rest of the specimen lying 1–2 mm lower in the sediment. Just posterior of this ridge, on the ventral surface, the dorsal margins of seven podomeres of a second appendage are visible (F2 in Fig. 18B). Posterior of the appendages is an area consisting of several layers of cuticle with associated setal structures preserved as faint lines or small ridges (S? in Fig. 18B). These setal structures do not form distinct body segments, and are poorly preserved and incomplete (because of the overlying rock that could not be prepared away without destroying an overlying fossil). They could be either the anterior setal structures seen in other specimens, or unclear posterior body segments. A distinct break marked by a different orientation of body segments and a ridge of sediment (Rd2 in Fig. 18B) separates these unclear setal structures from the better-defined posterior body segments. Posteriorly, the dorsal area of the body consists of at least seven repeated units of prominent dorsolateral setal structures (S1 to S7 in Fig. 18B) in an arrangement unique to this specimen. The setal structures are solid and relatively wide, and appear to run up the lateral side of the animal, fold over at the dorsal margin and continue running down the other lateral side. These ‘loops’ of seta are not attached to each other or to any obvious cuticle or body segment, and the two sides of the setal loop are separated only by sediment. This specimen may represent a moult as opposed to a carcass, based on the absence of any obvious body holding these segments together. A third ridge of sediment separates body segment 7 from the rest of the body (Rd3 in Fig. 18B). In the posterior region of the specimen, the ventral margin is extended into swimming flaps in at least seven places (L1 to L7 in Fig. 18B). One of the swimming flaps (L3 in Fig. 18B) preserves clear strengthening rays (SR in Fig. 18B), however most are poorly preserved and incomplete. The seven swim flaps do not necessarily line up with the setal structures in the dorsal region of the body. The body terminates in a small piece of smooth cuticle that may represent part of the tail flap (T? in Fig. 18B). ROM 60017. This relatively large specimen (Fig. 19A, B) represents an articulated assemblage in dorsal aspect with the H-element, mouthparts and frontal appendages displaced from the rest of the body. No P-elements are preserved. The part preserves features more clearly than the counterpart, and coating the specimen in ammonium chloride and lighting it at an extremely low angle is the best way to distinguish details (Fig. 19A). The total length of the specimen is 187 mm, with 45% of that length accounted for by the H-element. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia The H-element of this specimen was included in the outline analysis, and plots in the Hurdia victoria shape group. It does not preserve any reticulate pattern, but has wrinkles along one lateral margin (A in Fig. 19B) and tears and cracks in the central region. Posterior of the Helement are mouthparts and two frontal appendages. The most complete appendage (F1 in Fig. 19B) has a total of nine podomeres, with five elongated and anteriorly curved ventral spines on podomeres 2 to 6 (VS1–VS5 in Fig. 19B), and two short, straight ventral spines on podomeres 7 and 8 (VS6 and VS7 in Fig. 19B). Podomere 9 is elongated into one long terminal spine. The second appendage (F2 in Fig. 19B) consists of four poorly preserved podomeres with four complete ventral spines extending from them. The tip of a fifth ventral spine is also present but highly incomplete. The mouthparts of this specimen (M in Fig. 19B) are preserved slightly oblique to the sediment such that the domed shape of the mouthpart is visible. Twenty-two outer plates are preserved, two of which are larger than the rest and separated from each other by seven smaller outer plates. Many outer plates bear 1–2 small spines that project into the central opening. No extra teeth are visible within the central opening, but very little of it is visible. Posterior of the cephalic structures, the body consists of 10 poorly delineated body segments (S1 to S10 in Fig. 19B), with wide bands of setal structures located anterior to smooth pieces of cuticle. Body segments are separated from one another by thin layers of sediment (Sm in Fig. 19B). The setal structures extend over much of the dorsolateral region of the body, but it is unclear if the blades extend directly over the dorsal mid-line. The quality of preservation deteriorates towards the posterior end, and although some faint impressions of cuticle are visible beyond segment 10, no clear tail fan is present. ROM 42986. ROM 42986 is a relatively small specimen (Fig. 19C, D) preserved in lateral aspect, representing a moult as opposed to a carcass, as is evident from the disarticulated arrangement of the frontal carapace complex. The specimen is complete except for the mouthparts and is 79 mm long, with the frontal carapace complex comprising 41% of that length. The frontal carapace complex was displaced from the rest of the body as a unit such that the H- and P-elements are oriented with their long axes perpendicular to the length of the body and the pointed tip of the H-element directed dorsally (H and P in Fig. 19D). The H-element is similar to Hurdia victoria, but is too incomplete to be included in the outline analysis. There is no observable reticulate pattern, but wrinkles are present along one lateral margin (A in Fig. 19D). A partial P-element lies next to the H-element, but is mostly covered by overlying sediment. It consists of the flat posterior region and the base of the beak. The lateral margin of the H-element overlies the dorsal margin of the P-element. 773 Posterior of the frontal carapace complex, on the ventral surface, two frontal appendages are partially preserved (F in Fig. 19D). The more complete appendage (F1 in Fig. 19D) has podomeres 1 to 6, with complete ventral spines on podomeres 2 to 5 (VS1–VS4 in Fig. 19D). The ventral spines are straight and thin. The second frontal appendage (F2 in Fig. 19D) has only two or three ventral spines preserved at an oblique angle to bedding. Dorsal of the appendages is a darkened area of unclear carapace fragments. The darkening may be associated with setal structures, based on the preservation of the rest of the body, but no individual lanceolate blades are visible. It is possible that this region represents the anterior setae (An? in Fig. 19D). The posterior region of this specimen is flattened such that the seven body segments (S1–S6 and B7 in Fig. 19D) are not preserved on different sedimentary layers but are instead distinguished based on alternating areas of smooth cuticle and associated bands of setal structures along the length of the body. Lanceolate blades of the setal structures are visible in at least three different places (S1, S2 and S5 in Fig. 19D). The setae extend three-quarters of the way down the body from the dorsal margin, and the ventral margin consists of featureless cuticle that in some places extends down into pointed areas that could represent swimming flaps (L? in Fig. 19D). Posterior of body segment 7, the body terminates in two irregular tail lobes (T in Fig. 19D). ROM 60010. This is a possible dorsolateral view specimen with relatively poor preservation from Stanley Glacier (Fig. 21A, B). Frontal appendages and mouthparts are not present. The frontal carapace complex comprises 45% of the total body length of 78 mm and consists of a relatively complete H-element oriented with its point towards the anterior (H in Fig. 21B). It has wrinkles running along one margin (A in Fig. 21B), and preserves several single, unbranching burrows along its surface (Bu in Fig. 21B). It partly overlies two partial P-elements in lateral orientation. One P-element (P1 in Fig. 21B) consists of an elongated rectangular fragment with many wrinkles (A in Fig. 21B) to the side of the H-element, and the other is an irregular fragment preserved posterior to the H-element (P2 in Fig. 21B) with wrinkles along one side (A in Fig. 21B) and several burrows (Bu in Fig. 21B). Posterior to the frontal carapace, the body of this specimen is poorly preserved with very few recognizable details. It consists of a darkened body outline, with dark grey to black regions within. There are no distinct levels of setal structures or smooth cuticle, although three triangular flaps (L1 to L3 in Fig. 21B) bearing strengthening rays (SR in Fig. 21B) are visible on the left side of the specimen. The irregularly shaped black areas within the body cavity may represent setal structures, but individual lanceolate blades cannot be identified. Several burrows pass through the body and are often preserved in dark orange colour (Bu in 774 A. C. Daley et al. Fig. 21B). The body terminates in a single rounded flap (T in Fig. 21B). Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Disarticulated assemblages ROM 60012. This small specimen from UE preserves a nearly complete body with lateral orientation, and has disarticulated mouthparts, one frontal appendage and a fragment of frontal carapace (Fig. 21C, D). The length of the body is 32 mm, not including the highly incomplete fragment of frontal carapace. The frontal appendage (F in Fig. 21D) consists only of the seven most distal podomeres (P3 to P9), with the first two being covered by the mouthparts. Podomere 4 has a dorsal spine that is anteriorly directed (DS in Fig. 21D). There are five elongated ventral spines with anteriorly curved distal ends and bearing up to five auxiliary spines. Two shorter, straight ventral spines are just anterior of the larger spines. Podomere 9 terminates in a single, large, thick terminal spine (TS in Fig. 21D). The mouthparts (M in Fig. 21D) have a typical outer arrangement of plates, which appears to have undergone distortion such that the most ventral large plate is overlapped by the two smaller plates on either side of it, and the general outline of the mouthparts is slightly oval. Within the rectangular central opening, four rows of inner plates with teeth (X in Fig. 21D) are preserved as dark denticulate lines. A broken fragment of featureless carapace is located dorsal to the mouthparts, which likely represents an H- or P-element (H/P? in Fig. 21D). The body consists of a series of setal structures and smooth cuticle. The setal structures are darkened compared to the rest of the body parts. Just dorsal to the mouthparts, frontal appendage and carapace, there are four sets of dark bands associated with the bases of the anterior setal structures (An1 to An4 in Fig. 21D). Lanceolate blades are visible in An1. A thin strip of dark brown material runs between these darkened bands, but no other structures are preserved. Body segments 5 to 11 preserve large ventral swimming flaps (L1 to L6 in Fig. 21D) in association with smooth, featureless cuticle material that extends over the dorsal surface. The swimming flaps are rounded triangles in most cases. The first body segment has a darkened band associated with it that is similar to those seen in the anterior setal structures, as well as some evidence of short lanceolate blades. Segments 2 to 6 preserve wide bands of setae along the dorsal margin that extend onto some of the swimming flaps (e.g. L3 in Fig. 21D). The posterior end of the specimen ends in a rounded piece of smooth cuticle with two protrusions (T in Fig. 21D). ROM 42985. This specimen (Fig. 21E, G) from UE is missing the posterior section due to rock breakage. It has two well-preserved frontal appendages adjacent to fragmented mouthparts (Fig. 21G), and a poorly preserved body consisting mostly of overlapping setal structures. The cara- pace complex is not visible, but the total preserved length of the specimen is 88 mm. The most complete frontal appendage (F1 in Fig. 21F) has 10 podomeres, with five elongated ventral spines (rVS1–rVS5 in Fig. 21F), one shorter straight ventral spine (rVS6 in Fig. 21F), and a long terminal spine (rTS in Fig. 21F). The second frontal appendage (F2 in Fig. 21F) preserves only the anterior podomeres and the ventral spines. There are five elongated ventral spines (lVS1–1VS5 in Fig. 21F) with auxiliary spines (AS in Fig. 21F), two shorter ventral spines anterior to these (lVS6–lVS7 in Fig. 21F), and an elongated terminal spine (lTS in Fig. 21F). Posterior of the frontal appendages is a fragment of the mouthparts in lateral orientation (M in Fig. 21F), including 15 outer plates that are curved as though the mouthparts were domed (Fig. 21G). The body is situated beside the mouthparts and appendages, and consists mainly of eight well-preserved bands of setal structures (S1 to S8 in Fig. 21F) in association with some smaller pieces of smooth cuticle, arranged in reverse imbrication. Individual lanceolate blades are preserved as dark grey or black fine lines on the rock. They form thick, wide bands that extend almost the full width of the specimen, and there are at least eight different levels present. Lanceolate blades emerge from a straight margin of attachment and arch towards the posterior. Rounded areas of smooth cuticle adjacent to the setal blades may represent swimming flaps (L? in Fig. 21F), and the posterior of this specimen terminates in small fragments of smooth cuticle that extend outward into two possible tail lobes (T? in Fig. 21F). ROM 60038. This specimen from RQ (Fig. 20A–C) shows an unusual configuration of Hurdia animal parts, with the frontal carapace complex splayed out and the mouthparts and appendages dislocated and shifted backwards. It includes two frontal appendages, dome-shaped mouthparts, two P-elements, one H-element and some small pieces of body segments with setal structures. These parts, however, are not in their original positions, and the body is too fragmented for length measurements. The frontal carapace complex is splayed out such that the P-elements are joined at the beaks and lying flat (P1 and P2 in Fig. 20B), and the H-element is preserved with its point facing the attached beaks (H in Fig. 20B). The Hand P-elements have excellent preservation of reticulations (Re in Fig. 20B) and wrinkles indicating previous relief (A in Fig. 20B). Although incomplete, the H-element has a similar posterior outline to that of Hurdia victoria. The faint relief of two large setal structures can be seen beneath the H-element (S in Fig. 20B). The frontal appendages and mouthparts (Fig. 20C) are preserved as a unit in between the H-element and the P-element on the right (P2 in Fig. 20B). The two frontal appendages are on either side of the mouthpart, with their anterior margins pointing in the same direc- Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia tion (Fig. 20C). The more complete frontal appendage (F1 in Fig. 20B) preserves four elongated ventral spines (VS in Fig. 20B) that curve towards the anterior, two of which have auxiliary spines preserved (AS in Fig. 20B), and two additional short, straight ventral spines anterior to this (VS5–6 in Fig. 20B), and a small terminal segment with single spine (TS in Fig. 20B). The second frontal appendage is incomplete and only a few podomeres are visible (F2 in Fig. 20B). The mouthparts are preserved in lateral orientation, such that the original domed shape of the structure is visible (M in Fig. 20B). A total of 29 outer plates are present, but only two of the large plates can be identified (Pl in Fig. 20B). The posterior end of a Leanchoilia specimen (Le in Fig. 20B) is preserved in sediment approximately 1 mm below the mouthparts. Several fragments of setae (S in Fig. 20B) and smooth cuticle extend from the mouthparts and frontal appendages into the area between the ventral margins of the P-elements and between the left P-element (P1 in Fig. 20B) and the H-element. These areas are highly disarticulated, and no clear body segments can be identified. Setal blade structures are preserved as a series of closely spaced ridges making up some fragments. There are at least three distinct setal fragments in these areas (S in Fig. 20B). The rest are smooth, featureless cuticular fragments. ROM 60029. This specimen (Fig. 20D–F) from RQ is a highly distorted assemblage that includes a frontal appendage, fragments of the frontal carapace complex, and body segments with setae and swimming flaps. Length of the body and frontal carapace complex cannot be determined. Three partially preserved carapaces are assumed to be the H- and P-elements but their outlines are too incomplete to make a positive identification (H/P in Fig. 20E). Outer margins, when present, are rounded and often lined with wrinkles (A in Fig. 20E), and reticulate pattern is preserved on two of the carapaces as highly reflective polygons (Re in Fig. 20E). Next to the frontal carapaces is a nearly complete frontal appendage, which has 10 podomeres with elongated ventral spines projecting from the second to fifth (VS2 to VS5 in Fig. 20E). The sixth podomere has no ventral spine, and the seventh (VS7 in Fig. 20E) and eighth (VS8 in Fig. 20F) have a short, straight ventral spine. The appendage terminates in a single terminal spine (TS in Fig. 20E). Posterior of the frontal appendage and next to the frontal carapaces, there is a large area with interlayered setal structures and smooth cuticle (Fig. 20E). This region of the Hurdia body is highly distorted, and no clear body outline can be determined. On the outer side there are seven rectangular lateral flaps arranged in a series (rL1 to rL7 in Fig. 20E), some of which have clear strengthening rays (SR in Fig. 20E). Bands of setal structures, with lanceolate blades preserved as a series of raised ridges, are associated with the bases of swimming flaps 4 to 6 (rS4–6 in Fig. 20E). On the 775 other side of the body next to the frontal carapaces, the four most posterior swimming flaps are visible as rectangular cuticular fragments (lL4–lL7 in Fig. 20E), with strengthening rays on the first and third most posterior lobes (SR7 and SR5 on Fig. 20E). The two swimming flaps anterior of those are highly distorted (lL2–3 in Fig. 20E), and the swimming flap for the first body segment is missing. Setal structures are preserved in association with the second and third swimming flaps on this side (lS2 and lS3 in Fig. 20E). General anatomy A description of the general anatomy of Hurdia combines information provided by both the isolated individual body elements and the assemblages, and is largely unchanged from that shown in the previous Hurdia reconstruction (Daley et al. 2009, fig. 3). The orientation of the two P-elements and one H-element that make up the frontal carapace can be reconstructed from their relative positions in disarticulated and articulated assemblages (Fig. 22A–C, E). P-elements are often paired and joined at the narrow beak regions, either laying flat out in dorsoventral specimens (Figs 9C, H, 22A, E), or overlying each other with the beaks pointing in the same direction when preserved laterally (Figs 9G, 22B). These two elements were joined at the narrow beak ends, as opposed to along their dorsal margin as originally suggested by Rolfe (1962). The Helement is either oriented with its pointed end directed towards the joined beaks of the P-elements (Fig. 22A, E), or lies parallel to the dorsal margins of overlying paired P-elements with the pointed end directed in the same direction as the narrow beaks of the P-elements (Figs 18A, B, 22C). In life, the frontal carapace complex attached to the anterior of the body, with the H-element located dorsally and the two P-elements positioned laterally on either side of it. The pointed end of the H-element and the joinedbeak ends of the P-elements were directed anteriorly, and the dorsal margins of the P-elements were overlain by the lateral margins of the H-element. The single notches in the dorsoposterior corner of both P-elements were lined up with the two posterior notches of the H-element, creating a space through which the pair of stalked eyes protrudes. The eyes are large, oval and smooth, situated on thick annulated stalks. A pair of frontal appendages was presumably attached to the ventral surface of the cephalon, but lack of a ventrally preserved specimen prevents a more complete description of its position. In disarticulated assemblages, paired frontal appendages are often found in close association