87Sr/86Sr, Î´13C and Î´18O evolution of Phanerozoic seawater · 2016. 12. 2. · volume Hayes...

Ž .Chemical Geology 161 1999 59–88www.elsevier.comrlocaterchemgeo

87Srr86Sr, d13C and d18O evolution of Phanerozoic seawater

Jan Veizer a,b,), Davin Ala b,c, Karem Azmy b, Peter Bruckschen a, Dieter Buhl a,´Frank Bruhn a,d, Giles A.F. Carden a,e, Andreas Diener a,f, Stefan Ebneth a,g,Yves Godderis b,h, Torsten Jasper a, Christoph Korte a, Frank Pawellek a,

Olaf G. Podlaha a,i, Harald Strauss a,ja Institut fur Geologie, Ruhr UniÕersitat, Bochum 44780, Germany¨ ¨

b Ottawa-Carleton Geoscience Centre, UniÕersity of Ottawa, Ottawa, ON, Canada K1N 6N5c 17A Street SW, Calgary, AB, Canada T3T 4S4

d Leibniz-Labor fur Altersbestimmung und Isotopenforschung, Christian-Albrechts UniÕersitat, 24158 Kiel, Germany¨ ¨e Research and DeÕelopment SerÕice Offices, The UniÕersity of Warrick, CoÕentry CV4 7AL, UK

f Fuhlrott Museum, Auer Schulstrasse 20, Wuppertal-Elberfeld 42103, Germanyg F.D. Zublin, Mathias-Bruggen Strasse 79, Cologne 50827, Germany¨ ¨

h Laboratoire de Physique Atmospherique et Planetaire, UniÕersite de Liege, Liege 4000, Belgium´ ´ ´ ` `i Shell International Exploration and Production, PO Box 60, 2280 AB Rijswijk, Netherlands

j Geologisch-Palaoutologiselies Institut und Museum, Westphalische Wilhelms-UniÕersitat, 48149 Munster, Germany¨ ¨ ¨ ¨

Received 11 December 1997; accepted 12 October 1998

Abstract

A total of 2128 calcitic and phosphatic shells, mainly brachiopods with some conodonts and belemnites, were measuredfor their d18O, d13C and 87Srr86Sr values. The dataset covers the Cambrian to Cretaceous time interval. Where possible,

Ž .these samples were collected at high temporal resolution, up to 0.7 Ma one biozone , from the stratotype sections of allcontinents but Antarctica and from many sedimentary basins. Paleogeographically, the samples are mostly from paleotropi-

Ž .cal domains. The scanning electron microscopy SEM , petrography, cathodoluminescence and trace element results of theŽ .studied calcitic shells and the conodont alteration index CAI data of the phosphatic shells are consistent with an excellent

preservation of the ultrastructure of the analyzed material. These datasets are complemented by extensive literaturecompilations of Phanerozoic low-Mg calcitic, aragonitic and phosphatic isotope data for analogous skeletons. The oxygen

18 Ž .isotope signal exhibits a long-term increase of d O from a mean value of about y8‰ PDB in the Cambrian to a presentŽ .mean value of about 0‰ PDB . Superimposed on the general trend are shorter-term oscillations with their apexes coincident

with cold episodes and glaciations. The carbon isotope signal shows a similar climb during the Paleozoic, an inflexion in thePermian, followed by an abrupt drop and subsequent fluctuations around the modern value. The 87Srr86Sr ratios differ fromthe earlier published curves in their greater detail and in less dispersion of the data. The means of the observed isotope

87 86 18 13 34 Ž .signals for Srr Sr, d O, d C and the less complete d S sulfate are strongly interrelated at any geologicallyŽ .reasonable 1 to 40 Ma time resolution. All correlations are valid at the 95% level of confidence, with the most valid at the

99% level. Factor analysis indicates that the 87Srr86Sr, d18O, d13C and d 34S isotope systems are driven by three factors.

) Corresponding author. Department of Earth Sciences, Ottawa-Carleton Geoscience Centre, University of Ottawa, Ottawa, ON, CanadaK1N 6N5. Tel.: q1-613-562-5800; fax: q1-613-562-5192; E-mail: [email protected]

0009-2541r99 q 1999 Elsevier Science B.V.Ž .PII: S0009-2541 99 00081-9

Open access under CC BY-NC-ND license.

http://creativecommons.org/licenses/by-nc-nd/3.0/

( )J. Veizer et al.rChemical Geology 161 1999 59–8860

The first factor links oxygen and strontium isotopic evolution, the second 87Srr86Sr and d34S, and the third one the d13Cand d34S. These three factors explain up to 79% of the total variance. We tentatively identify the first two factors as tectonic,

Ž . Žand the third one as a biologically mediated redox linkage of the sulfur and carbon cycles. On geological timescales G1. Ž .Ma , we are therefore dealing with a unified exogenic litho-, hydro-, atmo-, biosphere system driven by tectonics via its

Ž .control of bio geochemical cycles. q 1999 Elsevier Science B.V.

Keywords: Seawater; Phanerozoic; Isotopes; Strontium; Oxygen; Carbon

1. Introduction

The Sr isotopic composition of past seawater, asmonitored by marine carbonate shells and rocks, canbe utilized as a proxy parameter for tectonic evolu-tion of the Earth System, because the 87Srr86Srvariations reflect principally the waxing and waning

Ž .of Sr input from rivers ‘continental’ flux vs. theinput from the submarine hydrothermal systemsŽ . Ž‘mantle’ flux Faure, 1986; Taylor and Lasaga, this

. 87 86issue . The general trends of Srr Sr variations forPhanerozoic seawater have been sketched already in

Ž .the pioneering studies of Peterman et al. 1970 ,Ž . Ž .Veizer and Compston 1974 and Burke et al. 1982 .

