Comparison of geochemical and distinctive mineralogical features associated with the Kinzers and Burgess Shale formations and their associated units

Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

j ourna l homepage: www.e lsev ie r.com/ locate /pa laeo

Comparison of geochemical and distinctive mineralogical features associated withthe Kinzers and Burgess Shale formations and their associated units

Wayne PowellDepartment of Geology, Brooklyn College, 2900 Bedford Avenue, Brooklyn, NY 11210, USA

E-mail address: [email protected].

0031-0182/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.palaeo.2009.02.016

a b s t r a c t

Article history:

Keywords:Herringbone calciteCambrianBurgess ShaleKinzersGeochemistry

This paper compares two N(BST): the Middle CambriaMiddle Cambrian Kinzers FSc) indicate that palaeo-oxOgygopsis-bearing fauna inmetal signature. Burgess Shwhereas the BST-fossil bedsthat consistently indicates

orth American occurrences of fossils that exhibit Burgess Shale-type preservationn Burgess Shale Formation of southeastern British Columbia, and the Lower tom of southeastern Pennsylvania. Redox-sensitive trace elements (Mo, Ni/Co, V/ygenation conditions varied with respect to faunal assemblage and location.both the Burgess Shale and Kinzers formations are associated with a dysoxicale-type preservation in the Kinzers Fm was deposited in a fluctuating oxycline,of Fossil Ridge and Mt Stephen (Burgess Shale Fm) yield a geochemical signatureoxic conditions. Thus, sustained low-oxygen conditions do not correlate with

higher quality of fossil preservation at these sites. Ogygopsis is associated with dysoxic conditions in bothlocalities, and may therefore be a faunal indicator of such environments in Middle Cambrian strata.Carbonate units associated with both the Burgess Shale and Kinzers formations contain large, generallyconcordant, near-surface cavities that contain herringbone calcite cements that show no significantenrichment in the potential calcite inhibitor ions such as reduced Fe, Sr, S, Pb, and Zn. The abundance ofmicro-inclusions of dolomite suggests that Mg may have been the calcite inhibitor ion responsible forherringbone cement in these Cambrian occurrences. Another distinctive mineralogical feature is thepresence of doubly-terminated, euhedral, zoned, authigenic quartz as a conspicuous part of a carbonate host-rock replacement assemblage. The association of both herringbone calcite and zoned quartz crystals in unitsassociated with the Burgess Shale and Kinzers formations suggest a common feature in the depositional anddiagenetic environment of these two regions in which Burgess Shale-type fossil preservation occur; syn-depositional brine flow is the likely cause of these distinct mineralogical features. Brine seeps in thedepositional environment may have provided nutrients for stable high-abundance faunal communities in aslope setting, thereby accounting for the abundance of fossil material. Furthermore, high-salinity brine in thepore waters of carcass-bearing Cambrian muds may have further limited shallow burrowing, scavenging, anddecomposition rates, accounting for the exceptional preservation in these two localities.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Cambrian strata host more localities of fossils of non-mineralizedfauna than anyother interval inEarthhistory (Allison andBriggs,1993a).The factors and mechanisms that led to the exceptional preservation offossils during the Cambrian continue to be debated since the pioneeringwork of Butterfield (1990). Suggested controls on preservation includeanoxia (e.g., Allison, 1988), rapid burial (e.g., Babcock et al., 2001),matrix clay composition (e.g., Butterfield, 1990), lack of burrowing (e.g.,Allison and Briggs, 1993b), salinity variation (e.g., Babcock et al., 2001;Johnston et al., 2009-this volume), reduced permeability of hostsediments (e.g., Gaines et al., 2005), and combinations of these factors.

Despite being more common in the Cambrian, sites of exceptionalfossil preservation are still a rarity. Even within regions that are

ll rights reserved.

fossiliferous, such as the Burgess Shale Formation of southeasternBritish Columbia and the Kinzers Formation of southeastern Pennsyl-vania, fossil occurrences have a patchy distribution and are restrictedto specific horizons. Accordingly, the set(s) of processes and/orconditions responsible for fossil preservation must be temporally andspatially limited, and rare. Prior inquiries into the controls on soft-tissue preservation in Cambrian slope deposits have focused onchemical and mineralogical features in or around the fossilsthemselves; if fossilization is a result of broader chemical character-istics of the depositional/diagenetic environment (e.g., Johnston et al.,2009-this volume), then documenting and comparing chemical,mineralogical, and petrological features that are locally associatedwith Cambrian lagerstätten may prove a fruitful means of identifyingsome of the critical factors.

As a beginning to such comparative analysis, this paper discussestwo North American localities with Burgess Shale-type preservation(BST): the Burgess Shale Fm of southeastern British Columbia, and the

Fig. 1. Generalized map of Laurentia and its shelf and slope facies during the Middle Cambrian.

128 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

Kinzers Fm of southeastern Pennsylvania (Fig. 1). This paper servesboth to present new data on trace metal signatures in shales of theKinzers Fm, and occurrences of quartz megacrysts in associatedcarbonate units, as well as expanding upon descriptions of herring-bone calcite in the Eldon and Cathedral formations of BritishColumbia. In addition, this paper compares the characteristics ofthese chemical and mineralogical features with similar features thatoccur in the corresponding fossil locality; comparison relies primarilyupon previously published data, although some additional dataexpand the existing data sets of these localities. The possibleimplications of shared chemical and mineralogical features withrespect to fossilization processes and environments of deposition anddiagenesis are discussed.

2. Analytical methods

Microprobe analyses and elemental mapping was conducted at theAmerican Museum of Natural History-Lamont-Doherty Earth Obser-vatory Electron Microprobe facility which houses a 5-spectrometreCameca SX-100 electronmicroprobe. Analyses were obtained using anaccelerating voltage of 15 keV and a beam current of 10 nA.

Carbonate samples were analyzed by Activation LaboratoriesLimited of Ancastar, Ontario, Canada. Samples are prepared andanalyzed in a batch system. Each batch contains a method reagentblank, certified reference material and 17% replicates. Calibration isperformed using 7 prepared USGS and CANMET certified referencematerials. One of the 7 standards is used during the analysis for everygroup of ten samples. Major oxides, Be and Sr were analyzed by fusioninductively coupled plasma spectroscopy (ICP). One gram sampleswere mixed with a flux of lithiummetaborate and lithium tetraborateand fused in an induction furnace. The molten melt was immediatelypoured into a solution of 5% nitric acid containing an internal standard,and mixed continuously until dissolved completely (~30 min). Thesamples were run on a combination of simultaneous/sequentialThermo Jarrell-Ash ENVIRO II ICP or a Spectro Cirros ICP. Totaldigestion ICP was used for S, Cd, Ni, Pb, and Zn; 0.25 g samples weredigested with four acids beginning with hydrofluoric, followed by amixture of nitric and perchloric acids, heated using precise program-mer controlled heating in several ramping and holding cycles whichtakes the samples to dryness. After drynesswas attained, sampleswerebrought back into solution using hydrochloric acid. Samples wereanalyzed using a Perkin Elmer Optima 3000 ICP. Neutron activationwas used to analyze for Co; an approximately 30 g aliquot wasencapsulated and weighed in a polyethylene vial and irradiated withflux wires and an internal standard (1 for 11 samples) at a thermalneutron flux of 7×1012 n cm−2 s−1. After a seven day decay to allow

Na-24 to decay the samples were counted on a high purity Ge detectorwith a resolution of better than 1.7 KeV for the 1332 KeV Co-60. Usingthe flux wires the decay corrected activities are compared to acalibration developed from multiple certified international referencematerials.

Shale samples were analyzed by XRAL Laboratories in Don Mills,Ontario, Canada. All elements were digested using the lithiummetaborate fusion method, and the resulting solutions analyzedusing ICP–AES. Organic C content was determined by coulometery.

3. Depositional settings

The continental margins of Laurentia evolved from active riftmargins in the Late Neoproterozoic and Early Cambrian, to passivemargins undergoing thermal subsidence during the latter Early toMiddle Cambrian (Bond et al., 1985). In the Appalachians of Virginiaand Pennsylvania, Mesoproterozoic bedrock gneiss is overlain uncon-formably by Neoproterozoic to Lower Cambrian rift-related sedimen-tary and volcanic rocks. A schematic representation of stratigraphicunits is presented in Fig. 2. Rift deposits accumulated in NE–SWtrending rift basins, including the Catoctin Rift which underlies theConestoga Valley region in which lie the fossiliferous strata of theKinzers Formation (Faill, 1999). By the earliest Cambrian, clasticsediments of the Chilhowee Group had filled the Catoctin Rift basin,and a subsiding passive margin was established (Faill, 1999). InPennsylvania, the Chickies Fm is overlain by the Vintage Fm, a thick-bedded to massive blue-grey, fine-grained, Early Cambrian crystallinedolostone which marks the onset of carbonate sedimentation in adeep-marine setting (Gohn, 1976). Correlative strata of the betterpreserved lower Shady Dolomite in Virginia include a thick basalsequence of deep-ramp nodular limestones that shallow upward intoa carbonate ramp fringed by low-relief mud mounds (Barnaby andRead, 1990).

The 20 to 50 m thick, laterally persistent Emigsville Mbr of theKinzers Fm overlies the Vintage ramp deposits. It hosts fossils of soft-bodied organisms (Campbell, 1969; Briggs, 1978; Capdevila andConway Morris, 1999; Skinner, 2005), and records regional hemi-pelagic deposition in a periplatformal environment (Taylor andDurika, 1990). The Emigsville Mbr consists of three facies: 1) alower pelitic facies that consists of thin, discontinuous beds thatcontain common fossils and a low intensity of bioturbation; 2) amiddle pelitic facies that consists of moderately to thickly beddedshale to silty shale inwhich bioturbation is relatively common; and 3)an impure carbonate facies that consists of tan carbonate-richmudstones and siltstones with common burrows and echinodermfossils (Skinner, 2005).

Fig. 2. Generalized stratigraphic columns for the Neoproterozoic to Middle Cambrian sections in the Conestoga Valley, PA and southeastern British Columbia. Section thickness is notdrawn to scale.

