Exceptionally preserved crustaceans from western Canada reveal a cryptic Cambrian radiation

Thomas H. P. Harveya,1, Maria I. Vélezb, and Nicholas J. Butterfielda

aDepartment of Earth Sciences, University of Cambridge, Cambridge CB2 3EQ, United Kingdom; and bDepartment of Geology, University of Regina, Regina, SK, Canada S4S 0A2

Edited by Steven M. Stanley, University of Hawaii, Honolulu, HI, and approved December 16, 2011 (received for review September 16, 2011) The early history of crustaceans is obscured by strong biases in fossil Geological Context preservation, but a previously overlooked taphonomic mode yields The Deadwood Formation (broadly defined, to include the Earlie important complementary insights. Here we describe diverse crus- and Finnegan formations) encompasses a broad expanse of shal- tacean appendages of Middle and Late Cambrian age from shallow- low-marine, Middle to Late Cambrian sandstones and mudstones marine mudstones of the Deadwood Formation in western Canada. extending through eastern parts of the Western Canada Sedi- The fossils occur as flattened and fragmentary carbonaceous cuticles mentary Basin, the Williston Basin, and into the Black Hills of but provide a suite of phylogenetic and ecological data by virtue of South Dakota, its type locality (15, 16). In Canada, the formation their detailed preservation. In addition to an unprecedented range occurs primarily in the subsurface, with all of the specimens in this fi of complex, largely articulated ltering limbs, we identify at least study recovered from petroleum exploration drillcores in south- four distinct types of mandible. Together, these fossils provide the west Saskatchewan and southeast Alberta. Unoxidized mudstones earliest evidence for crown-group branchiopods and total-group from Ceepee Riley Lake 3-4-39-13W3 and Ceepee Reward 4-28- copepods and ostracods, extending the respective ranges of these 38-24W3 (Middle/Late Cambrian, Saskatchewan) (16) and Rio clades back from the Devonian, Pennsylvanian, and Ordovician. De- Bravo Ronald 1-6-38-15W4 (Late Cambrian, Alberta) (15) were tailed similarities with living forms demonstrate the early origins gently dissolved in hydrofluoric acid and the isolated SCFs in- and subsequent conservation of various complex food-handling dividually collected from the rinsed residues (see Materials and adaptations, including a directional mandibular asymmetry that Methods and SI Text for details of sample distributions and age). has persisted through half a billion years of evolution. At the same Among the several thousand recovered specimens are significant time, the Deadwood fossils indicate profound secular changes in subpopulations of cuticle fragments that bear distinctively ar- crustacean ecology in terms of body size and environmental distri- thropodan spines and setae, including an exceptionally rich di- bution. The earliest radiation of crustaceans is largely cryptic in the versity of crustacean body parts. fossil record, but “small carbonaceous fossils” reveal organisms of surprisingly modern aspect operating in an unfamiliar biosphere. Fossil Description and Identification The Deadwood crustaceans are distinguished from other ar- arthropods | phylogeny | taphonomy | Paleozoic thropodan remains by diagnostic cuticular ornamentations. They come from nine samples representing three separate assemblages, rustaceans are the dominant arthropods in the modern marine one from each drillcore (Table S1). Mandibles are the most widely Crealm and are renowned for their diversity, disparity, com- distributed elements and fall into four distinct categories: bran- plexity, and ecologic range (1, 2). Their fossil record, however, is chiopod-type, copepod-type, ostracod-type, and an unidentified heavily skewed toward biomineralizing post-Cambrian forms (3), morphology. Other crustacean remains include comparatively obscuring the higher-level relationships of crustaceans and their delicate arrays of spines and setae, which are generally less

terrestrial mandibulate relatives, the myriapods and hexapods (4). abundant and informative, although one sample horizon has EVOLUTION Nonmineralizing (pan)crustaceans have been documented in the yielded a rich assemblage of extensively articulated branchiopod- Cambrian fossil record but, until recently, have been represented type limbs. almost exclusively by “Orsten-type” taxa of minute body size (< 2 Branchiopod-Type Mandibles. The first of two types of mandible mm) and limited appendage differentiation (5, 6). In contrast, the from the Riley Lake assemblage is distinguished by an extensive, larger-bodied crustacean-like forms preserved in Burgess Shale- D-shaped grinding (molar) surface (n =17)(Fig.1A–H). The type and other macroscopic assemblages are either assignable to specimens fall into at least three distinct “morphotypes” that ap- much deeper phylogenetic positions (1, 6, 7), or have yet to reveal pear to be independent of both size and preservational orienta-