The published ‘curve’ of the last authors servedsubsequently as a baseline for all follow-up studies.The last decade, however, witnessed a considerableadvance in instrumental precision, sample prepara-tion and in better geological constraints, particularlyin stratigraphic assignment of samples. This resultedin a considerably tighter delineation of the secularmarine isotope trend and opened the stage for utiliza-tion of Sr isotopes as a correlation and dating tool,‘isotope stratigraphy’. This stage of research has

Ž .been summarized in DePaolo and Ingram 1985 ,Ž . Ž . Ž .Elderfield 1986 , Veizer 1989 , McArthur 1994

Ž .and Smalley et al. 1994 . Nevertheless, the utility ofthis new tool has been restricted mostly to the Ceno-zoic, because of the steep and well-defined rise inmarine 87Srr86Sr ratio. Apart from the steep slope,the major reason for delineation of such a well-de-fined trend has been the fact that this young portionof the geologic record has a relatively well-resolvedand detailed stratigraphy and that sediments of thisage contain shells that were only marginally affectedby the vicissitudes of diagenesis.

In theory, it should be possible to define similarwell-constrained trends also for other Phanerozoic

intervals with equally steep slopes and thus extendthe utility of the ‘isotope stratigraphy’ into the

Ž .Mesozoic and Paleozoic Veizer et al., 1997a . Theconstraining factors in achieving this goal are not somuch the instrumental capabilities, because for manylaboratories intra- and interlaboratory reproducibility

y6 Žroutinely falls within "25=10 limits Thirwall,.1991; McArthur, 1994 , but the preservation and

resolution of the geologic record.13 ŽThe d C of ancient oceans e.g., Clayton and

Degens, 1959; Schidlowski et al., 1975; Veizer and.Hoefs, 1976 was for a long time regarded as essen-

tially invariant, but with a considerable noise, around0‰ PDB. Only in the 1980s was it realized that thisnoise contains a real oscillating secular signalŽScholle and Arthur, 1980; Veizer et al., 1980; Arthur

.et al., 1985; Shackleton, 1985 . In contrast to thebelow-discussed oxygen isotopes, the primary natureof the d13C secular trend has not been seri-carbonateously challenged. The reason is that diagenetic re-crystallization of carbonates is accomplished in asystem with a low waterrrock ratio for carbon, and a

Žhigh ratio for oxygen Banner and Hanson, 1990;.Banner, 1995; Jacobsen and Kaufman, this issue .

Diagenetic stabilization of carbonates therefore re-sults in transposition of carbon from a precursor to asuccessor mineral phase.

The above oscillations in the d13C trendcarbonateare being increasingly utilized for correlation and

Žisotope stratigraphy e.g., Knoll et al., 1986; Derry et.al., 1992 . This tool, however, is subject to addi-

tional limitations if compared to Sr isotopes. Themajor difference arises from the fact that d13Ccarbonateat any given time shows a considerable spread ofvalues due to spatial variability of oceanic d13CÝCO 2Ž .e.g., Kroopnick, 1985 and due to biological factors

Žof shell formation McConnaughey and Whelan,.1997 . The complications arise not so much during

Open access under CC BY-NC-ND license.

http://creativecommons.org/licenses/by-nc-nd/3.0/

( )J. Veizer et al.rChemical Geology 161 1999 59–88 61

logging of a single well or a profile, but can becomeconsiderable if unrelated sequences or wells arecompared. Nonetheless, large peaks can serve ascorrelation markers, particularly in the PrecambrianŽ .Kaufman et al., 1992 . In addition, the seculard13C trend is increasingly utilized as a proxycarbonatefor determining oceanic paleoproductivityrpreserva-tion patterns, ancient pCO and pO states, and2 2

Žsimilar paleoenvironmental phenomena Berner etal., 1983; Berner, 1994; Kump and Arthur, this

.issue .The 18 Or16O composition of ancient oceans is

one of the perennial and most controversial issues ofisotope geochemistry. Commencing with the nascentstages of isotope geochemistry, it has been realizedŽ .Baertschi, 1957; Clayton and Degens, 1959 thatoxygen isotopic composition of carbonates becomesprogressively more depleted in 18 O with increasing

Žage of the rocks. This trend for carbonates and other.chemical marine sediments has been confirmed by

Ž .many studies e.g., Veizer and Hoefs, 1976 and itsvalidity is not disputed. It is the implications of suchobservations that result in a sharp division within thescientific community. If essentially a primary fea-ture, the implications could be one or all of the

Ž . 18following: 1 seawater d O must have changed inŽthe course of the Phanerozoic e.g., Weber, 1965;

.Perry and Tan, 1972; Hudson and Anderson, 1989 ;Ž .2 the early oceans must have been much warmer

Žthan today, up to ;708C Knauth and Epstein,1976; Kolodny and Epstein, 1976; Karhu and Ep-

. Ž .stein, 1986 ; or 3 they were stratified, with deepwaters being generated by sinking of salty brines

Žfrom extensive evaporation at low latitudes Brass et.al., 1982; Railsback, 1990 . Objections to alternative

Ž .1 arise from studies of the oceanic crust and itsŽ .ancient counterparts Gregory and Taylor, 1981 and

Žfrom model considerations Muehlenbachs and Clay-.ton, 1976; Gregory, 1991; Muehlenbachs, 1998 ,

both claiming that the d18 O of ocean water is, andhas been, buffered at ;0‰ SMOW by rockrwaterinteractions in hydrothermal cells at oceanic ridges.

Ž .For alternative 2 , the persistence of essentially thesame faunal assemblages and the recurrence of iceages throughout the Phanerozoic are difficult to rec-

Žoncile with the advocated warm temperatures Veizer. Ž .et al., 1986 . Finally, for alternative 3 , the implied

permanent saline stratification would be difficult to

sustain and, even if sustainable, it could account forno more than ;1.5‰ 18 O depletion in the near-

Ž .surface oceanic layer Railsback, 1990 .The above considerations, and the observation

that advancing diagenetic recrystallization of carbon-Ž . 18ate and other phases results in O depletion, either

due to involvement of meteoric waters or to elevatedtemperatures, led one group of researchers to arguethat the d18 O secular trend is essentially a diagenetic

Žfeature Degens and Epstein, 1962; Killingley, 1983;.Land, 1995 . Resolution of this secondary vs. pri-

mary dilemma is of utmost importance, because thed18 O could potentially serve as a powerful paleo-ceanographic tracer for the entire Phanerozoic.