129W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

Themiddle member of the Kinzers Fm (York Mbr) consists of well-bedded limemudstone, tabular-bedded oolitic grainstones, andminormegabreccia that were deposited in a deep ramp or periplatformalenvironment (Taylor and Durika, 1990). The uppermost Kinzers Fm inYork County is assigned to the Greenmount Mbr, a dark, pyritic,siliciclastic-rich, laminated to nodular limestone that contains trilo-bites indicative of off-platform deposition (Taylor and Durika, 1990),and which represent the Poliella-Plagiura Zone of the lowermostMiddle Cambrian (Taylor et al., 1997). In Lancaster County, theuppermost Kinzers Fm is assigned to the vertically and laterallyvariable Longs Park Mbr, a unit that consists of black shale, dark sandylimestone, dark pyritiferous limestone, and grey crystalline dolostone(Campbell, 1969). This unit hosts an Ogygopsis-bearing fossil fauna(Campbell,1971). TheGreenmount and Longs Parkmembers representdeepening, or decline in carbonate influx, during deposition of theupper Kinzers Fm, but it is unclear how these two units correlate due toinsufficient biostratigraphic data (Taylor and Durika, 1990).

Shelf-margin carbonate rocks of the Middle Cambrian Ledger Fmconformably overlie the Kinzers Fm (Gohn, 1976). The extensivelydolomitized strata of the Ledger Fm contain relict oolitic textures andcross-stratification, similar to back-reef and shelf-margin facies of themiddle Shady Dolomite (Austinville Mbr) (Barnaby and Read, 1990;Taylor and Durika, 1990). The Willis Run Mbr is a limestone intervalwithin the Ledger Fm that is composed predominantly of bioturbated,

thin-bedded mudstone deposited in a deep shelf environment (Taylorand Durika,1990). The sparse fauna from this unit indicates depositionduring the Glossopleura Zone of the Middle Cambrian (Taylor et al.,1997). The Willis Run Mbr also contains micrite-free microbialitesintercalated with oolitic grainstones, that are interpreted to have beendeposited in a high-energy shelf-margin setting (de Wet et al., 2004).

The Conestoga Fm is a time-transgressive unit that is the deep-water,basinal equivalent to all units from the Lower Cambrian Vintage Fm tothe Upper Cambrian Conococheague Group (Rodgers, 1968). TheConestoga Fm is composed of thinly bedded limestone-shale rhythmites,lithoclastic grainstone, massive oolitic grainstone, graphitic limestone,phyllite, and polymictic megabreccia (Taylor and Durika, 1990).

Late Proterozoic rift sequences in the southeastern Cordillera arecomposed of the Neoproterozoic to Lower Cambrian Hamill Group(Selkirk and Purcell Mountains) and the correlative Lower CambrianGog Group (Rocky Mountains) that overlie the earlier NeoproterozoicWindermere Supergroup (Devlin and Bond,1988). The predominantlyorthoquartzite strata of the Gog Group and Upper Hamill Groupcontain common Early Cambrian trace fossils (e.g., Kubli and Simony,1992) and are capped by shallow marine strata that includearchaeocyathid-bearing limestone and other fossils indicative of theOlenellus-Bonnia Zone (Fritz et al., 1991).

Whereas Appalachian rocks indicate that a subsiding passivemarginwas established along the south coast of Laurentia by the Early

130 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

Cambrian, strata from theRockyMountains indicate that the north coastdid not reach a similar state until the Middle Cambrian (Bondand Kominz, 1984). In southeastern British Columbia, the shelf breakrepeatedly developed along a topographically high hinge zone that wasprobably controlled by underlying basement faults. This reoccurringlinear palaeotopographic feature is referred to as the Kicking Horse Rim(Aitken,1971). To thewest of theKickingHorseRim (oceanward), slope-facies shales and siltstones of the Middle Cambrian Naiset Formationunconformably overlie the rift-related sedimentary rocks of the GogGroup, and locally contain an Ogygopsis-bearing fauna and lenses ofseep-related chloritic phyllites (Powell et al., 2006). Landward of theKicking Horse Rim, the Mount Whyte Formation lies at the base of theMiddle Cambrian section. It is composed of shale with oolitic andoncolitic limestone beds (Aitken, 1997). The Mt. Whyte and Naisetformations lie within the Plagiura-Poliella Zone (Aitken, 1997).

The predominantly Glossopleura Zone Cathedral Formation is theoldest of the major Middle Cambrian cliff-forming carbonate units insoutheastern British Columbia (Fritz, 1971). Along the Kicking HorseRim the unit is extensively dolomitized; non-dolomitized strataconsist mainly of pellet grainstone, ooid grainstone, laminites ofmicrobial origin, and fenestral limemudstone, and herringbone calciteis a common cavity-fill in microbialite strata (Aitken, 1997; Pratt,2002). Along the outboard edge of the Cathedral platform, thesecarbonate rocks abut the younger (Bathyuriscus-Elrathina Zone)basinal strata of the Burgess Shale Fm along a steeply dipping contactknown as the Cathedral Escarpment, a gravity slide-scar that truncatedthe outer edge of the platform (Stewart, 1991). Behind the CathedralEscarpment the Cathedral Fm is overlain conformably by the StephenShale Fm (Fletcher and Collins, 1998) which accumulated during aslight deepeningeventwith increased terriginous input (Aitken,1997).

Platformal carbonate accumulation resumed with the Eldon Fm(Bathyuriscus-Elrathina Zone; Aitken, 1997). It consists of abundantoolitic grainstones along with stromatolitic and thrombolitic strataalong the Kicking Horse Rim (Aitken, 1997). The Field Mbr is a distinctrecessive, black unit within the Eldon Fm that consists of ribboncarbonates with interbeds and lenses of seep-related rock (Powell et al.,2006). Like the Cathedral Fm, the Eldon platformwas truncated to forma moderately dipping gravity-slide scar called the Eldon Escarpment(Stewart et al., 1993).

The Chancellor Group is a shale-dominated basinal sequence thatwas deposited oceanward of the Kicking Horse Rim, and includes theBurgess Shale Formation in which Fletcher and Collins (1998) definedten members (Fig. 2). The oldest unit, the Kicking Horse Shale Mbr, lieswithin the Glossopleura Zone, whereas the rest of the Burgess ShaleFormation is Bathyuriscus-Elrathina Zone (Fletcher and Collins, 1998).The five oldest calcareous shale-dominated members (Kicking Horse,Campsite Cliff, Walcott Quarry, Raymond Quarry, and Emerald Lake)host exceptionally preserved fossil beds, and the basal strata of theCampsite Cliff Mbr contain an Ogygopsis-bearing fauna (Fletcher andCollins, 1998). With the exception of the Emerald Lake Oncolite Mbr,these shaley units contain lenses of seep-related rock adjacent to theCathedral Escarpment (Powell et al., 2006). The Yoho River Limestoneand Wash Limestone members thicken toward the Cathedral Escarp-ment and consist of dark, thinly bedded limemudstone and debris-flowbreccias (Fletcher and Collins, 1998), and small carbonate mudmounds(Johnston and Collom, 2001). Shallow-water sedimentary structuresincluding oncolites and ripple laminations occur in the four uppermostmembers (Fletcher and Collins, 1998).

The Burgess Shale Fm overlies ribbon carbonates of the TakakkawTongue Fm, the slope-facies equivalent of the Cathedral Formation(Fletcher and Collins, 2003). The megatruncation surface that is theshallowly dipping extension of the Cathedral Escarpment stripped awaythe upper beds of this unit (Stewart, 1991). Similarly, the ribbon-carbonate dominated TokummUnit is the slope-facies equivalent of theEldon Fm, and ribbon carbonates lying above the extension of the EldonEscarpment lie within the Vermilion Subunit (Stewart, 1991) which

includes Ba-rich brine-seep related strata (Powell et al., 2006). TheVermilion Subunit is overlain conformably by the shaley soft-bodiedfossil-bearing Duchesnay Unit (Motz et al., 2001). Both the VermilionSubunit and Duchesnay Unit (Bolaspidella Zone; Motz et al., 2001)contain carbonate mud-mound-rich horizons (Powell et al., 2006).

The Cambrian sequences of the Conestoga Valley and southeasternBritish Columbia share a common tectonic and geographic setting;sediments in both localities accumulated on a continental margin thatwas undergoing post-rift subsidence in tropical latitudes, and so exhibit abroad similarity in composition and sequence of sedimentary units. Pooroutcrop in the Conestoga Valley means the regional geometry ofstratigraphic units in Pennsylvania is less certain. Nevertheless, sig-nificant differences in the architecture of these two localities exist. TheCathedral and Eldon formations of British Columbia indicate platformalbuild-up, whereas the Vintage and Kinzers carbonates accumulated on aramp with minor organic buildups along the shelf break. Onset ofplatformal development in the Appalachian section coincides with theLedger Formation.No evidence of regionally extensive collapse structuresakin to the Cathedral escarpment have been documented in theConestoga Valley section. However, breccia lenses in the York Memberof the Kinzers Fm, and within the Conestoga Fm (Taylor and Durika,1990), indicate that localized collapse of the carbonate strata occurredduring deposition of the Kinzers and Ledger formations.

4. Redox-sensitive trace metals

Skinner (2005) reports that bedding-parallel burrows are a rarityin shales of the Emigsville Mbr of the Kinzers Fm, and that sessilebenthic organisms were buried in situ on at least one horizon in thelower pelitic facies. Abundant organic carbon is rare, and theabundance of organic carbon correlates with the abundance of pyritein the Emigsville Mbr, indicating that pyrite formed during post-burialmicrobial activity rather than from anoxic seawater during deposition(Skinner, 2005). This data led Skinner (2005) to conclude that there islittle mineralogical or palaeontological evidence for widespread, long-term anoxia during deposition of the Emigsville Mbr, but rather thatthis unit probably was deposited in an exaerobic environment.