key diagnostic characters among the inner leg branches and tion/resolution. In the first morphotype (n =6)(Fig.1A–D), scaly EARTH, ATMOSPHERIC, mouthparts (8, 9). Notably, the only macroscopic Cambrian fossil lineations extend across the width of the molar surface, forming AND PLANETARY SCIENCES to exhibit convincing mandibles (“jaws”) is a Late Cambrian deep ridges at the straight/concave margin and a protruding fringe euthycarcinoid, a probable stem-group mandibulate (10). (sometimes also strong teeth) along the opposite edge (Fig. 1B). Despite this limited record, the identification of disarticulated The second morphotype (n =2)(Fig.1E and F)isdistinguished but unambiguously crustacean body parts among small carbona- by its opposite polarity (which is evident once images have been ceous fossils (SCFs) (11) in the Early Cambrian Mount Cap For- corrected for the “way-up” of slide-mounted specimens) and mation of NW Canada (12, 13) points to a cryptic but significant by lineations that do not extend across the width of the molar diversity of Cambrian crustaceans. Here we describe extensive SCF assemblages of exceptionally preserved filtering appendages and Author contributions: T.H.P.H., M.I.V., and N.J.B. performed research; T.H.P.H. analyzed mouthparts (mandibles) from the Middle and Upper Cambrian data; and T.H.P.H. and N.J.B. wrote the paper. ∼ Deadwood Formation of western Canada ( 488 to 510 Ma; The authors declare no conflict of interest. — Cambrian Series 3 Furongian) (14). By bridging a major tapho- This article is a PNAS Direct Submission. nomic gap in body size and preservational resolution, the Dead- 1To whom correspondence should be addressed. E-mail: [email protected]. wood fossils provide crucial phylogenetic and ecologic datapoints This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. for charting a major Cambrian radiation of crustaceans. 1073/pnas.1115244109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1115244109 PNAS | January 31, 2012 | vol. 109 | no. 5 | 1589–1594 Downloaded by guest on October 1, 2021 Fig. 1. Fossil crustacean mandibles from the Middle and Late Cambrian Deadwood Formation. (A–H) Branchiopod-type mandibles from the Riley Lake assemblage. Morphotypes one (A–D) and two (E and F) are interpreted as the right and left mandibles from a single taxon, and morphotype three (G and H) as a distinct form. See Fig. S1 for detailed images of A, E, and F.(I–O) Copepod-type mandibles from the Riley Lake assemblage; detail I′ shows the platform and dorsal seta. (P) An ostracod-type mandible from the Rio Bravo Ronald assemblage; detail P′ magnifies the gnathal edge. Images have been reversed from slide-orientation in C, E, F, and H to show true polarity, and in J, K, N, and O for purposes of comparison. Grains of diagenetic pyrite show as opaque objects. See Table S2 for specimen numbers. (Scale bar, 50 μmforA–P;30μmforI′ and P′.)

surface, but become confluent with an unornamented region various extant anostracan branchiopods (Fig. 2 A and B), which bounded by marginal nodes (Fig. S1). The third molar morpho- suggests that they come from a single taxon displaying a complex type (n =3)(Fig.1G and H) features a region with disconnected, pattern of mandibular asymmetry adapted for enhanced food- poorly aligned scales and no discrete bounding margin. In all three grinding efficiency (18, 21, 25). A comparable pattern of continu- morphotypes the mandibular profile, as far as it is preserved, ous scale rows on the right molar vs. a smooth region adjacent to appears to be similar: one or more long setae and a single stout dorsal marginal nodes on the left is a recognized synapomorphy spine are inserted in line with the more acute end of the molar (see character 15 in ref. 24) of extant anostracans and Lepidocaris, surface, beyond which the mandibular margin curves away form- a stem-anostracan from the Devonian Rhynie Chert (24, 25). The ing a pronounced “shoulder” (Fig. 1 A, C–E, G,andH). third fossil morphotype is sufficiently distinct to represent a sepa- Mandibles with extensive, scaly molar surfaces are known from rate—although still branchiopodan—taxon (18). among hexapods and myriapods as well as branchiopods, mala- Overall, the Deadwood molars range up to at least 230 μmlong, costracans, and remipedes (17). However, in both overall shape predicting a maximum body length of at least 10–15 mm based on and detailed ornamentation the fossil molars are conspicuously scaling relationships in extant anostracans (see figure S3 in ref. 13). similar to those of branchiopod crustaceans (Fig. 2 A and B). The presence in the first and second morphotypes of a moderately The pronounced posterior “shoulder” is characteristic of the sized posterior tooth and an asymmetric “tooth-groove” system post-molar profile in branchiopod mandibles (21, 22), and confirms points to an ecology of mixed benthic scraping and suspension that a distinct incisor process was absent during life. This condition feeding, as opposed to more exclusive predation or suspension is shared with branchiopods crown-wards of Rehbachiella (23), feeding (21). a Cambrian stem-group form (see character 14 in ref. 24). More- over, the first and second fossil morphotypes show striking simi- Copepod-Type Mandibles. A contrasting type of mandible from the larities to the right- and left-handed mandibles, respectively, of Riley Lake assemblage occurs as cuticle fragments bearing arc-