In this contribution, due to space restrictions, weconcentrate solely on documentation of the trends inorder to make the datasets available to the scientificcommunity. Their interpretations and implicationsfor the d13C are discussed by other authors in this

Žvolume Hayes et al., this issue; Kump and Arthur,. 18this issue . For the d O, the resolution of the pri-

mary vs. secondary nature of the Phanerozoic trendis a precondition for any argumentation dealing withconsequences. We shall therefore concentrate on thisissue only, leaving the model implications for futureconsideration.

Because of the potential utility of 87Srr86Sr, d13Cand d18 O secular trends for resolution of paleoceano-graphic, paleoecological and stratigraphic questions,it is essential to define these trends as tightly aspossible. Here, we introduce a new set of paleosea-water curves for the Phanerozoic. In order to mini-mize the impact of diagenetic resetting of the signal,we have concentrated on material that is relativelyresistant to diagenetic alteration. Similarly, in orderto attain high resolution, we resampled ‘complete’and well-dated sequences. Nevertheless, fulfillmentof these two requirements, in particular the temporalassignment of samples, is a complex task that re-quires some introductory discussion.

2. Completeness of the sedimentary record andisotopic ages

Any given stratigraphic section only intermittentlyŽrecords the passage of geological time Dingus,

.1984 . On a global scale, the volume of sedimentary


rocks decreases with age because the record becomesprogressively less complete with time. Within anysedimentary sequence, there are hiatuses which rep-resent periods of non-deposition, erosion andrordiagenesis. Stratigraphic sections spanning greatertemporal intervals have a greater opportunity of in-corporating more hiatuses and hiatuses spanning

Ž .longer periods of time Sadler, 1981; Korvin, 1992 .Determining the stratigraphic position and durationof a hiatus within a section is often problematic. Theend result is a progressive decrease in the complete-ness of the stratigraphic record with increasing age.

In general, sequences often have erosional ornon-depositional contacts which are difficult to iden-tify from lithological parameters alone. In these situ-ations, biostratigraphy, absolute dating and magne-tostratigraphy are often utilized. Unfortunately, eachmethod has its limitations. Presently, the highestresolution can be achieved with biostratigraphy, witha resolution limit rarely better than 0.5 Ma and moreoften on the order of 1 Ma or more. As a result, theupper limit for using this method for hiatus identifi-cation cannot be better than 0.5 Ma. In addition,bioturbation can effectively homogenize sedimentsin the upper 10–15 cm, rendering any stratigraphic

Žmethod meaningless at scales less than 15 cm Anders.et al., 1987 . If sediment accumulation is slow, then

the effect of bioturbation on the resolution maybecome geologically significant.

Stratigraphers define a complete section as onewhich has no gaps larger than each time unit of a

Ž .specific scale Anders et al., 1987 . As the resolutionŽ .improves corresponding to shorter time scales , there

is a corresponding decrease in the perceived com-pleteness of the section. Therefore, improved sam-pling accuracy will result in an overall decrease inthe perceived completeness of the stratigraphic sec-

Ž .tion Schindel, 1982; Dingus, 1984 . On average, itis estimated that only 1r30th of the elapsed time is

Žrepresented by sediments Algeo and Wilkinson,.1988; Miall, 1994 .

The above constraints are compounded by prob-lems with assignment of ‘absolute’, or more cor-rectly ‘isotopic’, ages to particular samples. In orderto enable comparison with other datasets, we consis-

Ž .tently utilized the time scale of Harland et al. 1990 ,despite the fact that many of these ‘ages’, includingthe commencement of the Phanerozoic at 570 Ma

Ž .ago rather than 543 Ma; Bowring and Erwin, 1998 ,are today not accepted anymore. The uncertainties in‘age’ estimates for epochs and stages are usuallyseveral million years, larger than the entire durationof many stages. For example, the estimated durationof the Eifelian of ;4 Ma is bracketed by isotopic

Žages of 386"5 to 382q9ry14 Ma Harland et.al., 1990 . This interval encompasses some five con-

odont biozones, each arbitrarily assigned an isotopicage by linear extrapolation between these two esti-mates that have large experimental errors.

Because of the above complications, we haveconcentrated on sedimentary sequences that were ascomplete as possible, preferably stratotypes, provid-

Ž .ing they contained suitable material shells for iso-topic studies. The correlation of samples from differ-

Žent localities was based on biostratigraphy bio-.zones and not on ‘isotopic ages’. The relative ‘age’

within the same sequence and biozone was based onstratigraphic superposition. Unfortunately, whencomparing distant Paleozoic and Mesozoic se-quences, the best that we could accomplish was acorrelation to a biozone resolution. This will beimportant when discussing higher level oscillationsin the isotope signal.

3. Sample selection

In order to constrain as much as possible post-de-positional alteration of the isotope signal, we haveconcentrated on samples that are relatively resistantto diagenesis. The preferred material is the low-Mgcalcite of marine skeletons, such as brachiopods,belemnites and foraminifera. Keeping in mind thatthe Cenozoic portion of the isotope record has beenextensively studied, chiefly due to DSDP, we haveconcentrated on the Paleozoic and Mesozoic. Theseleft brachiopods and belemnites as the only abundantand stratigraphically widespread fossils with low-Mgcalcitic skeletons. Furthermore, for brachiopods, we

Žhave utilized only the so called ‘secondary’ i.e.,.interior layer, because its elongated calcitic fibers

showed the highest degree of preservation, up tosub-micron levels. The details of our approach andexamples of the exceptional degree of preservationof many shells have been documented in Diener et


Ž . Ž .al. 1996 , Veizer et al. 1997a,b , Azmy et al.Ž . Ž .1998 and Bruckschen et al. 1999 . For the Meso-zoic belemnites, the approach has been similar, ex-cept that the samples were drilled within a single‘lamina’ of the rostrum, parallel to its elongationŽ .Podlaha et al., 1998 .