Analysis of redox-sensitive trace elements in shales of the Emigsvilleand Longs Park members of the Kinzers Formation was conducted forthis study to improve our understanding of palaeo-redox conditionsduring deposition of the Kinzers Fm, and to allow better comparisonwith the existing geochemical data available for the Burgess Shale(Powell et al., 2003). Mudstones deposited under anoxic conditions arecommonly enriched in a series of metals relative to average shalecompositions. Molybdenum, silver, zinc, nickel, copper, chromium,vanadium, uranium and platinum-group elements have all been notedto be significantly enriched both in modern muds deposited in anoxicbasins and in ancient black shales (e.g., Vine and Tourtelot,1970; Calvertand Pederson, 1993), thereby allowing oxic, dysoxic and anoxicdepositional settings to be identified in ancient shales. For example,the BST-bearing Ordovician Soom Shale of South Africa contains anaverage of 24.5 ppmMo (Gabbott, 1998). This contrasts with 2 ppmMoin an average shale (Taylor and McLennen, 1985) and providesgeochemical evidence for anoxic conditions during deposition of stratawith exceptional fossil preservation in the Soom Shale.

Metals that concentrate in low-oxygen environments include thosewhose valency can vary with local redox potential. Elements such as Cr,Mo, Re, U and V form highly soluble ions under oxygenated conditions,but under anoxic conditions produce insoluble materials in their lowervalency state (Calvert and Pederson, 1993). Redox-sensitive elementsalso include those that formhighly insoluble sulphides: Cd, Cu, Ni and Zn(Calvert and Pederson,1993).WhereasMo on its own has proven to be areliable redox indicator (e.g., Gabbott, 1998), many other metals areproblematic because their source may include a detrital component.Consequently, researchers have employed a series of metal ratios asproxy indicators of palaeo-redoxconditions. Thebehaviorof oneelement

Table 1Geochemical proxies for redox conditions in the Burgess Shale Formation.

Sample Location Fossils Org C Mo Ni Co Ni/Co V Sc V/Sc

Raymond 1a FR Tuzoia 0.28 b2 52 17.4 3.0 92 20 4.6Raymond 2a FR Margaretia 0.20 b2 54 18.5 2.9 99 20 5.0Raymond 3 MS Vauxia b2 44 20.4 2.2 103 22 4.7Raymond 4 MS Vauxia b2 33 12.0 2.8 105 20 5.3Raymond 5 MS Vauxia b2 24 8.2 2.9 105 20 5.3Raymond 6 MS Nisusia b2 51 23.0 2.2 96 18 5.3Raymond 7 MS Priscansermarinus b2 53 14.6 3.6 91 17 5.4Raymond 8 MS Olenoides b2 23 8.9 2.6 112 20 5.6Walcott 1 FR Marrella 0.24 b2 35 30.1 1.2 101 18 5.6Walcott 2 FR Vauxia 0.26 b2 56 19.0 2.9 92 19 4.8Walcott 3 FR Unidentified 0.28 b2 54 48.0 1.1 89 20 4.5Walcott 4 MS Anomalocaris b2 34 19.8 1.7 101 20 5.1Walcott 5 MS Scenella b2 29 21.8 1.3 98 20 4.9Campsite 1a MS Ogygopsis 0.17 b2 57 19.4 2.9 98 19 5.2Campsite 2a MS Ogygopsis 0.18 b2 54 20.5 2.6 107 18 5.9Campsite 3a MS Ogygopsis 0.17 b2 45 11.3 4.0 103 17 6.1Campsite 4a MS Ogygopsis 0.25 b2 61 17.3 3.5 92 16 5.8Campsite 5 MS Ogygopsis b2 50 22.1 2.3 100 20 5.0Campsite 6a MS Anomalocaris 0.32 b2 68 31.5 2.2 93 20 4.7Campsite 7 MS Anomalocaris b2 47 12.3 3.8 91 19 4.8Campsite 8a MS None b2 51 14.6 3.5 106 13 8.2Campsite 9a MS Ogygopsis b2 49 23.6 2.1 75 13 5.8Campsite 10a MS Ogygopsis b2 52 17 3.1 80 17 4.7Campsite 11a MS None b2 33 10.5 3.1 104 11 9.5Campsite 12a MS None b2 32 10 3.2 79 16 4.9Campsite 13a MS None 3 32 6.1 5.2 122 b10 N12.2

Metals are measured in ppm. Organic carbon is measured in wt.%. Sample sites: FR = fossil ridge; MS = Mt Stephen.a Denotes analyses that were reported in Powell et al. (2003).

131W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

in the ratio (numerator) is dependent upon redox conditions, whereastheother co-varyingelement (denominator) in generally independentofsuch variables. Thus the value of the ratio will increase when the redox-dependent metal precipitates and is added to the sediment.

Not all elemental ratios are equally reliable. Jones and Manning(1994) evaluated the reliability of a set of proxy palaeo-redox indicatorsand concluded that themost reliable methods were U/Th, V/Cr and Ni/Co. Kimura andWatanabe (2001) suggest thatV/Sc ismore sensitive andreliable thanV/Cr. Itmayhaveadditional advantage for theBurgess Shalebecause V correlates better with Sc than Cr in shale at the Precambrian–

Table 2Geochemical proxies for redox conditions in the Kinzers Formation.

Sample Location Fossils Org C Mo

Longs Park 1 10C Ogygopsis 0.62 6Longs Park 2 10C Ogygopsis 0.73 3Longs Park 3 10C Ogygopsis 6Longs Park 4 10C Olenoides 6Longs Park 5 10C Elrathina 0.73 8Longs Park 6 10C Byronia 0.51 8Longs Park 7 10C Spicules 0.83 4Longs Park 8 10C Spicules 0.93 10Longs Park 9 Rte 30 Diagonialla 2.42 6Emigsville 1 22L Olenellus 0.19 b2Emigsville 2 22L Olenellus 0.23 b2Emigsville 3 22L Hyolithids 0.16 b2Emigsville 4 22L Tubullela 0.11 b2Emigsville 5 22L Margaretia 0.13 b2Emigsville 6 22L Marpolia b5Emigsville 7 22L Marpolia b5Emigsville 8 22L Morania b5Emigsville 9 Stoner Morania b5Emigsville 10 Fruitville Olenellus b2Emigsville 11 Fruitville Olenellus b2Emigsville 12 Getz Olenellus 0.14 b2Emigsville 13 Getz Olenellus 0.15 b2Emigsville 14 Getz Olenellus 0.18 b2Emigsville 15 Getz Olenellus b2Emigsville 16 Getz Olenellus b2Emigsville 17 Getz Olenellus b2Emigsville 18 Getz Olenellus b2

Metals are measured in ppm. Organic carbon is measured in wt.%. Location names correspo

Cambrian transition (H. Kimura, pers. comm. 2002). Powell et al. (2003)determined that Ni/Co and V/Sc preserved the most detailed record ofpalaeo-redoxconditions in theBurgess Shale. Algeo andMaynard (2004)noted that Mo and V behave differently from Ni with respect tocovariance with organic carbon; V and Mo exhibit stronger covariancewith TOC than does Ni in rocks with less than 10 wt.% TOC. All BurgessShale and Kinzers formation shales contain far less than 10 wt.% TOC(Tables 1 and 2). Values of V/Scb9.1 correspond to normal shale,whereas V/Sc≥9.1 corresponds to low-oxygen depositional environ-ments (Kimura and Watanabe, 2001). Values of Ni/Cob5 in shale

Ni Co Ni/Co V Sc V/Sc

24 4.7 5.1 341 13 26.221 4.2 5.0 375 13 28.822 4.2 5.2 292 14 20.926 5.8 4.5 320 13 24.619 3.7 5.1 471 14 33.623 5.6 4.1 469 15 31.320 4.9 4.1 395 13 30.424 4.9 4.9 389 13 29.926 8.1 3.2 233 b10 N23.315 4.9 3.1 103 19 5.413 4.8 2.7 101 20 5.122 3.9 5.6 101 21 4.810 3.5 2.9 104 23 4.519 3.4 5.6 104 21 5.013 4 3.3 114 21.2 5.417 4 4.3 112 22.1 5.17 3 2.3 117 21.4 5.533 17 1.9 108 18.7 5.821 5.3 4.0 113 19 5.934 14.1 2.4 115 18 6.413 5.0 2.6 101 20 5.114 4.3 3.3 99 19 5.211 5.5 2.0 107 20 5.418 12.6 1.4 99 19 5.218 4.7 3.8 103 20 5.214 4.4 3.2 96 19 5.113 4.7 2.8 101 20 5.1

nd to those of Campbell (1969).

Fig. 3. Cross plots of two geochemical proxies for redox conditions, Ni/Co and V/Sc. The white area represents agreement for oxic conditions between the two indicators. Light greyareas represent disagreement for redox conditions between the two indicators. Darker grey areas represent agreement for low-oxygen conditions between the two indicators.

132 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

indicate oxic conditions, whereas Ni/CoN7 corresponds to deposition inan anoxic environment (Jones and Manning, 1994). Intermediate valuesof Ni/Co (between 5 and 7) indicate deposition in the dysoxic zone(Jones and Manning, 1994). Accordingly, data herein is presented in theform of a plot of V/Sc vs. Ni/Co to highlight the most significantvariations in redox conditions.

Powell et al. (2003) studied geochemical proxies of seawater redox-conditions for four Burgess-Shale-Type (BST) fossil-bearing sectionsin southern British Columbia. Table 1 compiles geochemical data fromtwo of these sites (Fossil Ridge and Mount Stephen) as reported inPowell et al. (2003), as well as 13 new analyses, new organic carbondata, and a listing of fossils that were contained within the analyzedsamples. Trace metals indicate that oxygenated conditions prevailedin the overlying seawater during deposition of beds within which BSTpreservation developed, whereas strata from the Ogygopsis-bearingsection yielded values that indicate a more complex setting withdeposition under varying redox conditions (Powell et al., 2003).