1590 | www.pnas.org/cgi/doi/10.1073/pnas.1115244109 Harvey et al. Downloaded by guest on October 1, 2021 In contrast, comparisons with mandibles in other crustacean groups appear to be superficial. Certain cirripedes possess a tooth row that ends in a protrusive bristly region, although the teeth are fewer and more robust and there is no projecting seta (30). The series of cusped teeth found in some branchiopods are either re- stricted to very early ontogenetic stages [e.g., in anostracans (31)], or are much broader and closely packed, and unaccompanied by terminal platforms or setae [in notostracans and laevicaudatans (22, 32)]. Among the fossil Orsten-type crustaceans, the mandibles of Skara and Bredocaris are broadly similar in profile but do not exhibit bifurcated tooth cusps, platforms, or dorsal setae (33, 34), whereas Rehbachiella is distinguished by the disproportionate ex- pansion through ontogeny of a flattened grinding region (23). The specific similarities to copepod mandibles allow predictions of body size and diet in the Deadwood species. Scaling relationships between gnathal edge and body length in various living copepods (35, 36) predict a prosome length of around 4.5–7mm(andabody length ∼1–2 mm more) for the largest intact fossil (gnathal edge length ∼270 μm) (Fig. 1J). The fragmentary remains of larger mandibles (Fig. 1 N and O) point to still larger individuals, possibly in the centimetric range. A correlation between diet and mandib- ular morphology is well-established for living planktic calanoid copepods [Itoh’s “Edge Index” (37)]. Comparisons with the similar adaptations seen in the fossil taxon, notwithstanding its compara- tively large body size and unknown planktic or benthic habit, predict Fig. 2. Mandibles from modern crustaceans. (A and B) Gnathal edges a largely herbivorous diet for the larger specimens with compara- (molars) from the right and left mandibles, respectively, of Chirocephalus tively robust teeth, and a more omnivorous diet for the smaller diaphanus (Branchiopoda: Anostraca); reprinted with permission from specimens with elongate cusps, a possible ontogenetic distinction. ref. 18 (copyright 1991, Koninklijke Brill NV). Labels indicate anterior (A), posterior (P), dorsal (D), and ventral (V); a and b indicate matched Ostracod-Type Mandible. The third type of Deadwood mandible is opposing regions. (C) Gnathal edge of Calanus propinquus (Copepoda; represented by a single specimen from the Rio Bravo Ronald cranial side of female right mandible; image reversed); arrow indicates borehole that uniquely preserves the entire proximal mandibular dorsal seta. Image courtesy of Jan Michels. (D) Coxa with articulated palp body (coxa) along with its intact gnathal edge (Fig. 1P). The coxa of Macropyxis alanlordi (Ostracoda: Podocopa: Macrocyprididae). Image courtesy of Simone Nunes Brandão (19). (E) Mandibular (coxal) gnathal exhibits an elongate overall shape that narrows to an acute apex, edge of Danielopolina exuma (Ostracoda: Myodocopa: Halocypridina); a large proximal opening (the insertion point in life for soft tis- redrawn from ref. 20. (Scale bars, 100 μmforA, B,andD;50μmforC; E is sues), and a palp foramen (for the attachment of more distal parts, not drawn to scale.) which have not been preserved). The gnathal edge is particularly complex: it bears a raised toothed blade (or possibly two super- imposed blades) adjacent to three long setae set back from the shaped arrays of up to six robust teeth (n = 32) (Fig. 1 I–O). The edge, an intermediate region with short setae alongside a series of tooth row terminates in a protruding, bristly wedge-shaped plat- toothed cusps and a stout hooked tooth, and a bristly protruding form (n = 12) (Fig. 1 I, J, L,andM), confirming that the fossils platform (Fig. 1P′). Other mandibular remains in this assemblage