For Sr isotopes, in the initial stages, we haveconcentrated also on conodonts, believing that be-cause of their superior stratigraphic resolution andthe high Sr content of their phosphatic skeletons,they may yield a 87Srr86Sr curve for the Phanero-zoic seawater with the highest temporal resolutionattainable. Comparison with brachiopod data showed,however, that conodonts, even at a conodont alter-

Ž . 87 86ation index CAI as low as 1.5, had Srr Sr ratios

that were at best comparable to coeval brachiopods,but mostly the measurements were more radiogenicŽ .Ebneth et al., 1997; Veizer et al., 1997a . These,and the textural and proton-induced X-ray emissionŽ . Ž .PIXE trace element studies Bruhn et al., 1997 ,showed that conodonts exchanged about 1r3 of theirSr with the surrounding rock matrix. In suitablematrix, such as pure carbonates, conodonts may stillretain their near-original 87Srr86Sr values, but mostly

y5 Žthe values are ;5=10 more radiogenic Ebneth.et al., 1997 . Because of these reasons, we have

discontinued conodont studies, except for the timeintervals where the low-Mg calcitic fossils were rareor absent, such as the Cambrian and the early Trias-sic.

Ž . Ž .Fig. 1. The SEM photos of representative brachiopods. Note the excellent preservation of the secondary shell layer from Azmy, 1996 . aŽ .Isorthis amplificata B27; Wenlock, Much Wenlock, UK shell, with punctae filled by secondary calcite. The punctae calcite has intensive

yellowish cathodoluminescence, as opposed to no or only intrinsic blue luminescence of the shell. The isotopic and chemical characteristics18 13 Ž . Ž . Ž .of the shell are the following: d Osy5.50‰, d Csq0.39‰, Mns307 ppm, Srs1934 ppm. b As 3a at higher magnification. c

Ž . 18 13 Ž .Delthyris eleÕata EK 38-7; Pridoli, Ohesaare Cliff, Estonia : d Osy5.50‰, d Csy0.50‰, Mns104 ppm, Srs1328 ppm. d AsŽ .3c at higher magnification.


Ž . Ž .Fig. 2. Histograms of Sr ns1370 and Mn ns1267 concentra-tions in the studied brachiopods and belemnites.

The samples were screened for their degree ofŽpreservation by optical transmitted light, cathodolu-

.minescence and scanning electron microscopy, SEMtechniques. These have been documented in Bruhn et

Ž . Ž .al. 1995 and Bruckschen et al. 1995a,b . Fig. 1demonstrates an example of excellent preservation oftextures on sub-micron scale. This optical screeningwas complemented by trace element evaluation by

Žwet chemical techniques inductively coupled plasmaw xatomic emission spectroscopy ICP-AES , atomic ab-

w x.sorption spectroscopy AAS on aliquots remainingafter phosphoric acid liberation of CO for stable2

Ž .isotope measurements Coleman et al., 1989 and byPIXE spot analyses on thin sections. Based on thewet chemical approach, some 2r3 of all studiedsamples have Sr and Mn concentrations comparable

Ž .to modern low-Mg calcitic shells Fig. 2 . The re-maining 1r3, mostly with Sr content -800 ppm,may or may not have suffered some diagenetic lossof Sr, since concentrations as low as 200 ppm have

Žbeen reported for modern brachiopods Morrison and.Brand, 1986; Brand, 1989b . Note also that recrystal-

lization, unless at exceptionally high waterrrock ra-tios, does not necessarily result in resetting of87Srr86Sr or d13C values. For carbon, this is becauseof the earlier-discussed low waterrrock ratios of therespective diagenetic systems. For strontium, due to

Ž .the fact that D calcite is -1, recrystallizationSrhappens in a partially open system buffered by the

Ž .dissolving carbonate phase Veizer, 1983 . Alterna-tively, the lower Sr concentrations may be, in part,due to inclusion of distinct small domains of sec-ondary calcites, and not to partial recrystallization,leaving the bulk of the shell unaltered. The presenceof such secondary calcites can usually be detected by

Ž .optical means Fig. 3 , particularly cathodolumines-cence, and these calcites have also distinctive chemi-

Ž . Žcal Table 1 and isotopic signatures Bruckschen et.al., 1995a; Bruhn et al., 1995 .

4. Samples

The present study is based on several thousandsamples of Cambrian to Cretaceous ages. These were

Ž .collected in Canada Anticosti Island, Ontario , USAŽUtah, Oklahoma, Texas, Ohio, Kentucky, Missouri,

.New York , Ireland, England, Wales, Spain, Italy,ŽBelgium, Germany, Poland, Norway including

Ž . Ž .Fig. 3. Top Cathodoluminescence photomicrograph of a Carboniferous punctate brachiopod sample PB 192a . Note the bright yellowŽ .luminescence of the punctae that are filled by diagenetic calcite. Bottom PIXE elemental map of the dark square in the top figure. Step

Ž . Ž .size 10 mm. Note that the punctae are typified by Sr depletions light domains and simultaneous Mn enrichments dark domains . The traceelement data are in Table 1. Punctate brachiopods, except for demonstration purposes, were avoided in the present study. Modified from

Ž . Ž .Bruhn 1995 and Bruckschen et al. 1995a .