with mouthparts (Figs 2H, 21C, 22A, D), making it likely that these elements were situated close together on the ventral surface, possibly connected by tissues. In articulated assemblages with both mouthparts and frontal appendages preserved, the frontal appendages are situated slightly anterior of the mouthparts (Figs 3, 16, 19A, B, 21C, D). In one dorsal impression (Fig. 16) the mouthparts appear to be Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 776 A. C. Daley et al. Figure 22. Disarticulated assemblages of sclerotized Hurdia body elements from the Burgess Shale. A, H-element oriented with pointed end directed towards the joined beaks of two P-elements, and mouthparts in lateral aspect flanked by two frontal appendages, under low-angle incident lighting directed from bottom right, ROM 59255. B, H-element oriented with pointed end towards the joined beaks of two P-elements, in close association with mouthparts with extra teeth and two frontal appendages, under low-angle polarized lighting from the top right, ROM 60018. C, H-element oriented parallel to the dorsal margin of a P-element, under low-angle incident lighting from bottom right, ROM 61491. D, Mouthpart with extra teeth in close association with two frontal appendages, underwater and polarized lighting directed from above, ROM 60035. E, H-element oriented with pointed end oriented towards jointed beaks of two P-elements, under polarized lighting from top right, ROM 60024. Localities: RQ (B–E) and EZ (A). Scale bars 10 mm. For abbreviations see Appendix. oriented parallel to the ventral surface, as opposed to facing backwards. The mouthparts in one laterally preserved specimen (Fig. 3) are directed towards the posterior; however, this specimen has likely been subjected to decomposition and torsion so the relative positions of elements may be distorted. Lateral and posterior to the mouthparts and frontal appendages are paired setal structures that are approximately half the length of the diameter of the mouthpart outer plates. ROM 59252 (Fig. 16) and ROM 49930 (Fig. 17C, D) have three pairs of anterior setal structures and ROM 60012 (Fig. 21C, D) has two pairs, but four are visi- ble in USNM 274155 and USNM 274158 (Fig. 3). Since the anterior region of the latter specimen has been disrupted, as indicted by the posteriorly direct mouthparts, Hurdia likely had three pairs of anterior setal structures. The body segments posterior of the cephalic region consist of smooth cuticle and associated setal structures. The number of body segments is inconsistent between specimens, ranging between six (ROM 60012, USNM 274155 and USNM 274158) and 10 (ROM 60017), and there is no significant correlation between the length of the body and the number of segments (Pearson correlation, R = 0.16, p = 0.73, N = 8). The inconsistency in the number of Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia body segments is most likely due to poor preservation. The posterior trunk shows no obvious dorsoventral flattening, such as that exhibited by Anomalocaris and Peytoia, but is instead cylindrical in cross section. In dorsal aspect, it has a constant width with no distinct tapering. In lateral view, specimens have a convex dorsal curve. All swimming flaps show reverse imbrication, where any given lobe overlaps the posterior margin of the flap anterior to it. Segment boundaries and details are often difficult to discern in articulated assemblages. It is clear that all articulated assemblages have prominent bands of setal blades covering most of the surface area of the posterior trunk region, particularly visible in USNM 274155 and USNM 274158 (Fig. 3A, B), ROM 49930 (Fig. 17D) and ROM 60017 (Fig. 19A, B). The segments have regions of smooth cuticle underlying the setae, and in specimens such as ROM 59254 (L1 and L2 in Fig. 17B) and ROM 49930 (L1–L6 in Fig. 17D) this smooth cuticle extends laterally and ventrally into small, triangular flaps. The swimming flaps may exhibit thin, dark lines running parallel along the length of the triangular extension, particularly clear in ROM 59320 (SR in Fig. 18B), ROM 60010 (SR in Fig. 21B) and ROM 60029 (SR in Fig. 20E). These are similar to the ‘strengthening rays’ seen in other anomalocaridids (Whittington & Briggs 1985). Some specimens, however, completely lack the triangular extensions of their lateral flaps (Figs 3, 19A, B, 21E, F), possibly due to decay and/or angle of burial. The posterior region is reconstructed as a cylindrical body from which a series of smooth swimming flaps bearing setal structures protrude (Fig. 23). Swimming flaps start at nearly the dorsal midline and extend down to the ventral midline (Fig. 23A) at an oblique angle that imparts an anteriorly directed tilt to the flap (Fig. 23B). The height of the swimming flap is short near the dorsal midline and gradually lengthens to a rounded triangular point in the ventral region of the body. Swimming flaps are situated close enough that the anterior margin of any lobe overlies the posterior margin of the lobe more anterior to it. The setal blades attach to the anterior margin of the dorsal surface of any given flap, with their posteriorly directed, free-hanging ends overlying the lateral flap. Setal structures found in isolation often have a width that tapers at one end (Fig. 15A), the result of conforming to the shape of the triangular lateral flaps. The posterior trunk region terminates in a small, blunt segment that does not bear any setae. Two small lobeshaped outgrowths protrude from the terminal segment, but no obvious tail fan is visible. The posterior termination is usually poorly preserved in articulated assemblages. Spence Shale Hurdia specimens Frontal appendages from the Spence Shale Formation are well preserved and easily identifiable as typical Hurdia appendages (Fig. 24), indistinguishable from the Burgess Shale specimens. Three specimens show partial, isolated frontal appendages (Fig. 24B–D), but one has a pair of 777 Figure 23. Reconstruction of the posterior region of the Hurdia body with swimming flaps and setal structures. A, Lateral view of posterior region of body showing the orientation of swimming flaps, curving down the side of the body and angled towards the anterior, with lanceolate blades of the setae situated on the dorsal surface. B, Cross section through the middle of the posterior region of the Hurdia body, in a horizontal plane relative to the dorsoventral body orientation. Setal blades are attached to the anterior margin of the anteriorly directed swimming flaps, and free-hanging towards the posterior. appendages in close association with the mouthparts (Fig. 24A). The mouthpart (M in Fig. 24A) is only partially preserved, with just the central opening and some of the inner ends of the lateral plates visible, but extra rows of teeth within the central opening are also present (X in Fig. 24A). Two rows of teeth parallel to one side of the central opening, and one row on the opposite side are preserved as black scalloped lines. Discussion Global distribution of Hurdia Most Hurdia material is derived from the Burgess Shale and nearby localities, however, isolated carapaces, mouthparts and/or frontal appendages are also known from the Wheeler Formation (Briggs et al. 2008) and the Spence Shale Formation (this paper) in Utah, the Jince Formation in the Czech Republic (Chlupáč & Kordule 2002), possibly the Shuijingtuo Formation in West Hubei, China (Cui & Huo 1990), and the Lower Ordovician Fezouata Formation of Morocco (Van Roy & Briggs 2011). In North America, the Spence Shale is located in the Wellsville Mountains of Utah, and is slightly older than the Burgess Shale, implying that the presence of Hurdia in the localities on Fossil Ridge, Mount Stephen and Stanley Glacier can be continued down Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 778 A. C. Daley et al. Figure 24. Hurdia sp. appendages from the Spence Shale Member in Utah (Miners Hollow locality). A, Pair of appendages with mouthparts showing extra rows of teeth within the central opening, ROM 59633. B, Ventral spines of a frontal appendage, ROM 59650. C, Complete frontal appendage, ROM 59634. D, Mostly complete frontal appendage with clear ventral spines, ROM 59651. Scale bars 5 mm. For abbreviations see Appendix. into the lower part of Cambrian Series 3 (Robison 1991; Garson et al. 2012). Hurdia material from the Middle Cambrian (Series 3, Stage 5) Jince Formation includes a P-element (originally designated as Proboscicaris hospes, but synonymized with Hurdia herein) and an H-element (O. Fatka, pers. comm.). Another possible H-element was identified by Chlupáč & Kordule (2002) as Helmetia? fastigata, and consists of three partially preserved carapaces with an elongated point at one end, but all other margins are incomplete. The presence of paired marginal spines on either side of the elongated point prompted Chlupáč & Kordule (2002) to consider it to be a helmetiid, but the specimens could also represent the H-element of Hurdia, in particular a specimen similar in morphology to that of H. dentata, which is herein synonymized with H. victoria. Specimens from the Shuijingtuo Formation (Cambrian Series 2, Stage 3) of West Hubei, China (Cui & Huo 1990) may also represent Hurdia. Liantuoia inflata Cui & Huo, 1990 and Huangshandongia yichangensis Cui & Huo, 1990 are carapaces similar in shape to the P-element, but distinguished from it by being smaller, thinner and smooth (Cui & Huo 1990). Liantuoia has an elongated beak, similar to that which characterizes the P-element, but Huangshandongia does not. These taxa are typically 5–10 mm in length and could represent juvenile forms of this carapace element, with the lack of reticulate pattern owing to it not being well developed and/or preserved in such young specimens. These specimens would be the oldest examples of Hurdia. The youngest examples of Hurdia body elements are from the Early Ordovician Fezouata Formation of Morocco (Van Roy & Briggs 2011), and include a P-element, an Helement and possible frontal appendages. The P-element (Van Roy & Briggs 2011, figs 1d, S4a) is elongated and thin, with a well-defined posterior notch below a