represent entire gnathal edges rather than fragmentary incisors. are limited to two isolated gnathal edges that likely represent EVOLUTION Below the platform is inserted a papposerrate seta that is con- a fourth distinct type of Deadwood mandible, and are not con- spicuously longer and more robust than adjacent setae, and proj- sidered further (Fig. S2). ects in line with the tooth row (n = 5) (Fig. 1 I′ and M). Variation In both overall morphology and details of the gnathal edge, in tooth outline (from broadly conical to narrow and strongly the more complete Rio Bravo Ronald mandible compares most bicuspidate) and in the degree of secondary ornamentation (on the closely to those of ostracod crustaceans (Fig. 2 D and E). Similarly apical ridge and the basal slopes) depends in part on the angle of shaped, markedly elongate coxae with palp foramina of equivalent fossil compression, which varies from side-on to oblique or near- size and position are characteristic of both major living subgroups, fl “vertical,” but also exhibits a trend toward more robust and highly Myodocopa and Podocopa, presumably re ecting the distinctive orientation, musculature, and articulation of ostracod mandibles EARTH, ATMOSPHERIC, ornamented teeth in larger specimens. This observation aside, AND PLANETARY SCIENCES large and small specimens exhibit similar numbers of teeth and (38). The complexity and form of the gnathal edge appear to be shared in particular with halocyprid myodocopes, some of which similar relative proportions of the toothed edge and bristly plat- express a similar suite of characters including a raised toothed form, and are reasonably interpreted as ontogenetic variants of blade with adjacent long setae, an intermediate region with a a single species. hook-shaped spine, and a protruding grinding surface (Fig. 2E) Broadly comparable mandibles are widespread among crusta- fi (20, 39). The size of the fossil is consistent with an overall body ceans, but this particular combination of ne-scale elaborations (carapace) length of around 2 mm (19). (teeth, platform, and protruding seta), their numbers, positions, and proportions, and their overall ontogenetic consistency, are Branchiopod-Type Limbs. A contrasting assemblage of SCFs, from shared only with copepods—among which close matches for the asinglethin(∼5 mm) horizon in the Ceepee Reward borehole, fossils are numerous (26–29) (Fig. 2C). In particular, the prom- lacks mandibles but contains delicate setal armatures in unrivalled inent projecting seta (Fig. 1I′) is comparable in form and position abundance and degree of articulation (n = 150) (Fig. 3). Most to the potentially homologous “dorsal seta” (sometimes a pair of conspicuously, crustacean-type “filter plates” formed from a series setae) found in every major order of nonparasitic copepods [i.e., of coplanar plumose setae with intersetule distances of ∼1 μm, plus Calanoida, Cyclopoida, Platycopioida, Misophrioida, Harpacti- accessory setae, occur commonly as isolated structures (n > 45) coida and Mormonilloida (27, 29)] (Fig. 2C). (Fig. 3 A and B) and sometimes within extensive setal arrays up to

Harvey et al. PNAS | January 31, 2012 | vol. 109 | no. 5 | 1591 Downloaded by guest on October 1, 2021 Fig. 3. Fossil branchiopod-type limbs from the Middle Cambrian Deadwood Formation (Reward assemblage). (A and B) Isolated filter plates plus accessory setae; detail (A′) shows the diagnostic setulation of plumose filtering setae. (C and D) More extensive setal arrays preserving filter plates and other armatures in situ on limbs. (C) An array representing one or more appendages; detail C′ shows a filter plate (from center-right of image; rotated). (D) Part of a single appendage that preserves a filter plate (to left) and three protruding endites (arrowed); detail D′ shows the middle endite. See Table S2 for specimen numbers. (Scale bar, 60 μmforA and B;15μmforA′; 100 μmforC and D;25μmforC′; and 35 μmforD′.)

800 μm across that reveal their wider anatomical context (n ∼11) combinations of filter plates and protuberant endites are known, (Fig. 3 C and D and Fig. S3). Specimens that preserve a continuous for example, in the notostracan/diplostracan-like Castracollis from underlying cuticle are demonstrably derived from a single ap- the Devonian Rhynie Chert (24, 45). Reconstructing the Dead- pendage (Fig. 3D) and show that filter plates were borne on limbs wood fossils as a branchiopod crustacean with a long series of with a series of up to five nodose lobes (Fig. 3D′), along with filtering thoracic appendages, an overall body length of at least a diversity of contrasting armatures composed variously of pap- several millimeters is likely for the more articulated arrays, al- pose, coarse plumose and, most distinctively, bifurcating serrated though a centimetric body size is suggested by isolated filters con- (“saw-toothed”) setae (Fig. 3 C and D and Fig. S3). The absence of structed from substantially larger setae. A mixed scraping/filtering articulations between the nodose lobes identifies them as the ecology (rather than a wholly planktic mode of life) is suggested by endites of either an undivided limb stem or a poorly segmented the juxtaposition of filter plates and saw-toothed armatures (25). limb branch. Filter plates are widespread and multiply convergent structures Discussion among crustaceans, but the arcuate outlines of the Deadwood Cambrian arthropods have sometimes been “shoehorned” into examples and their arrangement on extensive lobose appendages modern clades, despite having character combinations that are shared only with the phyllopodous thoracic filters of bran- support deeper, more stem-ward phylogenetic positions (7, 46). chiopods (24). In contrast, the mouthpart filters found in certain Conversely, the Deadwood fossils risk being assigned to in- malacostracans and the filter-like structures in various ostracods appropriately derived positions because of their “modern” ap- are much larger in proportion to the overall appendage (40, 41), pearance but disarticulated condition. Therefore, we conserva- whereas the thoracic filters of euphausiacean malacostracans tively assign them to comparatively inclusive clades, identifying (“krill”) and leptostracans/phyllocarids are linear rather than ar- crown groups via a synapomorphy shared with a subset of the cuate, and are not associated with such diverse accessory arma- crown (46). tures (42–44). Among branchiopods, the Deadwood filters share To summarize, the Middle/Late Cambrian branchiopod-type a strictly coplanar setal arrangement with crown-group forms, in fossils can all be assigned to a subset of the branchiopod total- contrast to the more 3D armatures of Rehbachiella (23); similar group that excludes Rehbachiella. Furthermore, the mandibles that