Table 1Ž .The PIXE trace element concentrations for a Carboniferous punctate brachiopod sample PB 192a

Mn Fe Sr

Shell Punctae Shell Punctae Shell Punctae

33 1810 40 733 1042 51820 3699 35 9261 977 21727 2043 36 722 969 37936 5040 73 1069 953 223

Mean 29 3148 48 2946 985 334s 8 1501 22 4833 12 92

. Ž .Svarlbad , Sweden including Gotland , Estonia,Ž .Latvia, Lithuania, Ukraine Donetsk basin, Podolia ,

ŽAustria, Slovakia, Hungary, Russia Moscow basin,.St. Petersburg area, Urals, western Siberia , Mo-

87 86 ŽFig. 4. Srr Sr variations for the Phanerozoic based on 4055 samples of brachiopods ‘secondary’ layer only for the new BochumrOt-.tawa measurements , belemnites and conodonts, and nine samples of micritic matrix. Normalized to NBS 987 of 0.710240. The literature

Ž . Ž . Ž .data foraminifera, belemnites and conodonts are from the following sources: Peterman et al. 1970 , Dash and Biscaye 1971 , Veizer andŽ . Ž . Ž . Ž . Ž . Ž .Compston 1974 , Brass 1976 , DePaolo and Ingram 1985 , Koepnick et al. 1985 , Hess et al. 1986, 1989 , Popp et al. 1986c ,

Ž . Ž . Ž . Ž . Ž .McKenzie et al. 1988, 1993 , Hodell et al. 1989, 1990, 1991 , Brand 1991 , Martin and McDougall 1991, 1995 , Bertram et al. 1992 ,Ž . Ž . Ž . Ž . Ž .Dia et al. 1992 , Barrera et al. 1993 , Kurschner et al. 1993 , Whittaker and Kyser 1993 , Banner and Kaufman 1994 , Cummins and¨Ž . Ž . Ž . Ž . Ž . Ž .Elderfield 1994 , Denison et al. 1994 , Hodell and Woodruff 1994 , McArthur et al. 1994 , Oslick et al. 1994 , Quinn et al. 1994 ,Ž . Ž . Ž . Ž . Ž . Ž .Jones et al. 1994a,b , Chaisi and Schmitz 1995 , Farrell et al. 1995 , Ruppel et al. 1996 , Wenzel 1997 , Qing et al. 1998 .


Žrocco, south China, Australia Queensland, Northern.Territory and New Zealand. They have been col-

Ž .lected assembled and described by the followingŽ .researchers and publications: Cambrian S. Ebneth ,

Ž . ŽOrdovician G.A.F. Carden , Silurian Azmy, 1996;. ŽAzmy et al., 1998, 1999 , Devonian Diener, 1991;

Ebneth, 1991; Pawellek, 1991; Fauville, 1995; Golks,1995; Ala, 1996; Diener et al., 1996; Ebneth et al.,

. Ž1997 , Carboniferous Bruckschen et al., 1995b,.1999; Bruckschen and Veizer, 1997 , Permian

Ž . Ž .Jasper, 1999 , Triassic Korte, 1999 and JurassicrŽCretaceous Podlaha, 1995; Podlaha et al., 1998,

.Podlaha et al., this issue . Judging from paleogeo-Ž .graphic reconstructions of Scotese et al. 1994 , most

Žof these samples originate from tropical regions 308N.to 308S of the paleoceans, with few from higher

paleolatitudes. From this collection, some 1500 sam-Žples of brachiopods and belemnites and additional

.conodonts have been selected for further Sr, Oand C isotope studies and the details of their des-criptions, including chemistry, isotopes, geologyand location, are available in the above publi-cations or in the summary tables posted on the Web

site: http:rrwww-ep.es.llnl.govrgermrevolution.html or http:rrwww.science.uottawa.cargeologyrisotope datar.y

5. Phanerozoic 87Srrrrrr86Sr trend

The summary of Sr isotope data for PhanerozoicŽ .seawater Fig. 4 is based solely on measurements on

fossils. The Cenozoic and late Cretaceous part isfrom the literature and based chiefly on foraminifera.The remaining literature data, and all of BochumrOttawa measurements, are based on brachiopods,belemnites and conodonts. The sole exceptions are

Ž .the matrix micrite samples for the early Cambrianthat were incorporated because of the lack of suitablefossils. The ‘curve’ resembles the one by Burke et al.Ž .1982 , but it differs in detail and is more tightlyconstrained. Where larger spread of data exists, suchas the radiogenic mid-Permian Chinese samplesŽ .Jasper, 1999 , it is dealt with in the publicationdescribing these particular periods. However, the

87 86 Ž .Fig. 5. Srr Sr variations during the Devonian new BochumrOttawa data only .


Ž .clumping of data cf. Devonian and Carboniferousis mostly a reflection of a large number of measure-ments, with higher-order oscillations compressed

Žalong the temporal axis cf. Bruckschen et al., 1995b,.1999; Bruckschen and Veizer, 1997 .

Ž .Taking the Devonian as an example Fig. 5 , thehigher-order oscillations, based on 384 samples, arewell-illustrated. The secular trend shows a continu-ous decline through the Lower Devonian, a plateauin the Middle Devonian, a rise through the Frasnianand another plateau in the Upper Frasnian and Fame-nian. This figure also illustrates that conodonts areusually somewhat more radiogenic than coeval bra-chiopods. Furthermore, even at this resolution, someclumping still persists, such as at the EmsianrEifelianand FrasnianrFamenian transitions. Again, the major

culprit is the presence of still higher-order fluctua-tions in a compressed time scale. For the Middle

Ž .Devonian, this was illustrated in Diener et al. 1996Ž .their Fig. 6 , where it was also pointed out that thegeological reproducibility of the Sr isotope pattern inprofiles only several kilometers apart was no betterthan 5=10y5. Thus, only the oscillations in excessof this range could be resolvable on regional toglobal scales.