Redox-related geochemical characteristics of the Emigsville andLongs Parkmembers of the Kinzers Fm (Table 2) are broadly similar tothose of the Walcott Quarry and Campsite Cliff members of theBurgess Shale, respectively (Fig. 3). The Ogygopsis-bearing strata ofthe Longs Park Mbr are relatively enriched in Mo (up to 10 ppm), arerelatively rich in organic carbon (0.5 to 2.4 wt.%), and have moderateNi/Co (up to 5.1) and elevated V/Sc ratios (up to 33.6) (Table 2). Thecoincidence of several geochemical indicators of dysoxia in the LongsPark Mbr in numerous beds provides reasonable confidence that thisunit was deposited under sustained conditions of reduced oxygen.

Mo in all samples from the Emigsville Mbr is below detection limit,organic carbon is less than 0.23 wt.%, and V/Sc is between 4.5 and 6.4(Table 2), all indicative of an oxygenated depositional environment.Most samples that contain exceptional preservation of fossils areindistinguishable chemically from those with more typical skeletalfossil beds. However, two samples from the most prolific BST site (22Lof Campbell, 1969) have Ni/Co ratios that are indicative of the dysoxiczone as defined by Jones and Manning (1994). This pattern suggeststhat the Emigsville Mbr was deposited in a fluctuating oxycline.

Fig. 4. Features of herringbone calcite from the Cathedral Formation. All images are in sectiB) Cavity in oolitic rock that is completely filled with herringbone calcite. C) Cavity in midolomite (sd); D) Herringbone calcite-filled fracture network; E) Angular clasts with mencrustations in herringbone-filled cavity. G) Concentric zoning in cavity from radiaxial caherringbone calcite in Pre-Burgess Shale collapse breccia that is overlain by Takakkaw Tong

The basal Campsite Cliff Mbr on Mount Stephen and the Longs ParkMbr (10C site of Campbell,1969) both yield geochemical signatures thatare consistent with deposition in a dysoxic environment. In addition,both localities contain a similar fossil assemblage that includes the largetrilobite Ogygopsis klotzi (Campbell, 1971). In southeastern BritishColumbia, Ogygopsis klotzi also occurs within the pre-Burgess ShaleNaiset Fm where it is spatially associated with a lens of brine-pool-related strata (Powell et al., 2006). The spatial association betweendysoxic facies strata and Ogygopsis suggests that this trilobite inhabitedlower Middle Cambrian low-oxygen environments, similar to Elrathiakingii in the youngerMiddle Cambrian strata of theWheeler Fm in Utah(Gaines and Droser, 2003).

The Ogygopsis beds in the Longs Park Mbr yield trilobites,brachiopods, sponges, and worm tubes but lack non-mineralizedfossils (Campbell, 1970). The same assemblage dominates the MountStephen assemblage, with the addition of anomalocarid feedingappendages (Fletcher and Collins, 2003). Thus the degree of fossilpreservation in the Ogygopsis-bearing beds in both British Columbiaand Pennsylvania is less than that of the classic Burgess Shale-typelocalities such as the Walcott Quarry and 22L. Geochemical signaturesindicate that the Ogygopsis-bearing beds were deposited under loweroxygen conditions than that of the strata that host fossils of soft-bodied fauna. Thus sustained dysoxia/anoxia in the depositionalenvironment does not appear to correlate with Burgess Shale-typepreservation in either the Kinzers or Burgess Shale formations.

5. Unusual diagenetic mineral features

5.1. Herringbone calcite

Herringbone calcite is a form of carbonate cement and seafloorprecipitate that is composed of paired light and dark crenulated bandsthat are less than 1 mm thick (Fig. 4A). Herringbone calcite iscomposed of needles of calcite oriented perpendicular to the banding(Sumner and Grotzinger, 1996). The dark base of each crystal isoptically unoriented, but the optical c-axis rotates upward through

on view. A) Detail of fine, chevron texture of herringbone calcite, scale in centimetres.crobialite infilled with microbial encrustations (mc), radiaxial calcite (ra), and saddleicrobial crusts and herringbone calcite with dark micrite bands. F) Bushy microbiallcite (ra) to herringbone calcite (hb) to saddle dolomite (sd); H) Clast of dolomitizedue. The coin in C,E and H is 19 mm in diameter.

133W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

134 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

the length of the crystal until it is perpendicular to crystal length in thelight-colored crystal tops. Herringbone texture is preserved as low-Mgcalcite or dolomite, but a high-Mg calcite precursor is inferred basedupon textures and the common occurrence of dolomite micro-inclusions (Sumner and Grotzinger, 1996).

Herringbone calcite is most common in Archean strata, where itoccurs as both seafloor precipitates and cement due to the presence ofreduced iron in seawater (e.g., Hofman et al., 1985; Simonson et al.,1993; Sumner and Grotzinger, 1996). Younger strata rarely containherringbone calcite, that if present, occurs only as cement. MostPhanerozoic examples occur within small carbonate buildups that areisolated within organic-rich sediments associated with anoxia (Leh-man,1978), and thus the presence of reduced iron can be inferred. Onewell-documented exception is an occurrence in Eocene cave depositsin platformal carbonates in Egypt (Tourre, 2000).

Numerous ions other than Fe2+ can inhibit calcite growth.These include Zn 2+, Pb2+, Fe3+, Co2+, Mn2+, Be2+, Ni2+, Cd2+, Ba2+,Sr2+, SO4

2−, Mg2+, phosphates, and various organic compounds (Meyer,1984). The common correlation of herringbone calcite with anaerobicenvironments led Sumner and Grotzinger (1996) to conclude that thepresence of soluble Fe2+ in poorly oxygenated seawater is a factorcritical to the formation of most occurrences of herringbone calcite.However, Sr2+ and Mg2+ appear to be responsible for herringbonecalcite formation in caves in Middle Eocene carbonate platforms(Tourre and Sumner, 1999). Thus numerous chemical environmentscan lead to herringbone calcite development.

Although herringbone calcite is rare in Phanerozoic strata, thisdistinct form of cement is well-developed along the Kicking HorseRim in the Cathedral Formation, and to a lesser extent within theEldon Formation (Aitken, 1997; Pratt, 2002). This mineralogicalfeature was first documented in the region by McIlreath and Aitken(1976), who interpreted it as an encrusting organic sedimentarystructure to which they assigned the name Yoholaminites. Herring-bone cement occurs within predominantly horizontal cavities thatreach several metres in length and up to a half metre in height. Mostcavities are subparallel to bedding and have irregular, cuspatemargins, with the lower surface generally flatter than the uppersurface (Fig. 4B and C). However, many cavities are composed of anetwork of bedding-parallel fractures that are connected by dis-cordant veins, and contain angular clasts of the host rock (Fig. 4D andE). Shrubby microbial colonies up to 5 cm in diameter commonlyoccur along cavity walls, and are particularly well-developed on cavityceilings (Fig. 4C,E and F).

The cements that fill herringbone-bearing cavities vary. Somevoids have been filled entirely by herringbone calcite (Fig. 4B). Morecommonly, concentric radiaxial calcite layers line the cavity walls, andin turn are overgrown by herringbone calcite (Fig. 4G); in other cases,herringbone calcite overlies microbial encrustations directly (Fig. 4Eand F). Sparry dolomite is the youngest cementwithin cavities (Fig. 4Cand G). Dark micrite laminae separate herringbone bands, forming anirregular, swirly texture (Fig. 4E).

In thin-section it is evident that herringbone calcite occurs indiscrete bands 0.5 mm to 7 mm in width. Each band is separated by anarrow band along which dolomite inclusions are concentrated(Fig. 5). Within each band low-Mg calcite is intergrownwith dolomiteinclusions. The density of dolomite inclusions varies from band toband, and across the width of each band from bottom to top (Fig. 5);finer-grained bands have a higher percentage of dolomite inclusionsthan do coarser-grained, and dolomite inclusions increase in abun-dance from the bottom to the top of each band. There is no systematicvariation in the present composition of crystals of calcite and dolomitefrom band to band, and across each band (Table 4). Ba, Ni, Pb, Sr, andZn are below detection limits of electron microprobe analysis, and Mnis uniformly low (Table 4). Iron is concentrated in dolomite inclusionsand varies from non-detectable to 0.5 wt.% FeO (Table 4). Although Fecontent is low in Cathedral herringbone calcite (Table 3), X-ray

mapping indicates that minute Fe-bearing grains tend to be moreabundant in the dolomite-rich upper zone of finer-grained herring-bone bands (Fe X-ray map in Fig. 5).

Herringbone calcite in the Cathedral Formation exhibits weakenrichment in the potential calcite-inhibitor ions Fe, S and Sr relativeto associated radiaxial calcite andmicrobial calcite (Table 3). Zn is lowin both herringbone and microbial calcite; concentrations of Ba arelow, with radiaxial calcite containing a higher concentration of Ba thanin adjacent herringbone calcite; Be, Cd, Co, and Ni are below detectionlimits in all forms of cavity-fill calcite from the Cathedral Fm (Table 3).

Clasts of dolomitized herringbone calcite occur in collapse brecciasat the front of the Cathedral platform (Fig. 4H). The breccia is overlainby strata of the Takakkaw Tongue (Powell et al., 2006; Johnston et al.,2009-this volume). Pratt (2002) noted cross-laminated, geopetalooidal grainstone within herringbone-bearing cavities. These fieldrelationships indicate that herringbone cement formed in shallowwater, within near-surface cavities, very early in the diagenetic historyof the unit. Some of these cavities formed teepee structures which ledPratt (2002) to conclude that they formed in response to earthquakesalong the seismically active Kicking Horse Rim.

Herringbone-bearing cavities occur also within microbial reefs ofthe Lower Ledger Formation and have been described in detail by deWet et al. (1999, 2004), and deWet and Hopkins (2006). The host rockis a dark limestone that is dominated by stratiform microbial sheets,although domal stromatolites and thrombolites are present as well(de Wet and Hopkins, 2006). Channel-deposited grainstone lenses liebetween microbial build-ups (de Wet et al., 2004). These rocksaccumulated in a subtidal, well-winnowed environment, consistentwith a shallow ramp or shelf (de Wet et al., 2004).