1592 | www.pnas.org/cgi/doi/10.1073/pnas.1115244109 Harvey et al. Downloaded by guest on October 1, 2021 express anostracan-type right-left differentiation—a directional SCF assemblages, conceivably as indigestible remains sedimented asymmetry (47) conserved across half a billion years of evolution— via fecal pellets (26, 35, 36) or simply as biostratinomically re- can be assigned to the crown. The Deadwood fossils thus extend calcitrant seabed detritus (58). In any case, they offer clear po- the known range of crown-group branchiopods, as well as those tential for reconciling the Orsten forms with adults and larger- crown-wards of Rehbachiella, back some 80–100 Myr from the bodied relatives for a new, high-definition narrative of early Lower Devonian Rhynie Chert (48). Furthermore, filter plates and mandibulate evolution. scraping armatures that are strikingly similar to those preserved in the Deadwood assemblage occur in the Mount Cap Formation Evolving Crustacean Form and Function. The fresh taphonomic (13) (Fig. S4), extending the known range of total-group bran- perspective of SCFs provides the only direct evidence for sophis- chiopods back to the late Early Cambrian (∼510 Ma). ticated particle-handling in larger-bodied Cambrian arthropods. The Late Cambrian ostracod-type mandible can be assigned to This characteristically crustacean-type ecology at the interface the ostracod total-group and perhaps to the crown, based on the of micro- and macroscopic nutrient cycling has otherwise been halocyprid-like construction of the gnathal edge. Ostracod-type loosely inferred from overall body form (1) and the proxy record of carapaces are known from the Early Ordovician and may extend phytoplankton diversification (59). The detailed adaptations de- back to the Cambrian in the guise of particular bradoriids (49). scribed here represent the acme of Cambrian differentiation within However, the Deadwood mandible provides the only appendage- appendages, an alternative (and potentially correlative) measure based evidence for ostracods before the Silurian Herefordshire of evolving arthropod complexity to the larger-scale tagmosis that Lagerstätte, a unit that it predates by some 70 Myr (50). has been the focus of previous studies (e.g., ref. 2). The Middle/Late Cambrian copepod-type mandibles are as- In part, the new fossils reinforce a picture of early origination signed to the copepod total-group (stem or crown) based on and subsequent conservation in crustacean form and function a combination of characters including isometric growth and a dor- (60). At the same time, however, the small carbonaceous record sal seta. The Deadwood fossils thus extend the record of copepods provides evidence for unanticipated ecologic turnover. In the (broadly defined) back some 190–210 Myr from fragments ∼ modern oceans, branchiopods are represented by a just a few extracted from a Pennsylvanian ( 303 Ma) bitumen clast (51); species of small, secondarily marine cladocerans; larger forms, other pre-Holocene records are restricted to the Miocene and comparable in size to those of the Deadwood (up to ∼15 mm or Cretaceous (27). more) and Mount Cap (∼50 mm), are now entirely nonmarine Taken together, our results provide unambiguous evidence for (24). Furthermore, modern free-living copepods are almost all a substantial branching by the Late Cambrian of within-crown ∼ — much smaller than the 5- to 10-mm (plus) Deadwood taxon (27). (pan)crustacean lineages a largely cryptic component of the — fi — Cambrian “explosion”—and offer key calibrations for molecular In the modern world, visual predators especially teleost sh clocks and time-scaled phylogenies (48). drive down body size in planktic freshwater crustacean commu- nities (61) and strongly constrain the complex behaviors and Complementary Taphonomic Modes. A Cambrian radiation of crus- distribution patterns of krill (62, 63), a group that shares with the taceans is not evident in either the conventional “shelly” fossil re- Cambrian branchiopods the attributes of centimetric body size, cord or, apparently, macroscopic Burgess Shale-type biotas (1). marine habitat, and (by convergence) thoracic filtering. Signifi- However, it is revealed to a limited extent by the small-bodied (< 2 cantly, the Deadwood and Mount Cap fossils reveal a contrasting mm) forms preserved in Orsten-type assemblages (5, 6). Among pattern of crustacean distribution in the comparatively “unesca- these forms, Rehbachiella has been interpreted as a stem-bran- lated” Cambrian biosphere. chiopod (24) and others, notably Skara, Yicaris, Bredocaris,and (possibly) the metanauplius Wujicaris, as stem-group members of Materials and Methods various higher-level “entomostracan” taxa (52–55); pentastomid- Washed mudstone samples of 5–20 g were immersed in 40% hydrofluoric acid like Orsten fossils may also be crustaceans (48). Debates over the for 2–5 d before being flushed with water over a 30- or 63-μm sieve. Individual fi microfossils were picked from residues suspended in water using a pipette, phylogenetic af nities of the Orsten taxa have emphasized the EVOLUTION difficulty in interpreting larvae and miniaturized adults (56, 57), and rinsed in distilled water, and transferred to glass coverslips for mounting on glass microscope slides (using epoxy resin). Specimens were studied using it is conceivable that their apparently more plesiomorphic positions fi are an artifact of their developmental stage or smaller size (13). transmitted light microscopy, and nal images assembled from digital pho- tographs taken at different focal planes. Figured specimens are stored at the Importantly, the Deadwood fossils, like those of the Mount Cap Geological Survey of Canada (GSC), 601 Booth Street, Ottawa, ON, Canada, (13), reveal the microscopic anatomies of both micro- and mac- numbered sequentially from GSC 135369 to GSC 135393 (Table S2). roscopic (millimetric to centimetric) individuals and therefore circumvent a major taphonomic bias. That said, even the smallest ACKNOWLEDGMENTS. We thank staff at the Geological Subsurface Labora- Deadwood and Mount Cap individuals exhibit previously unseen tory, Regina, and Energy Resources Conservation Board, Calgary for help with