thin elongated extension, and a long beak terminating in a point. This Moroccan specimen is most comparable to P-elements from the Burgess Shale that plot in the upper region of the RDA scatterplot of EFA coefficients from the outline analysis conducted herein (Fig. 7B), which generally indicates specimens that have undergone compression and distortion. The H-element from the Fezouata Formation (Van Roy & Briggs 2011, figs 1e–i, S4b, c) has an incomplete outline so is Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia impossible to assign to either Hurdia species, but has a reticulate pattern (Van Roy & Briggs 2011, fig. 1g) with similar scale and morphology to that seen in the Burgess Shale specimens (Fig. 9). The Ordovician H-element possesses both coarse and fine tubercles (Van Roy & Briggs 2011, fig. 1h, i), neither of which have been described from Burgess Shale material. Two frontal appendages are also described from the Fezouata Formation, both of which have elongated ventral spines but do not exactly match the morphology of either Hurdia or Peytoia appendages as described herein. One of these specimens (Van Roy & Briggs 2011, figs 1l, S4f) has the same number of podomeres, the characteristic curved ventral spines, and the robust auxiliary spines of Hurdia appendages, but has elongated and inclined dorsal spines more similar to Amplectobelua symbrachiata (Hou et al., 1995, figs 14, 15) or Anomalocaris (e.g. Briggs 1979, fig. 3, text-fig. 10). The second specimen (Van Roy & Briggs 2011, fig. S3c, d) has normal Hurdia appendage dorsal and ventral spines, but 11 podomeres as opposed to the usual nine. An isolated setal structure described by Van Roy & Briggs (2011, figs 1j, S4d) is similar in morphology to isolated setal structures described herein (Fig. 15), but the Ordovician specimen has been attributed to Peytoia based on comparison with articulated body specimens (Van Roy & Briggs 2011, figs 1a, b, S2a, b, S3a, b). Comparisons with Peytoia and Anomalocaris In a phylogenetic analysis of stem-group arthropods, Hurdia was found to be the sister taxon to a clade composed of Peytoia nathorsti and Anomalocaris canadensis (Daley et al. 2009). Peytoia and Anomalocaris have the same number of body and cephalic segments, mouthparts without extra rows of teeth, and large flaps (as opposed to the truncated lateral flaps of Hurdia). Peytoia and Hurdia uniquely share a frontal appendage with elongated ventral spines, mouthparts with 32 plates, and a frontal carapace complex, suggesting that these features may be plesiomorphic for the clade and that Anomalocaris is the most derived member of the group. Detailed morphological descriptions of whole-body specimens of Anomalocaris are limited to the single GSC specimen originally described by Whittington & Briggs (1985), and several specimens in the ROM collection are only partially described (Collins 1996). These specimens show that the mouthparts of Anomalocaris are not a typical 32-plate tetraradial oral cone, but consist of three large plates with a variable number of medium and smaller plates (Daley & Bergström 2012). As further details of the morphology of Anomalocaris have yet to be described, Hurdia is more easily compared to the several whole-body USNM specimens of Peytoia originally described by Whittington & Briggs (1985). Cephalic region. As discussed above, the ‘Morph A’ appendage, previously attributed to Hurdia (Daley et al. 2009) likely belongs to Peytoia nathorsti (Figs 13, 14A–C). 779 This appendage is distinguished from the Hurdia appendage by the robust nature of its 11 large podomeres, the five straight ventral spines and the presence of lateral spines. A second species of Peytoia may be represented by an appendage from the S7 locality that has some similarities to the Peytioa nathorsti appendage (Daley & Budd 2010). This appendage has 11 robust podomeres, but six straight ventral spines, an additional thin seventh ventral spine, and pronounced terminal spines that curve over the end of the appendage. Collins (1996) distinguished the mouthparts of Peytoia as rectangular in outline with a square central opening, with long spines or teeth on the four large plates. Anomalocaris mouthparts are round in outline and have only three large plates, which also have large spines (Daley & Bergström 2012). The general shape of the Hurdia mouthparts is also circular (Figs 2A, B, I, J, 9A), and, like Peytoia and Anomalocaris, the mouthparts of Hurdia have the largest teeth on the four large outer plates (Fig. 2B). Both mouthpart outline and the number of teeth were probably affected by taphonomic factors, such as the angle of burial and compression, and the teeth on the outer plates are often overlapping and poorly preserved, making their exact morphology difficult to describe. Morphometric outline or landmark analysis would be required to test for quantitative variations between the different mouthparts among anomalocaridids. The presence of inner rows of teeth within the central opening, which is not seen in Peytoia or Anomalocaris, seems to be a diagnostic feature of Hurdia mouthparts. The prominent frontal carapace complex of Hurdia is a structure not found in other anomalocaridids, or in fact in any modern or fossil arthropod. In Peytoia, thin strips of cuticular material along the anterior and lateral margins of the head are visible in some specimens preserved in ventral aspect (Fig. 14E, H), and may indicate a head shield was present in this genus as well. These cuticular elements extend posteriorly only as far as the dorsolateral eyes, as in Hurdia, but do not extend anteriorly from the head to the same extent as those of Hurdia. The purported flexibility of the cephalic region in Anomalocaris (Collins 1996) makes it unlikely to have possessed such a structure, the secondary loss of which is a derived feature of this taxon. Anomalocaris saron Hou et al., 1995 from the Chengjiang biota of China possesses a broad pre-oral plate (Chen et al. 1994, 2007; Daley et al. 2009) that may represent a hypostome, and both Anomalocaris and Peytoia from the Burgess Shale exhibit a relatively small arcuate sclerite as the most anterior structure of their head (Collins 1996, fig. 4.3; Daley et al. 2009). The latter may be homologous to the eyebearing cephalic sclerites identified in upper stem-group arthropods, such as Fuxianhuia Hou, 1987 (Budd 2008). The lack of ventrally preserved specimens of Hurdia make the identification of a hypostome or anterior sclerite difficult, although small pieces of featureless carapace are often found near the head region of articulated assemblages and Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 780 A. C. Daley et al. in disarticulated assemblages (e.g. C1 in Fig. 18B). Despite their close association with the eyes, the large frontal carapace complex of Hurdia is not thought to be homologous to the anterior sclerites of these upper stem-group arthropods because of its unique nature and dorsolateral position. The paired anterior setal structures in Hurdia, located lateral and posterior to the mouthparts and appendages, have a lanceolate blade structure similar to that on the trunk of the body of both Hurdia and Peytoia, but are much smaller in size and in some cases not obviously separated by smooth cuticle material or swimming flaps. Peytoia and Anomalocaris both have three pairs of relatively small swimming flaps and associated setae as the most posterior structures in the head (Whittington & Briggs 1985). These may be homologous to the anterior setal pairs found in Hurdia, although the associated swimming flaps are either absent or highly reduced. Posterior region. The swimming flaps (lateral lobes) and associated setal structures have long been an enigmatic aspect of the morphology of both the anomalocaridids and the closely allied taxon Opabinia Walcott, 1912 (Whittington 1974). The setal structures of Peytoia were originally described as a series of lamellae enclosed in a chamber beneath the dorsal covering of the body (Whittington & Briggs 1985), and those of Opabinia as imbricated sheets of folded structures on the dorsal surfaces of the lateral lobes. Bergström (1986, 1987) first pointed out the similarities between Opabinia and Peytoia, suggesting that the setal structures of both were dorsal coverings of lanceolate blades that were attached along one margin and interlayered with imbricated body plates. In Peytoia, the lanceolate blades were suggested to arise from mineralized strips of annulated transverse rods that run across the central region of the Peytoia body (Bergström 1986, 1987), although these structures have also been described as stiff supports that extended into the swimming flaps to help their undulatory movements (Collins 1996). Evidence for the lanceolate blade structure morphology of the setae in Opabinia was presented by Budd (1996) and Budd & Daley (2011), who reconstructed the attachment of the setae to the dorsal surfaces of the lateral lobes. Anomalocaridids from the Chengjiang fauna, including Anomalocaris saron, were also described as having sheets of lanceolate blades or scales along the dorsal surface of the body (Hou et al. 1995), and setal structures with similar morphology are present in whole-body specimens of Amplectobelua symbrachiata Hou et al., 1995 (described as ‘new anomalocaridid animal 2’ in Chen et al. 1994), as well as in close association with a disarticulated appendage of the Burgess Shale taxon Amplectobelua stephenensis Daley & Budd, 2010. Setal structures are exceptionally well preserved in Hurdia, and their morphology matches that originally described by Bergström (1986, 1987) based on the setal structures of Opabinia. Individual lanceolate blades are readily seen in several articulated assemblage specimens (Figs 3, 19C, D, 21C, D), and they are also preserved in isolation (Fig. 15). Specimens of Peytoia show that the setal structures in this taxon have the same morphology (Daley et al. 2009, fig. S2I). The conclusion that the mineralized transverse rods in Peytoia are the point of origin for the lanceolate blades is based on the similar morphology of the tips of the lanceolate blades in Hurdia