Herringbone calcite of the Ledger Fm occurs in roughly horizontalcavities that have relatively flat floors and upper surfaces that aremore irregular and dentate (de Wet et al., 2004) (Fig. 6). Thesehorizontal cavities, which may be up to 2 m in length and 0.5 mwide,are commonly associated with vertical, strata-cutting fractures, andupward-cutting fissures (de Wet et al., 2004). Despite the large size ofthese cavities, de Wet et al. (2004) concluded that these horizontalvoids are stromatactis cavities that resulted from episodic escape ofseawater that was trapped in the rapidly accumulating microbialite.

Microbial encrustations always occur along the walls of herring-bone-bearing fractures (Fig. 6), and are overlain by a sequence offibrous calcite, herringbone calcite, and baroque dolomite (de Wetet al., 1999). Geopetal sediments that contain shell fragments mayoverliemicrobial encrustations or cement layers (deWet and Hopkins,2006).

No herringbone clasts have been documented yet within theLedger Formation, but fragments of radiaxial calcite cement form thecores of ooids in the herringbone-bearing sequence (de Wet et al.,2004). The occurrence of clasts of herringbone-associated cements,and the intercalation of geopetal sediments with herringbone calcitein cavities suggests that these cavities and their cements formed veryearly in the history of the Ledger microbial reefs (de Wet et al., 2004).

Similar to the herringbone carbonates of the Cathedral Fm, thoseof the Ledger Fm are composed of discrete bands separated byconcentrations of microdolomite inclusions. Similar inclusions occurwithin bands of herringbone calcite and radiaxial calcite, and dolomiteis concentrated along the boundary between herringbone bands.Increase in density of dolomite inclusions across bands is not as welldeveloped as in Cathedral Fm examples. Nevertheless, there is anasymmetric distribution of dolomite inclusions across the boundariesof bands. Dolomite inclusions are interpreted to have developedduring diagenetic stabilization of low-Mg calcite from a high-Mgprecursor (Sumner and Grotzinger,1996; deWet et al., 2004), and thatthe transformation of high-Mg to low-Mg calcite occurred in a closedsystem, at least with respect to Mg.

Fe, Mg, Mn, and Sr are below detection limits for EDAX elementalmapping (deWet and Hopkins, 2006). New ICP-based analyses indicate

Fig. 5. BSE and elemental maps (Ca, Mg, Fe, Sr) across herringbone calcite bands from the Cathedral Formation. Light areas in BSE are calcite and dark areas are dolomite. Labeledpoints correspond to spot analyses in Table 4.Lighter colors in elemental maps represent higher elemental concentrations. Note that Fe concentration is very low, but Fe micro-grainconcentrate at the top (right) of bands. Sr is uniformly low.

135W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

that herringbone calcite in the Ledger Formation contains lowerconcentrations of Fe and Zn than associated radiaxial calcite (Table 3).Although low in concentration, herringbone calcite bands are moreenriched in the potential calcite-inhibitor ions Co, S, and Sr than theassociated radiaxial calcite (Table 3). Be, Cd, Ni and Pb are belowdetection limits in both herringbone and radiaxial calcite (Table 3).

Herringbone calcite occurrences in British Columbia and Pennsyl-vania share many similarities; they occur in cavities of similar size andshape that formed early in the diagenetic history of the hostlimestones; they are associated with well-developed microbial calcitestructures; they are composed of low-Mg calcite with abundantdolomite inclusions that increase in abundance from the bottom to top

Table 3Composition of carbonate minerals in herringbone calcite-bearing cavities.

CaO MgO Fe2O3 MnO P2O5 S Ba Be Cd Co Ni Pb Sr Zn

wt.% wt.% wt.% wt.% wt.% wt.% ppm ppm ppm ppm ppm ppm ppm ppm

FICP FICP FICP FICP FICP DICP FICP FICP DICP NA DICP DICP FICP DICP

Cath 1 Coarse HB 49.05 5.80 0.24 0.02 0.01 0.020 6 b1 b0.5 b1 b1 b5 385 2Cath 2 Fine HB 50.15 5.35 0.19 0.02 0.01 0.020 6 b1 b0.5 b1 b1 b5 377 2Cath 3 Microbialite 47.90 7.13 0.15 0.02 b0.01 0.015 5 b1 b0.5 b1 b1 b5 325 1Cath 4 RAC 54.27 2.43 0.11 0.01 0.06 14 b1 322Ledger 1 HB 56.33 0.99 0.11 0.01 b0.01 0.065 11 b1 b0.5 3 b1 b5 298 b1Ledger 2 HB 51.13 5.86 0.06 b0.01 b0.01 0.068 8 b1 b0.5 5 b1 b5 385 b1Ledger 3 Dark RAC 52.33 2.40 0.14 b0.01 b0.01 0.022 8 b1 b0.5 b1 b1 b5 261 5Ledger 4 Light RAC 56.18 0.90 0.12 b0.01 b0.01 0.029 7 b1 b0.5 b1 b1 b5 228 3

HB = herringbone, RAC = radiaxial calcite. FICP = fusion ICP, DICP = total digestion ICP, NA = activation.

136 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

of individual laminae; with regard to composition, these herringbonecalcite occurrences show little or no enrichment in the commoncalcite-inhibitor ions Fe, Mn, S, and Sr, as well as the rarer metals Ba,Be, Cd, Co, Ni, Pb, and Zn.

The Cambrian herringbone calcite occurrences contrast sharplywith most other Phanerozoic occurrences of herringbone calcite; themicrobial reefs of the Cathedral, Eldon, and Ledger formations are notspatially associated with organic-rich sediments or other evidence ofregional anoxia, do not contain elevated concentrations of iron, andare not associated with iron-bearing minerals. Thus it is unlikely thatthe development of herringbone calcite in these localities resultedfrom the presence of reduced iron. Unlike the Eocene Egyptianoccurrence of herringbone calcite with an average Sr concentration of3200 ppm (Tourre, 2000), Sr enrichment in the Middle Cambrianherringbone calcite cements (up to 385 ppm; Table 3) is low relative

Fig. 6. Examples of herringbone calcite-bearing cavities in the Ledger Formation. Notetheflat bottoms, cuspate tops, and shrubbymicrobial encrustations. The coin is 24mm indiameter. (mc = microbial encrustation, ra = radiaxial calcite, sd = saddle dolomite).

to adjacent forms of calcite (up to 325 ppm), and so it seems unlikelythat Sr was responsible for herringbone cement growth in theCambrian localities. Nevertheless, the occurrence of well-developedherringbone calcite with broadly similar characteristics in theCathedral, Eldon, and Ledger formations suggest that these geogra-phically distant localities shared a similar, and unusual, chemicalenvironment in the very shallow subsurface during deposition ofthese units.

Tourre and Sumner (1999) documented that herringbone calcite inEgyptian cave deposits is enriched in both Sr andMg (5.5 mol%MgCO3

average), and that both elements may have been responsible for theanomalous cement. Dolomite exsolved from a high-Mg herringbonecalcite protolith during diagenesis of the Cathedral Fm, and so theexisting compositions of calcite in these localities cannot be compareddirectly with that of the more pristine Egyptian example. However, ifone assumes that the bulk chemistry of the Cambrian rocks has notchanged significantly on the scale of centimetres, then a pre-diageneticcomposition of approximately 13 mol% MgCO3 can be estimated forherringbone calcite of the Cathedral Formation (Table 4).

Although there is no distinct signature of calcite-inhibitor ions suchas Fe2+ and Sr2+, the Cambrian herringbone samples are consistentlyenriched inMg2+. Thus the Cathedral, Eldon, and Ledger occurrences ofherringbone calcite share compositional similarities with those of theEgyptian cave deposits that were documented by Tourre (2000). InBritish Columbia, the extensive outcrop allowed for the documentationof lenses of chloritic strata that formed as a result of seeps of Mg-richbrines during the period that spanned the deposition of the Naiset,Cathedral, Burgess Shale, and Eldon formations, and these Mg-rich

Table 4Composition of calcite and dolomite in herringbone calcite from the CathedralFormation, WDS analysis.

CaO MgO MnO FeO NiO ZnO SrO BaO PbO

wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.% wt.%

CA1 61.523 0.323 0.000 0.000 b0.1 b0.2 b0.1 b0.1 b0.2CA2 56.605 0.369 0.000 0.000 b0.1 b0.2 b0.1 b0.1 b0.2DA1 34.625 21.050 0.000 0.048 b0.1 b0.2 b0.1 b0.1 b0.2DA2 33.700 20.478 0.000 0.022 b0.1 b0.2 b0.1 b0.1 b0.2CB1 57.759 0.332 0.026 0.000 b0.1 b0.2 b0.1 b0.1 b0.2CB2 59.444 0.298 0.024 0.051 b0.1 b0.2 b0.1 b0.1 b0.2CB3 62.023 0.367 0.000 0.000 b0.1 b0.2 b0.1 b0.1 b0.2CB4 61.221 0.673 0.013 0.028 b0.1 b0.2 b0.1 b0.1 b0.2DB1 34.552 21.313 0.000 0.000 b0.1 b0.2 b0.1 b0.1 b0.2DB2 33.441 20.790 0.017 0.030 b0.1 b0.2 b0.1 b0.1 b0.2DB3 33.714 21.498 0.040 0.097 b0.1 b0.2 b0.1 b0.1 b0.2DB4 31.713 20.922 0.022 0.277 b0.1 b0.2 b0.1 b0.1 b0.2DB5 32.027 21.133 0.031 0.250 b0.1 b0.2 b0.1 b0.1 b0.2DB6 33.567 20.155 0.000 0.198 b0.1 b0.2 b0.1 b0.1 b0.2DB7 32.502 20.247 0.078 0.559 b0.1 b0.2 b0.1 b0.1 b0.2DB8 34.552 21.313 0.000 0.000 b0.1 b0.2 b0.1 b0.1 b0.2DB9 33.011 21.233 0.000 0.011 b0.1 b0.2 b0.1 b0.1 b0.2DB10 34.461 21.113 0.043 0.229 b0.1 b0.2 b0.1 b0.1 b0.2

137W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

brines flowed through porous rocks such as the Cathedral limestones(Powell et al., 2006). Mg-rich hydrothermal activity associated withseafloor deposition of dololaminites and early dolomitizationwithin theCathedral and Eldon formations has also been documented by Moore(1994) and Jeary (2002). Regionally extensive evidence for the flow ofMg-rich brines from the subsurface during deposition of the Cathedraland Eldon formations, and the Mg-rich nature of herringbone calcitewithin these units suggests that Mg2+ was the calcite-inhibitor ion thatwas responsible for the development of herringbone cements insoutheastern British Columbia.