morphologies, perhaps because they lived in comparatively shal- core sampling; geoLOGIC for generous access to subsurface data; Pier Binda for EARTH, ATMOSPHERIC,

discussion of Deadwood microfossils; and Jan Michels, Simone Nunes Brandão, AND PLANETARY SCIENCES low-marine environments that are undersampled by both Burgess and Graziella Mura for providing images. This work is supported by Sidney Shale-type and Orsten-type preservation. Mandibles, at least, are Sussex College, Cambridge, and Natural Environment Research Council emerging as a widespread and reasonably abundant component of Grant NE/H009914/1.

1. Budd GE (2001) The Ecology of the Cambrian Radiation, eds Zhuravlev AYu, Riding R 7. Budd GE (2008) Head structure in upper stem-group euarthropods. Palaeontology 51: (Columbia University Press, New York), pp 404–427. 561–573. 2. Adamowicz SJ, Purvis A, Wills MA (2008) Increasing morphological complexity 8. Fu D, Zhang X (2011) A new arthropod Jugatacaris agilis n. gen. n. sp. from the Early in multiple parallel lineages of the Crustacea. Proc Natl Acad Sci USA 105: Cambrian Chengjiang biota, south China. J Paleontol 85:567–586. 4786–4791. 9. Collette JH, Hagadorn JW (2010) Three-dimensionally preserved arthropods from 3. Sepkoski JJ, Jr. (2000) Crustacean biodiversity through the marine fossil record. Cambrian Lagerstätten of Quebec and Wisconsin. J Paleontol 84:646–667. Contrib Zool 69:213–221. 10. Vaccari NE, Edgecombe GD, Escudero C (2004) Cambrian origins and affinities of an 4. Rota-Stabelli O, et al. (2011) A congruent solution to arthropod phylogeny: Phylo- enigmatic fossil group of arthropods. Nature 430:554–557. genomics, microRNAs and morphology support monophyletic Mandibulata. Proc Biol 11. Butterfield NJ, Harvey THP (2012) Small carbonaceous fossils (SCFs): A new measure of Sci 278:298–306. early Paleozoic paleobiology. Geology 40:71–74. 5. Maas A, et al. (2006) The ’Orsten’—More than a Cambrian Konservat-Lagerstätte 12. Butterfield NJ (1994) Burgess Shale-type fossils from a Lower Cambrian shallow-shelf yielding exceptional preservation. Palaeoworld 15:266–282. sequence in northwestern Canada. Nature 369:477–479. 6. Walossek D (1999) Crustaceans and the Biodiversity Crisis, eds Schram FR, von Vaupel 13. Harvey THP, Butterfield NJ (2008) Sophisticated particle-feeding in a large Early Klein JC (Brill, Leiden), pp 3–27. Cambrian crustacean. Nature 452:868–871.