Another constraining issue is the problem of cor-relations within a single biozone. As discussed in

Ž . Ž .Veizer et al. 1997a their Fig. 8 , the scatter of87Srr86Sr values at the FrasnianrFamenian transi-tion, the Kellwasser ‘event’, results from real oscilla-tions within one or two biozones. Yet, on largerregional or global scales, it is difficult to match these

Fig. 6. 87Srr86 Sr variations during the Devonian based on conodont biozones. Explanations: circlesmean; boxs"1s ; verticallinesminimum and maximum. The 2s in the lower right corner is an average 2s for the NBS 987 standard. Note that only brachiopods

Žare included in this figure. The biozones are the following Weddige, 1977, 1988; Carls, 1988; Sandberg et al., 1989; Ziegler and Sandberg,.1990 : 1 — woschmidti, 2 — postwoschmidti, 3 — pesaÕis, 4 — sulcatus, 5 — kindleri, 6 — pirinea, 7 — deniscens, 8 — gronbergi, 9

— laticostatus, 10 — serotinus, 11 — patulus, 12 — partitus, 13 — costatus, 14 — australis, 15 — kockelianus, 16 — esensisŽ .arkonensis, obliquimarginalus, bipecatus , 17 — hemiansatus, 18 — Õarcus, 19 — hermani — cristatus, 20 — disparitis, 21 —falsoÕialis, 22 — transitans, 23 — punctata, 24 — hassi, 25 — jamieae, 26 — rhenana, 27 — linguiformis, 28 — triangularis, 29 —crepida, 30 — rhomboidea, 31 — marginifera, 32 — trachytera, 33 — postera, 34 — expansa, 35 — presulcata.


wiggles in different profiles. In effect, one is limitedby constraints of biostratigraphy and the matching isin the term of biozones. This is not to say that someof the observed 87Srr86Sr variations may not be aresult of undetected post-depositional alteration, buta portion of these variations is real. It is thereforedifficult to claim some specific 87Srr86Sr value for agiven biozone. The alternatives can rely on eitherselecting the least radiogenic value or accepting all‘good’ data as representative of that biozone. Weprefer the second approach as more representative ofthe real situation. We therefore plot the mean, the 2srange and the total range of values for each of the 35

Ž .Devonian conodont biozones Fig. 6 . The Phanero-zoic 87Srr86Sr marine trend based on the aboveprinciples is, in our view, a more realistic reflectionof the natural situation and as such should be con-structed separately for each period, not only for the

Ž .Devonian or Silurian Azmy et al., 1999 .

6. Correlation potential

Bearing in mind the above qualifications, Sr iso-topes can serve as a useful complementary correla-tion tool, particularly for sequences where the bios-tratigraphic approach is of little value, due either to

the dearth of fossils or to their limited stratigraphicutility. Nevertheless, the technique can lead to im-proved correlations even for sequences with good

Ž .biostratigraphic control. Azmy et al. 1999 demon-strated that for the Silurian, the 87Srr86Sr ratio en-ables stratigraphic assignment of samples from inde-pendent sets to within "1 graptolite biozones. Thisis no minor accomplishment considering the diffi-culty of correlation for profiles with differing sedi-mentation rates and taking into account the uncer-tainty of the ‘first and last appearance’ of a given‘Leitfossil’ in the sedimentary record. The issue canbe illustrated by the present study of the GivetianrFrasnian transition in the Eifel Mountains of Ger-many. These well-studied sections, in close proxim-ity to the Belgian stratotypes, have a well-established

Ž .biostratigraphy Weddige, 1977; Struve, 1982 . Un-fortunately, the GivetianrFrasnian transition isstrongly dolomitized and it was necessary to samplethe coeval sections at the Teerstrassenbau Wahl-heimrFriesenrath quarry in Aachen and at the tunnel

Ž .near Behringhausen, east of the Rhine Ostsauerland .87 86 Ž .The measured Srr Sr values Fig. 7 define a

rising trend of ;3=10y4 in about 4 Ma that spansŽ .the gap between the Õarcus No. 18 and rhenana

Ž .No. 26 biozones in the Eifel. In detail, however, itappears that the steepest rise in the curve, of about

Fig. 7. 87Srr86 Sr trend at the GivetianrFrasnian transition. Conodont biozones as in Fig. 6.


y4 Ž1=10 in F1 Ma, appears in the falsioÕalis No..21 conodont biozone in Aachen and in the transi-

Ž .tans No. 22 biozone in Behringhausen. The criticalŽportion of the profile in Aachen is at its top Reis-

. Ž .sner, 1990 and it was assigned to the lower assy-metricus biozone in the older stratigraphic assign-

Ž .ment Ziegler, 1962 that encompassed the presentfalsioÕalis and transitans biozones. If so, that por-tion of the Aachen profile that contained the bra-chiopods measured for 87Srr86Sr should perhaps be

Ž .assigned to the transitans biozone No. 22 , and becorrelative with the Behringhausen.

A third approach of utilizing Sr isotopes for corre-lation purposes is based on inflexions in the trendŽ .McArthur, 1994; Azmy et al., 1999 . The higher-order oscillations are not necessarily reproducible,particularly in their amplitude. This is because thesedimentary sequences are not an unbroken recorderof time and the 87Srr86Sr evolutionary trend of

seawater can thus be truncated at any time prior toreaching the apex of an oscillation. An inflexion, onthe other hand, can be recognized even if somecorresponding strata are missing. Note that unrecog-nized hiatuses may appear also as abrupt ‘jumps’ in

Žthe strontium isotope trend cf. Diener et al., 1996,.their Fig. 7 .

7. Phanerozoic d13C trend

The secular Phanerozoic trends based on the newŽ .BochumrOttawa dataset Fig. 8 or on the compila-

Ž .tion of literature data Fig. 9 resemble each other interms of their overall shape and in the considerablespread of values. As for Sr isotopes, some of thisscatter is due to higher-order natural oscillationsclumped together by compressed time scale. This isparticularly the case for the literature dataset with itsmuch poorer temporal resolution. Additional scatter

13 Ž . Ž .Fig. 8. Phanerozoic d C trend based on 1564 brachiopod secondary layer and belemnite laminae pelucidae measurements at BochumŽ . Ž .and Ottawa. Three Triassic samples two corals and one bivalve from Cassian Beds and Kossener Schichten Alps are still preserved as¨

aragonite.