Fig. 7. Features of authigenic quartz crystals. A) Abundant quartz crystals in dolomitized mudSouth; coin is 19 mm in diameter. B) Photomicrograph of quartz crystal overgrowing Epipcrystals in alteration selvage of fracture in Greenmount Mbr. D) BSE image of assemblagecrystals in Greenmount Mbr. F) BSE showing zoning of quartz crystals in Greenmount Mbrdolomite, Ms = muscovite, Py = pyrite, and Qtz = quartz.

Pratt (2002) attributed the development of herringbone-calcite-cemented cavities to seismic activity along the subsiding KickingHorse Rim. Intermittent seismic activity in the region not onlyaccounts for the formation of large cavities, but also provides amechanism for the coeval pumping of subsurface brines (seismicpump and fault valvemechanisms of Sibson, 2001). In the near surfaceenvironment in which the cavities opened, the warm brine wouldhave been cooled and diluted by mixing with seawater. As calciteprecipitated, the solution became increasingly magnesian whichresulted in higher mol% MgCO3 of the precipitating calcite as the

mound associated with Mg-rich brine seep deposits, Vermilion Subunit, Haiduk Cirquehyton in dolomitized mud mound Vermilion Subunit, Haiduk Cirque South. C) Quartzin silicified fracture, Greenmount Mbr. E) Photomicrograph showing zoning of quartzwith inclusions of calcite and pyrite. Abbreviations: Ap = apatite, Cc = calcite, Do =

138 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

band thickened. An influx of a new pulse of seawater would haveresulted in the initiation of a new band of calcite with a lower Mgcontent. During diagenesis the high-Mg calcite converted to low-Mgcalcite with dolomite inclusions; heterogeneity of mol% MgCO3 acrossbands of herringbone calcite can account for the heterogeneousdistribution of dolomite micro-inclusions.

The similarity between the chemical composition and microtex-tures of herringbone calcite, as well as the physical form of host-cavities and associated carbonate mesoscopic textures, suggests acommon origin.

5.2. Authigenic quartz clusters

Stose and Stose (1944) noted dense nests of needle-like crystals ofsmoky quartzwithin limestone of the Ledger Fm in the south quarry of J.E. Baker Company. Dark, hollow, elongate quartz crystals with euhedralgrowth laminae were also documented in limestone beds of theGreenmount Mbr (Taylor and Durika, 1990). Field examination of theoccurrence of quartz crystals in the Greenmount Mbr during this studydocumented that they occur in alteration selvages that surround darkgrey, foliated, anastomosing fractures (Fig. 7C). The dark grey fracturebands are composed of a matrix of fine grained calcite withinwhich areset patches muscovite, quartz, dolomite, pyrite and fluorapatite(Fig. 7D), consistent with the whole-rock major element composition(Table 5). Muscovite is fine-grained (b50 μm), displays a felted textureand contains an average of 0.25 wt.% F. Quartz crystals within the darkgrey fractures are euhedral, lack zoning, are less than 50 μm long, andinclusions of calcite, dolomite and pyrite are common. Apatite formsanhedral, spongy clusters up to 100 μm across. Pyrite occurs in clusters.

Outside of the dark grey fractures, quartz crystals are much larger(up to 6 mm long), and occur in the medium-grey limestone thatimmediately surrounds the fractures. They comprise up to 15% of therock. Quartz crystals are randomly oriented. Growth zoning is welldeveloped and defined by concentric inclusion-rich and inclusion-poor bands (Fig. 7E and F). Calcite is the most abundant inclusion, butpyrite inclusions are also common (Fig. 7D). Muscovite crystals up to200 μm long, are disseminated through the quartz-bearing limestone.Muscovite contains an average of 0.57 wt.% F.

The hollow quartz crystals of the Greenmount Mbr have beenattributed to the reconstitution of siliceous sponge spicules (Taylorand Durika, 1990). However, sponge spicules are rare to absent in theassociated limestone. Furthermore, the direct association of quartzcrystals with the selvages of fractures, and a fluorine-rich mineralassemblage of dolomite-muscovite-apatite-pyrite indicates that thequartz crystals are related to fracture-controlled brine flow.

Similar clusters of abundant 10 mm long euhedral, doubly-terminated quartz crystals that have a hollow appearance in outcrop(Fig. 7A) occur in the Cathedral and Eldon formations along theKicking Horse Rim (Simandl and Hancock,1991; Powell et al., 2006) inassociation with dolomitic alteration haloes of magnesian ore bodies(talc and magnesite). They also occur in dolomites of the Cathedral,Takakkaw Tongue, Eldon formations, and the Vermilion subunit inadjacent to Mg-rich, chloritic strata and feeder veins associated withthe seepage of brines (Powell et al., 2006).

Table 5Composition of quartz crystal-bearing carbonate rocks, ICPMS analysis.

SiO2 TiO2 Al2O3 Fe2O3 MnO M

wt.% wt.% wt.% wt.% wt.% w

Eldon Qtz 1 2.78 0.04 0.42 0.90 0.06 20Cath Qtz 1 9.10 0.05 1.57 0.51 0.04 19Cath Qtz 2 4.09 b0.01 0.19 0.63 0.10 21Kinz Qtz 1 22.83 0.84 10.75 1.42 0.05 1Kinz Qtz 2 32.88 0.75 11.48 1.78 0.04 3Kinz Qtz 3 24.98 0.89 13.93 1.5 0.08 3

As described by Powell et al. (2006), the quartz crystals aredistinctly zoned and have overgrown, and partly replaced, dolostonehost rock (Fig. 7B), including primary microbial structures. In anygiven crystal, all zones of quartz are in optical continuity. Rhombohe-dral dolomite inclusions occur within quartz crystals. These texturesindicate that the quartz crystals grew episodically by replacement of adolomitized limestone host. Fine-grained pyrite is a minor componentof the quartz-bearing dolostones. A predominantly dolomitic rockwith quartz, and minor pyrite is consistent with major elementcomposition (Table 5).

Within the Vermilion subunit, several quartz-bearing clasts occurin a slump breccia on the flank of a quartz-bearing carbonate mudmound, indicating that, at least locally, quartz crystals grew at orimmediately below the seafloor. In contrast, quartz crystals associatedwith talc pods on Mt. Stephen occur within the selvages of veinletsthat emanate from talc bodies.

The direct association of quartz euhedra with highly porous orfractured carbonate rocks that are adjacent to seep-related magnesianstrata andMg-rich ore bodies led Powell et al. (2006) to conclude thatMg-rich brine was essential to the formation of high volume of quartzeuhedra observed. Fabricius (1984) documented another correlationbetween Mg-rich brines and euhedral quartz crystals in Permian saltdeposits of the Danish Trough. In these salt beds, quartz crystals up to1.7 mm long are abundant within carnallite-bearing strata (KMgCl3 6-H2O), whereas Mg-poor salt horizons lack such crystals. Fluidinclusions in these quartz crystals commonly contain highly concen-trated MgCl2 solutions (Fabricius, 1987).

Zoned, euhedral quartz crystals from the Jurassic Zuloaga Fm ofMexico are illustrated by Scholle andUlmer-Scholle (2004), p.141). Thisunit is underlain by the Minas Viejas evaporite. Chafetz and Zhang(1998) noted the association of authigenic, doubly-terminated, euhedralquartz crystals in Pleistocene sabkha dolomites of the Arabian Gulf inassociationwith sulphates. Thus various compositions of brinemay leadto the development of authigenic quartz.

Whereas there is evidence to link quartz crystal growth in theKinzers Fm to chemically reactive brines, the lack of extensivedolomitization in the quartz-bearing host-rock, and themore complexalteration assemblage indicates that the composition of the fluid musthave differed from the extreme Mg-rich brines of the Kicking HorseRim and Danish examples. A source for the silicifying brines of thePennsylvania may have been the Early to Middle Cambrian Rome Fm,which consists of redbeds that accumulated in response to upliftassociated with rifting of the Rome Trough to the west of theConestoga Valley (Read, 1989). This unit consists mainly of sand-stones, siltstone, and shale (Samman, 1975), but silicified evaporitesoccur in outcrop (Read, 1989), and evaporites are known to lie in thesubsurface (Read, 1989; Raymond, 1990). Migration of evaporite-related brines through the coeval Kinzers and Ledger formations, mayhave led to the development of quartz crystals by means similar tothat of the Zuloaga Fm. Future analysis of fluid inclusions in Kinzersand Ledger quartz crystals may help to better constrain the origin ofthe brines responsible for quartz growth.

Pope and Grotzinger (2003, p. 289) documented zoned quartzcrystals in the Palaeoproterozoic Stark Fm of Canada's Northwest

gO CaO Na2O K2O P2O5 Cr2O3 LOI

t.% wt.% wt.% wt.% wt.% wt.% wt.%

.99 27.89 b0.01 0.10 b0.01 b0.01 45.30.72 28.72 0.03 0.27 0.01 b0.01 38.90.08 30.89 0.03 0.02 b0.01 b0.01 43.10.04 32.10 0.04 3.36 0.33 b0.01 26.8.43 22.02 0.04 3.47 0.50 b0.01 21.7.92 24.15 0.05 4.30 0.82 b0.01 24.7

139W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

Territories that appear remarkably similar to those of the CambrianLaurentian examples. These quartz crystals occur in dolomite pseudo-morphs after halite, and have been interpreted to have grown very soonafter deposition through the introduction of silica with saline brines(Pope and Grotzinger, 2003). These brine-related quartz crystals of theStark Fmoccur in the unit that stratigraphically overlies anoccurrence ofherringbone calcite-bearing Neptunian dykes in the inner platformalfacies limestones of the Pethei Group. Thuswehave at least threewidelyspaced North American examples in which there exists a spatialassociation between zoned quartz crystals and herringbone calcitecements. This correlation reinforces the conclusion that brinesmay playa role in the formation of these enigmatic cements.