Harvey et al. PNAS | January 31, 2012 | vol. 109 | no. 5 | 1593 Downloaded by guest on October 1, 2021 14. Ogg JG, Ogg G, Gradstein FM (2008) The Concise Geologic Time Scale (Cambridge 38. Yamada S, Matzke-Karasz R (2011) Skeleto-musculature of the mandible and its University Press, Cambridge). function in podocopid ostracodes exemplified by Loxoconcha pulchra (Cytheroidea: 15. Hein FJ, Nowlan GS (1998) Regional sedimentology, conodont biostratigraphy and Loxoconchidae) and Fabaeformiscandona tyrolensis (Cypridoidea: Candonidae). J correlation of Middle Cambrian-Lower Ordovician(?) strata of the “Finnegan” and Morphol 272:1342–1352. Deadwood formations, Alberta subsurface, Western Canada Sedimentary Basin. Bull 39. Richter S, Kornicker LS (2006) The mandibles of a halocyprid ostracode (Halocypridina: Can Pet Geol 46:166–188. Halocypridae)—A new record of mandibular gnathal edges with a “lacinia mobilis”. J 16. Dixon J (2008) Stratigraphy and facies of Cambrian to Lower Ordovician strata in Crustac Biol 26:113–118. Saskatchewan. Bull Can Pet Geol 56:93–117. 40. Cannon HG, Manton SM (1929) On the feeding mechanism of the syncarid Crustacea. 17. Edgecombe GD, Richter S, Wilson GDF (2003) The mandibular gnathal edges: Ho- Trans R Soc Edinb 56:175–189. mologous structures throughout Mandibulata? Afr Invertebr 44:115–135. 41. Cannon HG (1933) On the feeding mechanism of certain marine ostracods. Trans R Soc 18. Mura G (1991) Scanning electron microscopic study of the molar surfaces of the Edinb 57:739–764. mandibles of Chirocephalus diaphanus Prévost (Anostraca). Crustaceana 60:178–185. 42. Hamner WM (1988) Biomechanics of filter feeding in the Antarctic Krill Euphausia 19. Brandão SN (2010) Macrocyprididae (Ostracoda) from the Southern Ocean: Taxo- superba: Review of past work and new observations. J Crustac Biol 8:149–163. nomic revision, macroecological patterns, and biogeographical implications. Zool J 43. Pabst T, Scholtz G (2009) The development of phyllopodous limbs in Leptostraca and – Linn Soc 159:567 672. Branchiopoda. J Crustac Biol 29:1–12. 20. Kornicker LS, Iliffe TM (1998) Myodocopid Ostracoda (Halocypridina, Cladocopina) 44. Martin JW, Christiansen JC (1995) A morphological comparison of the phyllopodous from anchialine caves in the Bahamas, Canary Islands, and Mexico. Smithson Contrib thoracic limbs of a leptostracan (Nebalia sp.) and a spinicaudate conchostracan – Zool 599:1 93. (Leptestheria sp.), with comments on the use of Phyllopoda as a taxonomic category. 21. Mura G (1995) Morphological features of the mandible related to feeding habits of Can J Zool 73:2283–2291. – some Anostraca species. Crustaceana 68:83 102. 45. Fayers SR, Trewin NH (2003) A new crustacean from the Early Devonian Rhynie chert, 22. Richter S (2004) A comparison of the mandibular gnathal edges in branchiopod Aberdeenshire, Scotland. Trans R Soc Edinb Earth Sci 93:355–382. crustaceans: Implications for the phylogenetic position of the Laevicaudata. Zoo- 46. Budd GE, Jensen S (2000) A critical reappraisal of the fossil record of the bilaterian morphology 123:31–44. phyla. Biol Rev Camb Philos Soc 75:253–295. 23. Walossek D (1993) The Upper Cambrian Rehbachiella and the phylogeny of Bran- 47. Palmer AR (2009) Animal asymmetry. Curr Biol 19:R473–R477. chiopoda and Crustacea. Fossils and Strata 32:1–202. 48. Edgecombe GD (2010) Arthropod phylogeny: An overview from the perspectives of 24. Olesen J (2009) Phylogeny of Branchiopoda (Crustacea)—Character evolution and morphology, molecular data and the fossil record. Arthropod Struct Dev 39:74–87. contribution of uniquely preserved fossils. Arthropod Syst Phylogeny 67:3–39. 49. Williams M, Siveter DJ, Popov LE, Vannier JMC (2007) Biogeography and affinities of 25. Fryer G (1985) Structure and habits of living branchiopod crustaceans and their the bradoriid arthropods: Cosmopolitan microbenthos of the Cambrian seas. Palae- bearing on the interpretation of fossil forms. Trans R Soc Edinb 76:103–113. ogeogr Palaeoclimatol Palaeoecol 248:202–232. 