13 Ž .Fig. 9. Phanerozoic d C trend compiled from 3918 measurements for LMC brachiopods, belemnites, oysters, foraminifera and 96Ž . Ž . Ž .measurements for A mollusc shells. The sources of the data are the following: Compston 1960 , Douglas and Savin 1971, 1973 , Veizer

Ž . Ž . Ž . Ž . Ž . Ž .and Hoefs 1976 , Brand and Veizer 1981 , Popp et al. 1986a,b , Veizer et al. 1986 , Miller et al. 1987 , Adlis et al. 1988 , BrandŽ . Ž . Ž . Ž . Ž . Ž .1989b , Delaney et al. 1989 , Rush and Chafetz 1990 , Grossman et al. 1991, 1993 , Middleton et al. 1991 , Jones 1992 , Wadleigh and

Ž . Ž . Ž . Ž . Ž . Ž .Veizer 1992 , Brand and Legrand-Blain 1993 , Lavoie 1993 , Anderson et al. 1994 , Qing and Veizer 1994 , Carden 1995 , CarpenterŽ . Ž . Ž . Ž . Ž . Ž .and Lohmann 1995 , Mii 1996 , Rao 1996 , Samtleben et al. 1996 , Wenzel and Joachimski 1996 , James et al. 1997 , Mii et al.

Ž .1997 .

is due to oceanographic and biological factors, men-tioned already in the introductory part, and well-pre-

Ž .served ancient aragonitic Anderson et al., 1994 asŽ .well as modern calcitic representatives Fig. 9 are

no exception to this pattern. Both sets of data, aswell as the secular d13C curve based on wholecarbonate

Ž .rocks Veizer et al., 1980; Lindh, 1983 , show ageneral increase in d13C throughout the Paleozoic,followed by an abrupt decline and subsequent oscil-lations around the present-day value in the course ofthe Mesozoic and Cenozoic. Because of this coher-

Žence, we feel justified in pooling both datasets Figs..8 and 9 in order to generate a secular trend for the

Ž .entire Phanerozoic Fig. 10 . The running mean basedon this combined set was calculated for 5 Ma incre-

mental shifts, because this is the realistic resolutionfor global biostratigraphic correlations and for ourcombined dataset. The bands around this mean incor-porate 68 and 95% of all measured samples, respec-tively. Note also that superimposed on the overalltrend are pronounced, but somewhat dampened dueto the 20 Ma window, second- and higher-orderoscillations. Considering the global nature of thedatabase, with samples originating from five conti-nents and a multitude of sedimentary basins, theobserved peaks are likely of global significance. Anexception may or may not be the Aptian peak at;115 Ma that is based on a small number of

Ž .measurements cf. Podlaha et al., 1998 . It is likelythat even many of the shorter-term peaks may still be


Fig. 10. Phanerozoic d13C trend for combined BochumrOttawa and literature data for LMC shells. The running mean is based on 20 MaŽ .window and 5 Ma forward step. The shaded areas around the running mean include the 68% "1s for a strictly Gaussian distribution and

Ž .95% "2s of all data.

of global significance, as will be discussed later forthe terminal Ordovician.

8. Phanerozoic d18 O trend

In analogy to carbon isotopes, the d18 O trends forŽ .the BochumrOttawa Fig. 11 and the literature

Ž .datasets Fig. 12 overlap, including their similaritiesŽin the second-order structure e.g., the SilurianrDe-

.vonian dip followed by Carboniferous rise . Wetherefore feel justified in combining them into a

Ž .general Phanerozoic trend Fig. 13 . Several featuresare immediately apparent from this presentation.

Ž . 181 The data define a trend of increasing d Ovalues, from about y8‰ at the onset of thePhanerozoic to about 0‰ at present.

Ž . 132 As for the d C, the trend is not a linearfeature, but a band. Note, however, that the observed

d18 O variations in modern tropical brachiopods atŽshelf depths are ;4‰ Carpenter and Lohmann,

.1995; Bruckschen et al., 1999 and for the specimensŽfrom temperate climates they span about 8‰ Fig.

. 1812 . A similar range of d O values was observed byŽ .Anderson et al. 1994 for ancient belemnites and

molluscs, the latter still preserved as pure aragoniteŽ .Fig. 13 , within a single member of the JurassicOxford Clay Formation. From this perspective, it isindeed noteworthy that the collections of fossils fromany one biozone, each representing populations witha much larger temporal and spatial range than themodern one, do not show a markedly larger disper-

18 Ž . 18sion of d O values Fig. 13 . The existing Odepleted outliers in the Carboniferous and PermianŽ .see Bruckschen et al., 1999 for further discussionrepresent only ;1% of the total population, thusextending somewhat the lower range of the 95%limit. Such outliers have only a negligible impact on


18 Ž . Ž .Fig. 11. Phanerozoic d O trend based on 1654 brachiopod secondary layer and belemnite laminae pellucidae measurements at BochumŽ .and Ottawa. Three Triassic mollusc samples are still preserved as aragonite see Fig. 8 .

the mean, the 2s range, or the upper 95% boundaryof the band. The latter boundary is sharp, with

Ž .almost no samples above it Figs. 11 and 12 . Conse-quently, even an arbitrary retention of only the‘heaviest’ values would not invalidate the secularPhanerozoic trend.

Ž . Ž .3 Statistical treatment of the data Fig. 13suggests an existence of second-order oscillationswith a frequency of ;150 Ma, their apexes coinci-dent with cold intervals and major glaciations. Ad-vancing global cooling, intermittently compoundedby generation of ice caps, could have been theprincipal reasons for the observed 18 O enrichments.

Ž .4 The three Carboniferous molluscs entombed inasphalt and still well-preserved as nacreous aragoniteŽ .Brand, 1989a , as well as the 122 Jurassic and

ŽTriassic aragonitic molluscs Anderson et al., 1994;.Korte, 1999 , all plot well within the main trend

Ž .Fig. 13 .

All the above points will be of importance fordiscussion of the primary vs. secondary nature of thed18 O Phanerozoic trend.