6. Conclusions

1. Redox-sensitive trace elements indicate that palaeo-oxygenationconditions varied with respect to faunal assemblage and location.Ogygopsis-bearing fauna in both the Burgess Shale and Kinzersformation are associated with a dysoxic metal signature. BurgessShale-type preservation at the 22L site of the Kinzers Fm wasdeposited in a fluctuating oxycline, whereas the BST-fossil beds ofFossil Ridge and Mt Stephen yield a geochemical signature thatindicates sustained oxic conditions. Thus BST-preservation does notcorrelate with specific dissolved oxygen content of bottom waters,and sustained low oxygen levels do not correlate with increasedqualityof fossil preservation in either theBurgess Shale or theKinzersformations.

2. Herringbone calcite cements occur in microbial carbonate reefs informations that are associated with both the Burgess Shale andKinzers formations. The herringbone calcite occurs in large, generallyconcordant, near-surface cavities lined with large shrubby microbialencrustations. Lack of evidence of anoxia in the immediate deposi-tional environment, lack of enrichment of distinct calcite inhibitorions such as Fe, Sr, S, Pb, and Zn in the herringbone cements, andabundance of dolomite inclusions in both localities, and extensivefield evidence forMg-rich brine seeps in the British Columbia sectionsuggest that Mg-enrichment of pore water was responsible for thedevelopment of herringbone calcite in these localities.

3. Doubly-terminated, euhedral, authigenic quartz occurs in associa-tion with fractures in the Kinzers, Ledger, Cathedral, and Eldonformations, and in association with Mg-rich orebodies and brine-pool deposits in the British Columbian units. These texturallydistinct quartz crystals, along with halogen-bearing minerals(chlorite, muscovite and apatite), are evidence of brine flowthrough these Cambrian units. Variation in the mineral assem-blages associated with the quartz crystals in the Cathedral andKinzers formations indicate that the composition of the brine wasdifferent at each site.

4. The occurrence of both herringbone calcite and zoned quartz crystalswithin units that are associated with the Burgess Shale and Kinzersformations suggest that brines may represent a common feature inthe depositional and diagenetic environment of these two regions inwhich Burgess Shale-type fossil preservation occur. The greaterabundance of both quartz crystals and herringbone calcite cement inthe Canadian locality suggests that brine flow was more widespreadin the Burgess Shale than in the Kinzers Formation. Brine seeps in thedepositional environmentof theseunitsmayhaveprovidednutrientsfor stable high-abundance faunal communities in a slope setting (e.g.,Johnston et al., 2009-this volume), thereby accounting for theabundance of fossil material. Furthermore, high-salinity brine inpores waters of carcass-bearing Cambrian muds may have furtherlimited shallow burrowing, scavenging, and decomposition rates, allof which are factors that have been cited repeatedly as essential topreservation in Burgess Shale-type deposits (e.g., Allison and Briggs,1993b; Orr et al., 2003), and so may account for the degree ofpreservation in these fossiliferous units.

5. Texturally distinct minerals such as herringbone calcite and zonedauthigenic quartz crystals, and/or chemically distinct mineralssuch as halogen-bearing phyllosilicates, are not present at all BST-deposits. For example, no such features have been documented inthe fossil-bearing strata of Utah's Marjum andWheeler formations,or associated units. The presence of such minerals may provide ameans to help identify fossil deposits in which brines play asignificant preservational role, from thosewhere other factors (e.g.,fluctuating oxycline) may have led to exceptional preservation.

Acknowledgements

Funding for this project was provided by a National Science andResearch Council of Canada grant to P.A. Johnston and two PSC-CUNYgrants to W. Powell. Thanks go to Parks Canada for permitting thisstudy within Yoho National Park. Thanks to the colleagues whoprovided advice and discussion: R. Thomas, C. de Wet, P. Johnston, C.Collom, and K. Johnston. Invaluable assistance in obtaining themicroprobe analyses reported in this paper was provided by C.Mandeville and N. Gitahi of the Earth and Planetary ScienceDepartment of the American Museum of Natural History. My thanksto J. Zalasiewicz and an anonymous reviewer for their the reviews ofthe original manuscript.

References

Aitken, J.D., 1971. Control on lower Palaeozoic sedimentary facies by the Kicking HorseRim, southern Rocky Mountains, Canada. Bulletin of Canadian Petroleum Geology19, 557–569.

Aitken, J.D., 1997. Stratigraphy of the Middle Cambrian platformal succession, southernRocky Mountains. Geological Survey of Canada Bulletin 398 322 pp.

Algeo, T.J., Maynard, J.B., 2004. Trace-element behavior and redox facies in core shales ofUpper Pennsylvanian Kansas-type cyclothems. Chemical Geology 206, 289–318.

Allison, P.A., 1988. The role of anoxia in the decay and mineralization of proteinaceousmacro-fossils. Palaeobiology 14, 139–154.

Allison, P.A., Briggs, D.E.G., 1993a. Exceptional fossil record: distribution of soft-tissuepreservation through the Phanerozoic. Geology 21, 527–530.

Allison, P.A., Briggs, D.E.G., 1993b. Burgess Shale biotas; burrowed away? Lethaia 26,225–229.

Babcock, L.E., Zhang, W., Leslie, S.A., 2001. The Chengjiang biota: record of the EarlyCambrian diversification of life and clues to exceptional preservation of fossils. GSAToday 11, 4–9.

Barnaby, R.J., Read, J.F., 1990. Carbonate ramp to rimmed shelf evolution: Lower toMiddle Cambrian continental margin, Virginia Appalachians. Geological Society ofAmerica Bulletin 102, 391–404.

Bond, G.C., Kominz, M.A., 1984. Construction of tectonic subsidence curves for the earlyPalaeozoic miogeocline, southern Canadian Rocky Mountains: implications forsubsidence mechanisms, age of breakup and crustal thinning. Geological Society ofAmerica Bulletin 95, 155–173.

Bond, G.C., Christie-Blick, N., Kominz, M.A., Devlin, W.J., 1985. An Early Cambrian rift topost-rift transition in the Cordillera of western North America. Nature 316, 742–745.

Briggs, D.E., 1978. A new trilobite arthropod from the Lower Cambrian KinzersFormation, Pennsylvania. Journal of Palaeontology 52, 132–140.

Butterfield, N.J., 1990. Organic preservation of non-mineralizing organisms and thetaphonomy of the Burgess Shale. Palaeobiology 16, 272–286.

Calvert, S.E., Pederson, T.F., 1993. Geochemistry of recent oxic and anoxic marinesediments; implications for the geological record. Marine Geology 113, 67–88.

Campbell, L., 1969. Stratigraphy and Palaeontology of the Kinzers Formation, South-eastern Pennsylvania. MS Thesis, Franklin and Marshall College, Lancaster,Pennsylvania, 51 pp.

Campbell, L.D., 1971. Occurrence of “Ogygopsis shale” fauna in southeastern Pennsylva-nia. Journal of Palaeontology 45, 437–440.

Capdevila, D., Conway Morris, S., 1999. New fossil worms from the Lower Cambrian ofthe Kinzers Formation, Pennsylvania, with some comments on Burgess Shale-typepreservation. Journal of Palaeontology 73, 394–402.

Chafetz, H.S., Zhang, J., 1998. Authigenic euhedral megaquartz crystals in a quaternarydolomite. Journal of Sedimentary Research 68, 994–1000.

de Wet, C.B., Hopkins, D., 2006. Ooid shoals, microbialite reefs, and shoal shorefacesands, the Ledger Formation, LWB Refractories site, York County, Pennsylvania.Geological Society of America Northeastern Section, 41st Annual Meeting. Field TripGuides, pp. 1–26.

de Wet, C.B., Dickson, J.A.D., Wood, R.A., Gaswirth, S.B., Frey, H.M., 1999. A new type ofshelf margin deposit: rigid microbial sheets and unconsolidated grainstonesriddled with metre-scale cavities. Sedimentary Geology 128, 13–21.

de Wet, C.B., Frey, H.M., Gaswirth, S.B., Mora, C.I., Rahnis, M., Bruno, C.R., 2004. Origin ofmeter-scale submarine cavities and herringbone calcite cement in a Cambrianmicrobial reef, Ledger Formation (USA). Journal of Sedimentary Research 74,914–923.

140 W. Powell / Palaeogeography, Palaeoclimatology, Palaeoecology 277 (2009) 127–140

Devlin,W.J., Bond, G.C.,1988. The initiation of the early Palaeozoic Cordilleranmiogeocline:evidence from the uppermost Proterozoic–Lower Cambrian Hamill Group of south-eastern British Columbia. Canadian Journal of Earth Sciences 25, 1—19.

Fabricius, J., 1984. Formation temperature and chemistry of brine inclusions in euhedralquartz crystals from Permian salt in the Danish Trough. Bulletin deMineralogie 107,203–216.

Fabricius, J., 1987. Natural Na–K–Mg–Cl solutions and solid derivatives trapped ineuhedral quartz from Danish Zechstein salt. Chemical Geology 61, 95–112.

Faill, R.T., 1999. Palaeozoic. The Geology of Pennsylvania. Pennsylvania GeologicalSurvey Special Publication, vol. 1.

Fletcher, T.P., Collins, D.H., 1998. TheMiddle Cambrian Burgess Shale and its relationshipto the Stephen Formation in the southern Canadian Rocky Mountains. CanadianJournal of Earth Science 35, 413–436.

Fletcher, T.P., Collins, D.H., 2003. The Burgess Shale and associated Cambrian formationswest of the Fossil Gully fault zone on Mount Stephen, British Columbia. CanadianJournal of Earth Science 40, 1823–1838.