26. Turner JT (1978) Scanning electron microscope investigations of feeding habits and 50. Siveter DJ (2008) Ostracods in the Palaeozoic? Senckenbergiana Lethaea 88:1–9. mouthpart structures of three species of copepods of the family Pontellidae. Bull Mar 51. Selden PA, Huys R, Stephenson MH, Heward AP, Taylor PN (2010) Crustaceans from Sci 28:487–500. bitumen clast in Carboniferous glacial diamictite extend fossil record of copepods. 27. Huys R, Boxshall G (1991) Copepod Evolution (The Ray Society, London). Nat Commun 1:50. 28. Michels J, Schnack-Schiel SB (2005) Feeding in dominant Antarctic copepods—Does 52. Walossek D, Müller KJ (1998) Arthropod Fossils and Phylogeny, ed Edgecombe GD the morphology of the mandibular gnathobases relate to diet? Mar Biol 146:483–495. (Columbia University Press, New York), pp 185–231. 29. Ivanenko VN, Defaye D (2006) Planktonic deep-water copepods of the family Mor- 53. Zhang XG, Siveter DJ, Waloszek D, Maas A (2007) An epipodite-bearing crown-group monillidae Giesbrecht, 1893 from the East Pacific Rise (13 degrees N), the north- – eastern Atlantic, and near the North Pole (Copepoda, Mormonilloida). Crustaceana crustacean from the Lower Cambrian. Nature 449:595 598. 79:707–726. 54. Zhang XG, Maas A, Haug JT, Siveter DJ, Waloszek D (2010) A eucrustacean meta- – 30. Chan BKK, Hsu C-H, Tsai P-C (2009) Morphology and distribution of the acorn bar- nauplius from the Lower Cambrian. Curr Biol 20:1075 1079. nacle Tetraclita reni nom. nov. (Crustacea: Cirripedia) in Madagascar and adjacent 55. Haug JT, Olesen J, Maas A, Waloszek D (2011) External morphology and post-em- waters. Zootaxa 2019:57–68. bryonic development of Derocheilocaris remanei (Mystacocarida) revisited, with – 31. Davoodi P, Rahimian H, Agh N (2009) SEM study of the early development of the a comparison to the Cambrian taxon Skara. J Crustac Biol 31:668 692. mandible of Artemia urmiana and A. parthenogenetica (Branchiopoda, Anostraca). 56. Wills MA (1997) Arthropod Relationships, eds Fortey RA, Thomas RH (Chapman & – Crustaceana 82:257–266. Hall, London), pp 189 209. fi 32. Fryer G, Boxshall G (2009) The feeding mechanisms of Lynceus (Crustacea: Branchio- 57. Boxshall GA (2007) Crustacean classi cation: On-going controversies and unresolved – poda: Laevicaudata), with special reference to L. simiaefacies Harding. Zool J Linn Soc problems. Zootaxa 1668:313 325. 155:513–541. 58. Van Waveren IM (1994) Chitinous palynomorphs and palynodebris representing 33. Müller KJ, Walossek D (1985) Skaracarida, a new order of Crustacea from the Upper crustacean exoskeleton remains from sediments of the Banda Sea (Indonesia). Scripta Cambrian of Västergötland, Sweden. Fossils and Strata 17:1–65. Geologica 105:1–25. 34. Müller KJ, Walossek D (1988) External morphology and larval development of the 59. Butterfield NJ (1997) Plankton ecology and the Proterozoic-Phanerozoic transition. Upper Cambrian maxillopod Bredocaris admirabilis. Fossils and Strata 23:1–70. Paleobiology 23:247–262. 35. Baier CT, Terazaki M (2005) Interannual variability in a predator-prey interaction: 60. Maderspacher F (2010) Palaeontology: The new conservative. Curr Biol 20:R513–R515. Climate, chaetognaths and copepods in the southeastern Bering Sea. J Plankton Res 61. Hart RC, Bychek EA (2011) Body size in freshwater planktonic crustaceans: An over- 27:1113–1125. view of extrinsic determinants and modifying influences of biotic interactions. Hy- 36. Dalpadado P, Yamaguchi A, Ellertsen B, Johannessen S (2008) Trophic interactions of drobiologia 668:61–108. macro-zooplankton (krill and amphipods) in the Marginal Ice Zone of the Barents Sea. 62. Verity PG, Smetacek V (1996) Organism life cycles, predation, and the structure of Deep Sea Res Part 2 Top Stud Oceanogr 55:2266–2274. marine pelagic ecosystems. Mar Ecol Prog Ser 130:277–293. 37. Itoh K (1970) A consideration on feeding habits of planktonic copepods in relation to 63. Kaartvedt S (2010) Diel vertical migration behaviour of the Northern krill (Mega- the structure of their oral parts. Bull Plankton Soc Japan 17:1–10. nyctiphanes norvegica Sars). Adv Mar Biol 57:255–275.

1594 | www.pnas.org/cgi/doi/10.1073/pnas.1115244109 Harvey et al. Downloaded by guest on October 1, 2021