setal blades and the annulations of the transverse rods in Peytoia (Daley et al. 2009, figs S2D–F, I). The setal morphology of Anomalocaris canadensis from the Burgess Shale awaits description but is thought to be similar, based on the description of the GSC specimen (Whittington & Briggs 1985) and comparison with the Chengjiang taxon Anomalocaris saron (Hou et al. 1995). The body trunk of the anomalocaridids, in particular Hurdia, remains difficult to reconstruct. Wide swimming flaps, sometimes called lateral lobes, are readily apparent in Anomalocaris and Peytoia and consist of imbricated triangular cuticular elements protruding from the ventral or lateral surface of the body, with parallel lines, interpreted as strengthening rays, running across them (Whittington & Briggs 1985; Bergström 1986, 1987; Collins 1996). Setae and the mineralized rods from which they originate are confined to the central region of the body, so the original reconstruction of Peytoia (Whittington & Briggs 1985) had a rounded central region with internal setae and an unsegmented outer covering of smooth cuticle, and swimming flaps extending from the lateral margins of the body. Bergström (1986, 1987) removed the dorsal covering in his reconstruction, which instead had a segmented central region bearing smooth cuticle alternating with the bands of lanceolate blades, and separate swimming lobes consisting of a series of imbricated plates extending across the ventral surface of the body and protruding laterally. Collins (1996) added swimming flap supports to the ventral surface of the Whittington & Briggs (1985) reconstruction, reflecting his interpretation of the mineralized lateral rods, and reconstructed Anomalocaris canadensis for the first time, based on new specimens briefly described from ROM material. The Anomalocaris body consists of an unsegmented and dorsoventrally flattened central region with the imbricated lobes extending outwards laterally. The species description mentions that this taxon has setae, but they are not included in the reconstruction or discussed in the text, so confirmation of their morphology in Anomalocaris requires more detailed descriptions of whole-body specimens. Two ‘oblique compressions’ of Anomalocaris nathorsti were described by Whittington & Briggs (1985), one of which is herein reassigned to Hurdia based on the presence of a frontal carapace structure (Fig. 3). The other, USNM 274154 with the counterpart in two pieces, USNM 274156 and USNM 274161 (Fig. 14D), is the posterior region of an anomalocaridid body that completely lacks the head region. The lack of a wide tail fan suggests that it is not Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia Anomalocaris. Preservation of setae is similar to that of Hurdia, as is its rounded body shape, however the obvious and well-preserved swimming flaps are more typical of Peytoia. The specimen has been rotated along its length, such that the posterior portion is roughly dorsoventrally preserved and the anterior region is preserved in lateral aspect. The specimen is unique in having obvious smooth layers of cuticle, or tergites, between the setae, a feature not clearly observed in other anomalocaridid specimens, and which Whittington & Briggs (1985) chose to omit in their reconstruction. Bergström (1986, 1987), however, considered the morphology of this specimen to be one of the main lines of evidence supporting a reconstruction of the Peytoia body with separate dorsal and ventral sclerites. The posterior body morphology of Hurdia is unique from the other Burgess Shale anomalocaridids in that the swimming flaps are not prominent features in any specimen, and are not visible at all in several of the best-preserved whole-body specimens. Both dorsal specimens (Figs 16, 21A, B) have compact bodies that are no wider than the width of the H-element of the cephalic carapace, and the only evidence of swimming flaps are small, triangular fragments of cuticle, in some cases with parallel lines that could be the strengthening rays observed (SR in Figs 18B, 20E, 21B). Bands of setal structures are paired along the lateral side of the body, but do not extend uninterrupted over the dorsal midline. Specimens preserved laterally, e.g. ROM 59254 (Fig. 17A, B), ROM 60012 (Fig. 21C, D) and ROM 49930 (Fig. 17C, D), show a series of small triangular structures extending from the ventral region of the body, but the overlap direction and attachment to the body is unclear. These triangular swimming flaps are typically in parallel with lateral setae that extend, in some cases, onto the proximal region of the lobe (S1 in Fig. 17D; S3 in Fig. 21D). In specimens preserved obliquely, e.g. ROM 59320 (Fig. 18), deformed swimming flaps are ventral and prominent loops of setal structures in the dorsal and lateral region of the body are unattached to any other body part. Specimens ROM 60017 (Fig. 19A, B), USNM 274159 (Fig. 3) and ROM 42985 (Fig. 21E, G) show no evidence of swimming flap structures, with their posterior regions being dominated by setal blades that in some areas overlie each other with little evidence of intervening smooth cuticle. Thus, unlike Peytoia and Anomalocaris, where the wide swimming flaps are a dominant features of the posterior body and the setae are less visible, Hurdia has wide bands of setal blades that are the best preserved and most complete features of the body trunk, while the swimming flaps are reduced (Daley et al. 2009, fig. 3). Despite the reduced swimming flaps, the posterior body reconstruction for Hurdia is comparable to the reconstruction of Peytoia described by Bergström (1986, 1987). The dorsal and lateral region of the body is not an unsegmented covering, but is divided into a series of alternating smooth cuticles (‘tergites’) and lanceolate blade structures. In Hurdia, the 781 latter do not pass completely over the dorsal margin, as is reconstructed for Peytoia (Bergström 1986, 1987), and it lacks ventral tergites (or swimming flaps), which Bergström (1986, 1987) interpreted as separate structures from the dorsal tergites. Since a ventral view of Hurdia is lacking, and the dorsal views do not exhibit the distinct margins between ventral swimming flaps and dorsal setae and tergites seen in Peytoia, it is more parsimonious to interpret the swimming flaps as triangular ventral extensions of the dorsolateral setae-bearing tergites for this taxon. Comparison with Amiella Amiella ornata Walcott, 1911a is known only from the single, incomplete holotype specimen (Fig. 14F, G), described as closely related to Sidneyia inexpectans Walcott, 1911a. Simonetta (1963) attempted to synonymize Amiella with Sidneyia through the identification of a second specimen very similar to Amiella. Bruton (1981) confirmed the identification of this second specimen as Sidneyia, and argued that the holotype of Amiella was an oblique and telescoped example of the same genus. An anomalocaridid affinity for the holotype specimen was suggested by both Whittington & Briggs (1985, p. 606), who considered it to be an “unusually oriented compression” of Peytoia, and Hou et al. (1995), who also suggested it was related to Peytoia. The species Amiella prisca Mansuy, 1912 from Yunnan was considered by Hou & Bergström (1997) as more similar to Fuxianhuia Hou, 1987 or Chengjiangocaris Hou & Bergström, 1991 from Chengjiang. These authors restricted Amiella ornata to the type specimen, which they considered as a nomen dubium. Examination of the holotype of Amiella ornata essentially confirms the description of Whittington & Briggs (1985), as a series of overlapping plates, some of which are smooth but have prominent wrinkles (the “outer layer” of Whittington & Briggs 1985, p. 605) and the rest of which lack wrinkles (the “inner layer”). We further show, as indicated in the camera lucida drawing of Hou et al. (1995), that the ‘inner layer’ of plates consists of a series of closely spaced ridges, similar to those seen in ROM 60038, 60029 and 59254, herein interpreted as setal blades (Fig. 14G). Based on the presence of setae, and faint rays on some of the smooth cuticles (Whittington & Briggs 1985), it seems most likely that this specimen represents a portion of the posterior body of an anomalocaridid, probably an obliquely preserved specimen of Hurdia. It is difficult to make this identification with any certainty when the cephalic region is lacking, but its morphology is more similar to Hurdia than to any other anomalocaridid. Anomalocaridid species diversity and temporal distribution Continued collecting efforts in the Burgess Shale have revealed that the diversity of anomalocaridids was higher Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 782 A. C. Daley et al. than previously described, and have suggested that morphology was variable between taxa in the group. Hurdia is the most common anomalocaridid at many of the Burgess Shale localities (Daley et al. 2009) and its description helps clarify enigmatic features of the morphology of Anomalocaris and Peytoia, especially with regards to the frontal appendages and setae. The diversity of anomalocaridids from the Burgess Shale was further increased by the description of a new anomalocaridid, Stanleycaris hirpex Caron et al., 2010 from the Stanley Glacier site in the Canadian Rocky Mountains, about 40 km south-east of the type Burgess Shale localities (Caron et al. 2010), and two other new taxa based on isolated appendage material from the Burgess Shale localities on Fossil Ridge and Mount Stephen (Daley & Budd 2010). Caryosyntrips serratus Daley & Budd, 2010 is unique to the Burgess Shale, while Amplectobelua stephenensis Daley & Budd, 2010 represents the first example of this genus found outside of the Chengjiang fauna in China. Anomalocaris saron is also found in the Chengjiang fauna, giving Amplectobelua and Anomalocaris a wide global distribution, particularly the latter, which is also found in the Kinzers Formation of Pennsylvania, USA (Anomalocaris pennsylvanica Resser, 1929) and the Emu Bay Shale in Australia (Anomalocaris briggsi Nedin, 1995). In terms of anomalocaridid diversity and abundance, Hurdia is more common in the Cambrian Series 