Fritz, W.H., 1971. Geological setting of the Burgess Shale: North American Paleontolo-gical Convention. Chicago, 1969, Proceedings Part I, pp. 1155–1170.

Fritz, W.H., Cecile, M.P., Norford, B.S., Morrow, D., Geldsetzer, H.H., 1991. Cambrian toMiddle Devonian assemblages. In: Gabrielse, H., Yorath, C.J. (Eds.), Geology of theCordilleran Orogen in Canada. Geological Survey of Canada, Geology of Canada,vol. 4, pp. 151–218.

Gabbott, S.E., 1998. Taphonomy of the Ordovician Soom Shale Lagerstätten: an exampleof soft tissue preservation in clay minerals. Palaeontology 41, 631–667.

Gaines, R.R., Droser, M.L., 2003. Palaeoecology of the familiar trilobite Elrathia kingii: anearly exaerobic zone inhabitant. Geology 31, 941–944.

Gaines, R.R., Kennedy, M.J., Droser, M.L., 2005. A new hypothesis for organicpreservation of Burgess Shale taxa in the Middle Cambrian Wheeler Formation,House Range, Utah. Palaeogeography, Palaeoclimatology, Palaeoecology 220,193–205.

Gohn, G.S., 1976. Sedimentology, Stratigraphy, and Palaeogeography of LowerPalaeozoic Carbonate Rocks, Conestoga Valley, Southeastern Pennsylvania. Unpub-lished Ph.D. Thesis, University of Delaware, 315 pp.

Hofman, H.J., Thurston, P.C.,Wallace, H.,1985. Archean stromatolites fromUchiGreenstoneBelt, Northwestern Ontario. Evolution of Archean Supracrustal Sequences. GeologicalAssociation of Canada Special Paper, vol. 28, pp. 1125–1132.

Jeary, V.A., 2002. The origin of Middle Cambrian-Hosted Mid-Platform Dolomite;Cathedral Formation, Whirlpool Point, Alberta. Unpublished M.Sc. Thesis, Uni-versity of Calgary, Calgary, 205 pp.

Johnston, P.A., Collom, C.J., 2001. Ancient seeps in the Middle Cambrian Burgess Shale;paleoecological and evolutionary implications. Earth System Processes; Programmeswith Abstracts, Geological Society of America and Geological Society of London, June24–28, 2001. Edinburgh, p. 91.

Johnston, K.J., Johnston, P.A., Powell, W.G., 2009. Palaeontology and depositionalenvironments of ancient brine seeps in the Middle Cambrian Burgess Shale at TheMonarch, British Columbia, Canada. Palaeogeography, Palaeoclimatology, Palaeoe-cology 277, 86–105 (this volume).

Jones, B., Manning, D.A.C., 1994. Comparison of geochemical indices used for theinterpretation of palaeoredox conditions in ancient mudstones. Chemical Geology111, 111–129.

Kimura, H., Watanabe, Y., 2001. Ocean anoxia at the Precambrian–Cambrian boundary.Geology 29, 995–998.

Kubli, T.E., Simony, P.S., 1992. The Dogtooth High, northern Purcell Mountains, BritishColumbia. Bulletin of Canadian Petroleum Geology 40, 36–51.

Lehman, P.J., 1978. Deposition, Porosity Evolution and Diagenesis of the Pipe Creek Jr.Reef (Silurian) Grant County, Indiana. Unpublished Masters Thesis, University ofWisconsin–Madison.

McIlreath, I.A., and Aitken, J.D., 1976. Yoholaminites (Middle Cambrian), problematicalcalcareous sediment-stabilizing organism. Program with Abstracts — GeologicalAssociation of Canada; Mineralogical Association of Canada; Canadian GeophysicalUnion, Joint Annual Meeting 1, p. 84.

Meyer, H.J., 1984. The influence of impurities on the growth rate of calcite. Journal ofCrystal Growth 66, 639–646.

Moore, S.L.O., 1994. The Origin of Dolomite of the Middle Cambrian Eldon and Pikaformations in the Yoho Glacier area, Yoho National Park, British Columbia.Unpublished M.Sc. Thesis, University of Calgary, Alberta, Canada, 268 pp.

Motz, K.J., Johnston, P.A., Collom, C.J., Krause, F.F., 2001. New Burgess Shale-like biotasfrom the Upper Middle Cambrian Duchesnay unit (Chancellor Group) at HaidukPeak and Miller Pass, British Columbia. Canadian Palaeontological Conference,London, Ontario. Program with Abstracts, vol. 11, pp. 33–37.

Orr, P.J., Benton, M.J., Briggs, D.E.G., 2003. Post-Cambrian closure of the deep-waterslope-basin taphonomic window. Geology 31, 769–772.

Pope, M.C., Grotzinger, J.C., 2003. Palaeoproterozoic Stark Formation, AthapuscowBasin,Northwest Canada: record of cratonic-scale salinity crisis. Journal of SedimentaryResearch 73, 280–295.

Powell, W.G., Johnston, P.J., Collom, C.J., 2003. Geochemical evidence for oxygenatedbottom waters during deposition of fossiliferous strata of the Burgess ShaleFormation. Palaeogeography, Palaeoclimatology, Palaeoecology 201, 249–268.

Powell, W.G., Johnston, P.J., Collom, C.J., Johnston, K.J., 2006. Middle Cambrian brineseeps on the Kicking Horse Rim and their relationship to talc and magnesitemineralization, and associated dolomitization, British Columbia, Canada. EconomicGeology 101, 431–451.

Pratt, B.R., 2002. Tepees in peritidal carbonates: origin via earthquake-induced deforma-tion, with example from the Middle Cambrian of western Canada. SedimentaryGeology 153, 57–64.

Raymond, D.E., 1990. Evaporites in the Cambrian Rome Formation; new evidence fromthe shallow subsurface in Bibb County, Alabama. Geological Society of America,Southeast Section 39th Annual Meeting. Abstracts with Programs, vol. 22, p. 58.

Read, J.F., 1989. Evolution of Cambro–Ordovician passive margin, U.S. Appalachians. In:Hatcher, R.D., Thomas, W.A., Viele, G.W. (Eds.), The Geology of North America,Volume F-2, The Appalachian–Ouachita Orogen in the United States.

Rodgers, J., 1968. The eastern edge of the North American continent during theCambrian and EarlyOrdovician. Studies of Appalachian Geology. Northern andMaritime, New York, Wiley-Interscience, pp. 141–149.

Samman, N.F., 1975. Sedimentation and Stratigraphy of the Rome Formation in EastTennessee. Unpublished Ph.D. Thesis, University of Tennessee, Knoxville, p. 337.

Scholle, P.A., Ulmer-Scholle, D.S., 2004. Color guide to the petrography of carbonaterocks: grains, textures, porosity, Diagenesis. American Association of PetroleumGeologists Memoir 474 pp.

Sibson, R.H., 2001. Seismogenic framework for hydrothermal transport and oredeposition. In: Richards, J.P., Tosdal, R.M. (Eds.), Structural Controls On Ore Genesis.Reviews in Economic Geology, vol. 14, pp. 25–50.

Simandl, G.J., Hancock, K.D., 1991. Geology of the Mount Brussilof magnesite deposit,southeastern British Columbia. Geological Fieldwork 1990, A Summary of FieldActivities and Current Research. Province of British Columbia, Ministry of Energy,Mines and Petroleum Resources, Paper, vol. 1991–1, pp. 269–278.

Simonson, B.M., Schubel, K.A., Hassler, S.W., 1993. Carbonate sedimentology of the earlyPrecambrian Hamersley Group of Western Australia. Precambrian Research 60,287–335.

Skinner, E.S., 2005. Taphonomy and depositional circumstances of exceptionallypreserved fossils from the Kinzers Formation (Cambrian), southeastern Pennsyl-vania. Palaeogeography, Palaeoclimatology, Palaeoecology 220, 167–192.

Stewart, W.D., 1991. Stratigraphy and Sedimentology of the Chancellor Succession(Middle and Upper Cambrian), Southeastern Canadian Rocky Mountains. Unpub-lished Ph.D. Thesis, University of Ottawa, Canada, 583 pp.

Stewart, W.D., Dixon, O.A., Rust, B.R., 1993. Middle Cambrian carbonate-platformcollapse, southeastern Canadian Rocky Mountains. Geology 21, 687–690.

Stose, G.W., Stose, A.J., 1944. Geology of the Hanover–York district, Pennsylvania. UnitedStates Geological Survey, Professional Paper 204 84 pp.

Sumner, S.Y., Grotzinger, J.P., 1996. Herringbone calcite: petrography and environmentalsignificance. Journal of Sedimentary Research 66, 419–429.

Taylor, S.R., McLennen, S.M., 1985. The Continental Crust — Its Composition andEvolution. Blackwell Scientific Publication, Oxford, England.

Taylor, J.F., Durika, N.J., 1990. Lithofacies, trilobite faunas, and correlation of the Kinzers,Ledger and Conestoga formations in the Conestoga Valley. In: Scharnberger, C.K.(Ed.), Carbonates, Schists, and Geomorphology in the Vicinity of the SusquehannaRiver: Guidebook for the 55th Annual Field Conference of Pennsylvania Geologists,pp. 136–155.

Taylor, J.F., Brezinski, D.K., Durika,N.J.,1997. The Lower toMiddleCambrian transition in thecentral Appalachian region. Second International Trilobite Conference, BrockUniversity, St. Catherines, Ontario, August 22–25, 1997, Abstracts with Program, p. 48.

Tourre, S.A., 2000. Cave-filling Herringbone Calcite; Morphology and Geochemistry ofan Unusual Carbonate Cement from Egypt. Unpublished Masters Thesis, Universityof California at Davis.

Tourre, S.A., Sumner, D.Y., 1999. Cave-filling herringbone calcite from a middle Eocenecarbonate platform, Egypt; crystal morphology and geochemistry of an unusualcarbonate cement. Abstracts with Programs — Geological Society of America,Annual Meeting, 31, p. 280.

Vine, J.D., Tourtelot, E.B., 1970. Geochemistry of black shale deposits — a summaryreport. Economic Geology 65, 253–272.