3, while Anomalocaris is known from the Cambrian Series 2 and 3. Hurdia and Anomalocaris are the two longest-lived taxa of anomalocaridids. A rarefaction analysis of anomalocaridid appendages (Daley & Budd 2010) suggests that the diversity of anomalocaridids at the Burgess Shale localities decreased with time, such that the oldest locality S7 has the highest diversity of anomalocaridid taxa of any site in the world, including Peytoia, Anomalocaris, Hurdia, Amplectobelua and Caryosyntrips. The Chengjiang fauna, which is approximately 10 million years older than the Burgess Shale, also has a relatively high diversity of anomalocaridids from several different localities, consisting of at least Amplectobelua symbrachiata and Anomalocaris saron, plus another possible species of Anomalocaris and the possibly related taxa Cucumericrus decoratus Hou et al., 1995 and Parapeytoia yunnanensis Hou et al., 1995 (Chen et al. 1994; Hou et al. 1995). However, the Sirius Passet biota of Greenland (Conway Morris et al. 1987; Babcock & Peel 2007), which is at least as old as Chengjiang, if not older (Conway Morris & Peel 2010; Budd 2011), has only one possible anomalocaridid taxon (Daley & Peel 2010), Tamisiocaris borealis Daley & Peel, 2010, in addition to other stemgroup euarthropods unique to this locality, such as the gilled lobopodians Kerygmachela Budd, 1993 and Pambdelurion Budd, 1998. Anomalocaris pennsylvanica Resser, 1929 from the USA, and Anomalocaris briggsi Nedin, 1995 from Australia are anomalocaridids found at sites older than the Burgess Shale (Cambrian, Series 2, Stage 3). Hurdia is also described from the Wheeler Shale (Briggs et al. 2008), which is younger than the Burgess Shale (Conway Morris 1992) and the Spence Shale Member (this publication), which is slightly older than the Burgess Shale (Briggs et al. 2008). Anomalocaridids in the Lower Ordovician Fezouata Formation may consist of new species of Hurdia and/or Peytoia, and other new taxa (Van Roy & Briggs 2011), all of which are larger in size than their Cambrian counterparts. Schinderhannes bartelsi Kühl et al., 2009, a stemlineage arthropod with some anomalocaridid features, has been described from the Lower Devonian Hunsrück Slate, Germany. Ecology Like other anomalocaridids, Hurdia was a presumed predator and/or scavenger. Its relatively large size, prominent pair of large eyes, highly toothed mouthparts and frontal appendages suggest that it fed on mobile prey items. Visual predation in the anomalocaridids was recently corroborated by the discovery of putative Anomalocaris compound eyes from the Emu Bay Shale of Australia, which have a visual surface with at least 16,000 ommatidial lenses (Paterson et al. 2011). Appendages with long ventral spines, as seen in Hurdia and Peytoia, were likely used as a rigid net apparatus to sweep through the water column (Whittington & Briggs 1985) or sift through the flocculent layer of seafloor sediment (Daley & Budd 2010) to capture prey, which were then transferred to the mouthparts for ingestion. The high diversity of anomalocaridid frontal appendage morphologies in the Burgess Shale is interpreted to be the result of these contemporaneous predators attempting to relieve competition for food sources by employing different feeding strategies (Daley & Budd 2010). The mouthparts of Hurdia, with their extra rows of teeth, may have functioned differently to those of Anomalocaris and Peytoia, with the former relying on the rows of extra teeth within the buccal cavity to masticate food items and move them further into the digestive tract, and the latter using suction to bring food to their mouthparts and ingesting prey items, such as newly moulted trilobites, whole or with moderate slicing (Rudkin 1979, 2009; Nedin 1999). Unlike its fully nektonic sister taxa Anomalocaris and Peytoia, Hurdia was probably nekto-benthic. The reduced swimming flaps, relatively small tail fan and cylindrical body shape suggest that Hurdia would not have been as efficient a swimmer as the more streamlined Anomalocaris and Peytoia. The large eyes of all anomalocaridids suggest visual predation, so Hurdia must have been seeking prey on the seafloor or in the water column. Hurdia was likely living at the sediment–water interface, resting on the seafloor at times and swimming to acquire prey. Such a lifestyle may have necessitated the large presumably respiratory setal blades seen in Hurdia because of the increased need for oxygen required to sustain an ambush predation strategy. If Hurdia was nekto-benthic, its frontal Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia carapace structure could have been involved in feeding, since the frontal appendages and mouthparts are located near the base of where the frontal carapaces attach to the body. It may have served to funnel potential prey towards the mouthparts and frontal appendages. Like other fossil and modern arthropods, Hurdia presumably moulted throughout its life, but direct evidence for moulting behaviour is rare in the Burgess Shale (Garcı́aBellido & Collins 2004). The evidence for moulting in Hurdia is the large number of disarticulated assemblages with numerous closely associated body elements. It is unknown if Hurdia moulted its harder body elements, such as the frontal carapace complex, appendages, mouthparts and setal structures, consecutively, or if it moulted its entire body at once. The large number of disarticulated assemblages consisting only of the more sclerotized body elements listed above might suggest that these were being moulted as individual elements. No evidence of gut glands or traces of the alimentary canal have been observed in any Hurdia whole-body specimens, such that no single specimen can be definitely identified as a carcass as opposed to a moult. However, as the most complete articulated assemblages show little evidence of disruption or dislocation of body elements, they are best interpreted as carcasses. Some of the more disrupted articulated assemblages and relatively complete disarticulated assemblages could represent moults preserved immediately after being discarded or carcasses that have been scavenged or partially decomposed. The reticulate pattern on the H- and P-elements of Hurdia (Fig. 9) is of potential ecological significance. Reticulation is relatively common in both modern and fossil arthropods. Two categories of arthropod reticulation have been described (Rolfe 1962): large mesh, which in practice describes a pattern where reticulate cell diameter ≥2 mm (Lieberman 2003), and small mesh, where reticulate cell diameter is <1 mm. The reticulate pattern of the H- and P-elements belongs to the former category. Reticulate patterns are also found in other Cambrian arthropods from Burgess Shale deposits, such as Tuzoia Walcott, 1912 (Whittington 1985; Lieberman 2003; Vannier et al. 2007), Perspicaris recondita Briggs, 1977, Isoxys Walcott, 1890 (Vannier & Chen 2000), Retifacies Hou et al., 1989 (Hou & Bergström 1997; Hou et al. 2004) and the olenelloid trilobite Wanneria walcottana Wanner, 1901 (Lieberman 1999). The Ri values for both H- and P-elements (0.03–0.11 and 0.01–0.09, respectively) is at the small end of the ranges observed for ostracods (0.09–0.28) and Tuzoia (0.03–0.4) (Vannier et al. 2007). In modern ostracods, reticules arise at the boundaries between underlying epidermal cells, which are added in succession to increase the size of the carapace, making the trait homoplastic (Okada 1981). In the specimens studied here, reticulations are unlikely to represent individual cells because of their size (Lieberman 2003; Vannier et al. 2007). As was suggested for Tuzoia (Vannier 783 et al. 2007), the reticulate pattern of the frontal carapace of Hurdia may have been an optimal way of increasing the strength of the carapace while minimizing its overall weight and the amount of new carapace material that must be generated at each moulting stage. This type of carapace is well adapted for a pelagic lifestyle (Vannier et al. 2007). Summary and conclusions Hurdia is the most common anomalocaridid in the Burgess Shale, and the wealth of specimens provides a good basis for detailed morphological description. These numerous specimens provide insight into enigmatic features of the anomalocaridids, such as the appendages, setal blades and swimming flaps, and present a unique challenge to understanding its systematics owing to the disarticulated nature of the material. A combined approach utilizing morphometric statistics, stratigraphical data and detailed morphological observations has produced a comprehensive summary of the known anatomy and systematics of Hurdia. Two species distinguished by the shape of the H-element of their frontal carapace coexist in most Burgess Shale sites in variable relative abundances. H. victoria, with its elongated H-element, is found most abundantly in localities WQ and RQ on Fossil Ridge, while H. triangulata is more abundant at the older site (S7 – Tulip Beds) and the three youngest localities EZ, UE and StanG, indicating that these morphs do not represent sexual dimorphs but actual species. A previous phylogenetic analysis suggested that Hurdia is the sister taxon to a group composed of Anomalocaris and Peytoia (Daley et al. 2009), with the entire clade located in the basal stem lineage to Euarthropoda. Clarification of the enigmatic morphological features of this clade thus allows for a better understanding of arthropod evolution. The morphology of Hurdia differs from Anomalocaris and Peytoia in details of its frontal appendage and mouthparts, as well as the arrangement of its swimming flaps and setal structures. The presence of setae and lateral flaps, presumably used for swimming, is plesiomorphic for the entire anomalocaridid clade, but the Hurdia lineage emphasizes the presence of larger setal structures, while the Anomalocaris–Peytoia lineage favours wider lateral flaps. These differences may reflect different modes of life, mirrored also in the different morphologies of frontal appendages and cephalic carapaces. Hurdia is interpreted as a nektobenthic predator or scavenger, searching out slow-moving prey items from near the sediment–water interface, whereas the larger Anomalocaris was better adapted for swimming in the water column. The application of morphometric statistical methods had some success with Burgess Shale fossils. Simple carapaces are ideal for outline analysis, especially when they contain some degree of symmetry, allowing highly deformed specimens to be readily identified and excluded from the analyses. Such quantitative analyses are worthwhile methods to explore when attempting to characterize Burgess Shale 784 A. C. Daley et al. taxa, especially when combined with detailed stratigraphical information. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Acknowledgements We thank J. Bergström, D. Collins and G. D. Edgecombe for discussions. Comments from M. Streng and reviewers D. Briggs and B. S. Lieberman improved this manuscript. F. Collier and J. Dougherty provided access to specimens at the MCZ and GSC, respectively. D. Erwin, J. Thompson and M. Florence provided access to specimens at the USNM, and P. Fenton provided support at the ROM. Burgess Shale specimens were collected under several Parks Canada Research and Collecting permits delivered to the ROM (D. Collins: 1975 to 2000; J-BC: 2008). J. Loxton, M. Streng, R. Gaines and Parks Canada staff provided fieldwork support at Stanley Glacier in 2008. Funding was provided by the European Union Marie Curie Research and Training Network ZOONET (grant MRTN-CCT-2004–005624) to GEB and ACD, the Swedish Research Council (Vetenskapsrådet) to GEB, and Natural Sciences and Engineering Research Council of Canada Discovery grant to J-BC (#341944). Hurdia specimens from Utah were purchased through The Louise Hawley Stone Charitable Trust (ROM). This is Royal Ontario Museum Burgess Shale Research Project number #8B. Supplementary material Supplementary material is available 10.1080/14772019.2012.732723 online DOI: References Aitken, J. D. 1997. Stratigraphy of the Middle Cambrian platformal succession, Southern Rocky Mountains. Geological Survey of Canada Bulletin, 398, 1–322. Aitken, J. D. & McIlreath, I. A. 1984. The Cathedral Reef Escarpment: a Cambrian great wall with humble origins. Geos, 13, 17–19. Babcock, L. E. & Peel, J. S. 2007. Palaeobiology, taphonomy, and stratigraphic significance of the trilobite Buenellus from the Sirius Passet Biota, Cambrian of North Greenland. Memoirs of the Association of Australasian Palaeontologists, 34, 401–418. Bengtson, S. 2000. Teasing fossils out of shales with cameras and computer. Palaeontologia Electronica, 3, 14 pp. Bergström, J. 1986. Opabinia and Anomalocaris, unique Cambrian arthropods. Lethaia, 19, 241–246. Bergström, J. 1987. The Cambrian Opabinia and Anomalocaris. Lethaia, 20, 187–188. Briggs, D. E. G. 1977. Bivalved arthropods from the Cambrian Burgess Shale of British Columbia. Palaeontology, 20, 67–72. Briggs, D. E. G. 1979. Anomalocaris, the largest known Cambrian arthropod. Palaeontology, 22, 631–663. Briggs, D. E. G. & Mount, J. D. 1982. The occurrence of the giant arthropod Anomalocaris in the Lower Cambrian of Southern- California, and the overall distribution of the genus. Journal of Paleontology, 56, 1112–1118. Briggs, D. E. G. & Robison, R. A. 1984. Exceptionally preserved nontrilobite arthropods and Anomalocaris from the Middle Cambrian of Utah. University of Kansas Paleontological Contributions, 111, 1–23. Briggs, D. E. G., Lieberman, B. S., Hendricks, J. R., Halgedahl, S. L. & Jarrard, R. D. 2008. Middle Cambrian arthropods from Utah. Journal of Paleontology, 82, 238–254. Bruton, D. L. 1981. The arthropod Sidneyia inexpectans, Middle Cambrian, Burgess Shale, British Columbia. Philosophical Transactions of the Royal Society of London, Series B, 295, 619–653. Budd, G. 1993. A Cambrian gilled lobopod from Greenland. Nature, 364, 709–711. Budd, G. E. 1996. The morphology of Opabinia regalis and the reconstruction of the arthropod stem-group. Lethaia, 29, 1–14. Budd, G. E. 1998. Stem group arthropods from the Lower Cambrian Sirius Passet fauna of North Greenland. Pp. 125–138 in R. A. Fortey & Thomas R. H. (eds) Arthropod Relationships. Systematics Association Special Volume, Chapman and Hall, London. Budd, G. E. 2008. Head structures in upper stem-group euarthropods. Palaeontology, 51, 561–573. Budd, G. E. 2011. Campanamuta mantonae gen. et sp. nov., an exceptionally preserved arthropod from the Sirius Passet fauna (Buen Formation, Lower Cambrian, North Greenland). Journal of Systematic Palaeontology, 9, 217–260. Budd, G. E. & Daley, A. C. 2011. The lobes and lobopods of Opabinia regalis from the middle Cambrian Burgess Shale. Lethaia, 45, 83–95. Butterfield, N. J., Balthasar, U. & Wilson, L. A. 2007. Fossil diagenesis in the Burgess Shale. Palaeontology, 50, 537–543. Caron J.-B., Gaines R. R., Mángano G., Streng M. & Daley A. C. 2010. A new Burgess Shale-type assemblage from the “thin” Stephen Formation of the Southern Canadian Rockies. Geology, 38, 811–814. Chen, J. Y., Ramsköld, L. & Zhou, G. Q. 1994. Evidence for monophyly and arthropod affinity of Cambrian giant predators. Science, 264, 1304–1308. Chen, J. Y., Waloszek, D., Maas, A., Braun, A., Huang, D., Wang, X. Q. & Stein, M. 2007. Early Cambrian Yangtze Plate Maotianshan Shale macrofauna biodiversity and the evolution of predation. Palaeogeography, Palaeoclimatology, Palaeoecology, 254, 250–272. Chlupáč, I. & Kordule, V. 2002. Arthropods of Burgess Shale type from the Middle Cambrian of Bohemia (Czech Republic). Bulletin of the Czech Geological Survey, 77, 167–182. Collins, D. 1996. The “evolution” of Anomalocaris and its classification in the arthropod class Dinocarida (nov.) and order Radiodonta (nov.). Journal of Paleontology, 70, 280–293. Collins, D. 1999. Dinocarids: the first monster predators on Earth. Rotunda, 32, 24–29. Collins, D. & Stewart, W. D. 1991. The Burgess Shale and its environmental setting, Fossil Ridge, Yoho National Park. Pp. 105–117 in P. L. Smith (ed.) A Field Guide to the Palaeontology of Southwestern Canada. University of British Columbia, Victoria. Collins, D., Briggs, D. E. G. & Conway Morris, S. 1983. New Burgess Shale fossil sites reveal Middle Cambrian faunal complex. Science, 222, 163–167. Conway Morris, S. 1986. The community structure of the Middle Cambrian Phyllopod Bed (Burgess Shale). Palaeontology, 29, 423–467. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia Conway Morris, S. 1992. Burgess Shale-type faunas in the context of the ‘Cambrian explosion’: a review. Journal of the Geological Society, London, 149, 631–636. Conway Morris, S. & Peel, J. S. 2010. New palaeoscolecidan worms from the Lower Cambrian: Sirius Passet, Latham Shale and Kinzers Shale. Acta Palaeontologica Polonica, 55, 141–156. Conway Morris, S. & Robison, R. A. 1982. The enigmatic medusoid Peytoia and a comparison of some Cambrian biotas. Journal of Paleontology, 56, 116–122. Conway Morris, S. & Robison, R. A. 1988. More soft-bodied animals and algae from the Middle Cambrian of Utah and British Columbia. The University of Kansas Paleontological Contributions, 122, 1–48. Conway Morris, S., Peel, J. S., Higgins, A. K., Soper, N. J. & Davis, N. C. 1987. A Burgess Shale-like fauna from the Lower Cambrian of North Greenland. Nature, 326, 181– 183. Cui, Z. L. & Huo S. C. 1990. New discoveries of Lower Cambrian crustacean fossils from Western Hubei. Acta Palaeontologica Sinica, 29, 321–330. Daley, A. C. & Bergström, J. 2012. The oral cone of Anomalocaris is not a classic “peytoia”. Naturwissenschaften, 99, 501–504. Daley, A. C. & Budd, G. E. 2010. New anomalocaridid appendages from the Burgess Shale, Canada. Palaeontology, 53, 721–738. Daley, A. C. & Peel, J. S. 2010. A possible anomalocaridid from the Cambrian Sirius Passet Lagerstätte, North Greenland. Journal of Paleontology, 84, 352–355. Daley, A. C., Budd, G. E., Caron, J.-B., Edgecombe, G. D. & Collins, D. 2009. The Burgess Shale anomalocaridid Hurdia and its significance for early euarthropod evolution. Science, 323, 1597–1600. Ferson, S., Rohlf, F. J. & Koehn, R. K. 1985. Measuring shape variation in two-dimensional outlines. Systematic Zoology, 34, 59–68. Fletcher, T. P. 2011. The development of mid-Cambrian lithostratigraphical nomenclature in the vicinity of the Burgess Shale, Kicking Horse belt of Canada. Palaeontographica Canadiana, 31, 39–68. Fletcher, T. P. & Collins, D. 1998. The Middle Cambrian Burgess Shale and its relationship to the Stephen Formation in the southern Canadian Rocky Mountains. Canadian Journal of Earth Sciences, 35, 413–436. Fletcher, T. P. & Collins, D. 2003. The Burgess Shale and associated Cambrian formations west of the Fossil Gully Fault Zone on Mouth Stephen, British Columbia. Canadian Journal of Earth Sciences, 40, 1823–1838. Gabbott, S. E., Zalasiewicz, J. & Collins, D. 2008. Sedimentation of the Phyllopod Bed within the Cambrian Burgess Shale Formation of British Columbia. Journal of the Geological Society, 165, 307–318. Gaines, R. R. 2011. A new Burgess Shale-type locality in the “thin” Stephen Formation, Kootenay National Park, British Columbia: stratigraphic and paleoenvironmental setting. Palaeontographica Canadiana, 31, 73–88. Garcı́a-Bellido, D. & Collins, D. H. 2004. Moulting arthropod caught in the act. Nature, 429, 40. Garson, D. E., Gaines, R. R., Droser, M. L., Liddell, W. D. & Sappenfield, A. 2012. Dynamic palaeoredox and exceptional preservation in the Cambrian Spence Shale of Utah. Lethaia, 45, 164–177. Hill, M. O. & Gauch, H. G. Jr. 1980. Detrended Correspondence Analysis: an improved ordination technique. Vegetatio, 42, 47–58. 785 Hou, X.-G. 1987. Three large new arthropods from the Lower Cambrian Chengjiang, eastern Yunnan. Acta Palaeontologica Sinica, 26, 272–285. Hou, X.-G. & Bergström, J. 1991. The arthropods of the Lower Cambrian Chengjiang fauna, with relationships and evolutionary significance. Pp. 179–187 in A. M. Simonetta & S. Conway Morris (eds) The Early Evolution of Metazoa and the Significance of Problematic Taxa. Cambridge University Press, Cambridge. Hou, X.-G. & Bergström, J. 1997. Arthropods of the Lower Cambrian Chengjiang fauna, southwest China. Fossils & Strata, 45, 1–116. Hou, X.-G. & Bergström, J. 2006. Dinocaridids: anomalous arthropods or arthropod-like worms? Pp. 139–158, 847–850 in J.-Y. Rong, Z.-J. Fang, Z.-H. Zhou, R.-B. Zhan, X.-D. Wang, & Z.-L. Yuan (eds) Originations, Radiations and Biodiversity Changes – evidences from the Chinese fossil record. Science Press, Beijing. Hou, X.-G., Chen, J.-Y. & Lu, H.-Z. 1989. Early Cambrian new arthropods from Chengjiang, Yunnan. Acta Palaeontologica Sinica, 28, 42–57. Hou, X., Bergström, J. & Ahlberg, P. 1995. Anomalocaris and other large animals in the Lower Cambrian Chengjiang Fauna of southwest China. GFF, 117, 163–183. Hou, X.-G., Aldridge, R. J., Bergström, J., Siveter, David J., Siveter, Derek J. & Feng, X.-H. 2004. The Cambrian fossils of Chengjiang, China. Blackwell Publishing, Malden, Massachusetts, 233 pp. Johnston, P. A., Johnston, K. J., Collom, C. J., Powell, W. G. & Pollock, R. J. 2009. Palaeontology and depositional environments of ancient brine seeps in the Middle Cambrian Burgess Shale at The Monach, British Columbia, Canada. Palaeogeography, Palaeoclimatology, Palaeoecology, 277, 86–105. Kuhl, F. P. & Giardina, C. R. 1982. Elliptic Fourier features of a closed contour. Computer Graphics and Image Processing, 18, 236–258. Kühl, G., Briggs, D. E. G. & Rust, J. 2009. A great-appendage arthropod with a radial mouth from the Lower Devonian Hunsrück Slate, Germany. Science, 323, 771–773. Lankester, E. R. 1904. The structure and classification of the Arthropoda. Quarterly Journal of Microscopical Science, 47, 523–582. Liddell, W. D., Wright, S. H. & Brett, C. E. 1997. Sequence stratigraphy and paleoecology of the Middle Cambrian Spence Shale in Northern Utah and Southern Idaho. Brigham Young University Geology Studies, 42, 59–78. Lieberman, B. S. 1999. Systematic revision of the Olenelloidea (Trilobita, Cambrian). Bulletin of the Yale University Peabody Museum of Natural History, 45, 1–150. Lieberman, B. S. 2003. A new soft-bodied fauna: The Pioche Formation of Nevada. Journal of Paleontology, 77, 674– 690. Mansuy, H. 1912. Etude géologique du Yunnan oriental. Part 2. Paléontologie. Mémoires du Service géologique de l’Indochine, 1, 1–146. Nedin, C. 1995. The Emu Bay Shale, a Lower Cambrian fossil Lagerstätten, Kangaroo Island, South Australia. Memoirs of the Association of Australasian Palaeontologists, 18, 31– 40. Nedin, C. 1999. Anomalocaris predation on nonmineralized and mineralized trilobites. Geology, 27, 987–990. O’Brien, L. J. & Caron, J.-B. 2012. A new stalked filter-feeder from the Middle Cambrian Burgess Shale, British Columbia, Canada. PLoS ONE, 7, e29233. Okada, Y. 1981. Development of cell arrangement in ostracod carapaces. Paleobiology, 7, 276–281. Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 786 A. C. Daley et al. Paterson, J. R., Garcia-Bellido, D. C., Lee, M. S. Y., Brock, G. A., Jago, J. B. & Edgecombe, G. D. 2011. Acute vision in the giant Cambrian predator Anomalocaris and the origin of compound eyes. Nature, 480, 237–240. Resser, C. E. 1929. New Lower and Middle Cambrian Crustacea. Proceedings of the United States National Museum, 76, 1– 18. Rigby, J. K. & Collins, D. 2004. Sponges of the Middle Cambrian Burgess Shale and Stephen Formations, British Columbia. Royal Ontario Museum Contributions in Science, 1, 1–155. Robison, R. A. 1976. Middle Cambrian trilobite biostratigraphy of the Great Basin. Brigham Young University Geology Studies, 23, 93–109. Robison, R. A. 1991. Middle Cambrian biotic diversity: Examples from four Utah Lagerstätten. Pp. 77–98 in A. Simonetta & S. Conway Morris (eds) The Early Evolution of Metazoa and the Significance of Problematic Taxa. Cambridge University Press, Cambridge. Robison, R. A. & Richards, B. C. 1981. Larger bivalve arthropods from the Middle Cambrian of Utah. The University of Kansas Paleontological Contributions, 106, 1–28. Rohlf, F. J. 2004. tpsDig version 1.40. Department of Ecology and Evolution, State University, Stony Brook, New York [updated at http://life.bio.sunysb.edu/morph/], accessed 6 December 2006. Rolfe, W. D. I. 1962. Two new arthropod carapaces from the Burgess Shale (Middle Cambrian) of Canada. Breviora, 160, 1–9. Rudkin, D. M. 1979. Healed injuries in Ogygopsis klotzi (Trilobita) from the Middle Cambrian of British Columbia. Royal Ontario Museum, Life Sciences Occasional Paper, 32, 1–8. Rudkin, D. M. 2009. The Mount Stephen Trilobite Beds. Pp. 90–102 in J.-B. Caron & D. M. Rudkin (eds) A Burgess Shale Primer – History, Geology and Research Highlights. The Burgess Shale Consortium, Toronto. Simonetta, A. M. 1963. Osservazioni sugli Artropodi non Trilobiti della Burgess Shale (Cambriano medio). II Contributo: i generi Sidneyia ed Amiella Walcott 1911. Monitore Zoologico Italiano, 70–71, 97–108. Simonetta, A. M. & Delle Cave, L. 1975. The Cambrian non trilobite arthropods from the Burgess Shale of British Columbia. A study of their comparative morphology, taxonomy and evolutionary significance. Palaeontographica italica, 69(N.S. 39), 1–37. Slice, D. E. 1998. Morpheus et al.: software for morphometric research. Revision 01–01-00. Department of Ecology and Evolution, State University, Stony Brook, New York [updated at http://life.bio.sunysb.edu/morph/, accessed 8 October 2008]. Suzuki, Y. & Bergström, J. 2008. Respiration in trilobites: a reevaluation. GFF, 130, 211–229. Suzuki, Y., Kondo, A. & Bergström, J. 2008. Morphological requirements in limulid and decapod gills: A case study in deducing the function of lamellipedian exopod lamellae. Acta Palaeontologica Polonica, 53, 275–283. ter Braak, C. J. F. & Šmilauer, P. 2002. Canoco for Windows version 4.5. Biometrics – plate research international, Wageningen, The Netherlands. Vannier, J. & Chen, J.-Y. 2000. The Early Cambrian colonization of pelagic niches exemplified by Isoxys. Lethaia, 33, 295– 311. Vannier, J., Caron, J. B., Yuan, J. L., Briggs, D. E. G., Collins, D., Zhao, Y. L. & Zhu, M. Y. 2007. Tuzoia: morphology and lifestyle of a large bivalved arthropod of the Cambrian seas. Journal of Paleontology, 81, 445–471. Van Roy, P. & Briggs, D. E. G. 2011. A giant Ordovician anomalocaridid. Nature, 473, 510–513. Walcott, C. D. 1890. The fauna of the Lower Cambrian or Olenellus Zone. Reports of the U.S. Geological Survey, 10, 509–763. Walcott, C. D. 1911a. Cambrian geology and paleontology II. Middle Cambrian Merostomata. Smithsonian Miscellaneous Collections, 57, 17–40. Walcott, C. D. 1911b. Cambrian geology and paleontology II. Middle Cambrian holothurians and medusae. Smithsonian Miscellaneous Collections, 57, 41–68. Walcott, C. D. 1912. Middle Cambrian Branchiopoda, Malacostraca, Trilobita and Merostomata. Smithsonian Miscellaneous Collections, 57, 145–228. Walton, J. 1936. On the factors which influence the external form of fossil plants: with descriptions of the foliage of some species of the Palaeozoic equisetalean genus Annularia Sternberg. Philosophical Transactions of the Royal Society of London, Series B, 226, 219–237. Wanner, A. 1901. A new species of Olenellus from the Lower Cambrian of York County. Pennsylvania. Proceedings of the Washington Academy of Sciences, 3, 267–272. Whiteaves, J. F. 1892. Description of a new genus and species of phyllocarid Crustacea from the Middle Cambrian of Mount Stephen, B.C. Canadian Record of Science, 5, 205–208. Whittington, H. B. 1974. The enigmatic animal Opabinia regalis, Middle Cambrian Burgess Shale, British Columbia. Philosophical Transactions of the Royal Society of London, Series B, 271, 1–43. Whittington, H. B. 1985. The Burgess Shale. Yale University Press, New Haven, Connecticut, 151 pp. Whittington, H. B. & Briggs, D. E. G. 1985. The largest Cambrian animal, Anomalocaris, Burgess Shale, BritishColumbia. Philosophical Transactions of the Royal Society of London, Series B, 309, 569–609. Appendix. Abbreviations used in text and figures A: Compression wrinkles indicating previous relief An: Anterior setal structure AS: Auxiliary spines of frontal appendage AW: Talus material from the Walcott Quarry on Fossil Ridge in Yoho National Park, probably originating from RQ, UE or EZ B: Body segment Bu: Burrow C: Unidentified smooth carapace CZ: Jince Formation of the Czech Republic D: Dark fossil of unknown identity superimposed on Hurdia DS: Dorsal spine of frontal appendage E: Eye ES: Eye stalk EZ: Ehmaniella Zone, lower Collins Quarry on Fossil Ridge in Yoho National Park f : F-value F: Frontal appendage GPT: Monte Carlo Global Permutation Test GSC: Geological Survey of Canada Downloaded by [the Bodleian Libraries of the University of Oxford] at 02:08 14 November 2013 Morphology and systematics of Hurdia H: H-element L: Swimming flaps l: Left (used as a prefix when indicating setae, swimming flaps, etc. on the left side of the body) Le: Leanchoilia LS: Lateral spines of frontal appendage M: Mouthparts MCZ: Harvard University Museum of Comparative Zoology N: Number of specimens p: P-value of probability P: P-element Pl: Large outer plates of mouthpart Po: Podomere of frontal appendage r: Right (used as a prefix when indicating setae, swimming flaps, etc. on the right side of the body) R: Correlation coefficient Rd: Ridge of sediment that disrupts the fossil specimen RDA: Redundancy analysis Re: Reticulate pattern on H- and P-element Ri: Reticulate ratio between the surface area of the largest polygon and the valve length ROM: Royal Ontario Museum, Canada 787 RQ: Raymond Quarry on Fossil Ridge in Yoho National Park RT: Raymond Quarry talus on Fossil Ridge in Yoho National Park, probably originating from RQ S: Setal structures SD: Standard deviation Sm: Smooth carapace of posterior body region SP: Smallest outer plate of mouthpart SR: Strengthening rays on swimming flaps S7: Tulip Beds on Mount Stephen in Yoho National Park StanG: Stanley Glacier locality in Kootenay National Park T: Tail lobe TS: Terminal spine of frontal appendage UE: Upper Ehmaniella Zone, upper Collins Quarry on Fossil Ridge in Yoho National Park USNM: National Museum of Natural History, Smithsonian Institution VK: Jince Formation of the Czech Republic (type locality of Proboscicaris hospes) VS: Ventral spine of frontal appendage WQ: Walcott Quarry, Greater Phyllopod Bed on Fossil Ridge in Yoho National Park X: Extra rows of teeth within central opening of mouthpart