9. Correlation of oxygen and carbon isotopes

Ž .The crossplots of all Phanerozoic data Fig. 14show the relationships of d18 O and d13C for theentire Phanerozoic. Covariant d18 Ord13C trends,such as those of the Ordovician, Silurian, Permian orTriassic, are often interpreted as an indicator ofdiagenetic resetting due to 18 O and 13C depletion in

Ž .altered samples e.g., Bathurst, 1975 . While post-depositional resetting, particularly in diagenetic sys-tems dominated by bicarbonate generated from soilCO , can indeed generate such trends, their exis-2tence is not an a priori proof of diagenetic alteration.Modern shells, almost as a rule, display such positive


18 Ž .Fig. 12. Phanerozoic d O trend of compiled 3969 measurements for LMC shells brachiopods, belemnites, oysters, foraminifera and 122Ž .Carboniferous and Jurassic aragonitic mollusc shells Brand, 1989a; Anderson et al., 1994 , the latter listed in literature as 100% aragonite.

Ž .The sources of the data are as in Fig. 9 and Lowenstam 1961 .

covariance, principally due to isotope fractionationŽeffects in the process of shell secretion McCon-

.naughey and Whelan, 1997 . The pattern in theŽ .diagenetically unaffected Quaternary cluster Fig. 14

confirms this assertion.Another reason for the covariance, particularly

during the Paleozoic, is the fact that both isotopesshow a general secular increase with decreasing age.The Paleozoic cluster represents, therefore, with largedegrees of overlap, a trend from the isotopicallydepleted Cambrian to the enriched Permian. In de-tail, the covariance is particularly pronounced forperiods of fast secular change, such as the Ordovi-cian. Note, however, that the patterns of thisd18 Ord13C covariance shift toward positive d18 Ovalues with decreasing age. The spread of d13Cvalues, on the other hand, remains relatively invari-ant.

In the late Paleozoic clusters, a number of sam-ples fall outside the general covariance domain dueto excessively negative d18 O values. These samples,mostly of Carboniferous age, are the same that plotas 18 O depleted outliers in the d18 O secular trendŽ .Fig. 13 . They originate mostly from the DonetskBasin of the Ukraine and are further discussed in

Ž .Bruckschen et al. 1999 .

10. Primary or secondary trends

The primary nature of the d13C secular trend, atleast in terms of its major features, is no longerdisputed, but the issue is far from settled for the

18 Ž .d O cf. Hoefs, 1997 . Five lines of argumentationare pertinent to this issue: textural, geological, min-eralogical, chemical and isotopic. We shall discussthem in this order.


Fig. 13. Phanerozoic d18 O trend for BochumrOttawa and literature data for LMC shells. Aragonite specimens as in Figs. 11 and 12. Notethat benthic foraminifera are excluded from this figure, because they may not represent ecological analogues to ancient brachiopods and

Ž .belemnites. Glaciations and cold times after Frakes et al. 1992 . Further explanations as in Fig. 10.

Starting with the textural considerations, the ex-cellent textural preservation of the shells, even on a

Žsub-micron scale, is illustrated in Fig. 1 cf. also.Bruckschen et al., 1999 . Such excellent preservation

is by no means a rarity in the present dataset.Documentation can be found, e.g., in BrandŽ . Ž .1989a,b , Wadleigh and Veizer 1992 , Grossman et

Ž . Ž . Ž .al. 1993 , Qing and Veizer 1994 , Carden 1995 ,Ž . Ž . Ž .Copper 1995 , Azmy 1996 , Grossman et al. 1996 ,

Ž .Samtleben et al. 1996 , Wenzel and JoachimskiŽ . Ž . Ž .1996 , Wenzel 1997 , Azmy et al. 1998 and

Ž .others. Note that the depicted samples Fig. 1 havetrace element, d13C and 87Srr86Sr values as ex-pected for a Silurian marine calcite, yet their d18 Osare around y5‰ PDB, well within the Phanerozoic

Ž .trend Fig. 13 . For the sake of the argument, let usassume that the oxygen isotope values have beenaltered by meteoric waters with d18 O of say y5 to

y10‰ SMOW. In such a case, one-half to all ofoxygens in these calcite structures, i.e., up to 60% oftheir building blocks, would have to be replaced byextraneous oxygen without leaving behind the slight-est trace in the optical, mineralogical, chemical andisotopic make-up of the shells. This is difficult toconceive, irrespective of whether one invokes a solu-tionrreprecipitation process or, the less likely, soliddiffusion exchange.

The second line of argumentation is based ongeological consideration. Let us assume that the

18 Ž .trend of O depletion with age Fig. 13 is in thefirst instance a result of post-depositional resettingdue to interaction with meteoric andror warm wa-ters. Diagenetic stabilization of carbonate rocks in-deed results in 18 O depletion, but this depletion ismostly related to transformation of metastable poly-

Ž .morphs aragonite, high-Mg calcite into stable ones


Ž .low-Mg calcite and it usually happens very early inthe post-depositional history of the rocks. Subse-quently, and in deep sea environments, the alterationproceeds, at a slower rate, by pressure solution untilcomplete lithification. Once lithified, the subsequent18 O depletions — if any — are much less pro-nounced. Note also that all ancient carbonate rockshave undergone this stabilization step, thus self-cor-recting for it for rocks older than about Cenozoic. Asa result, such a post-depositional trend would have asteep slope during the Cenozoic, exponentially flat-tening out with increasing age. Furthermore, due tovagaries of post-depositional development of differ-ent sedimentary basins, the resetting towards 18 Odepleted values from the assumed modern day equi-librium would result in a fan-shaped pattern thatbroadens with age, with many outliers along thealteration pathway, i.e., between 0‰ PDB and themain Phanerozoic trend. Neither of these predictions

Ž .is supported by experimental data Figs. 11–13 .Disregarding the second-order oscillations, the rateof decline is about linear, the band of about equalwidth, and there are essentially no outliers aboÕethe upper limit of this band. Whatever outliers do

Ž .exist ;1% of samples , they plot mostly below themain trend, i.e., any post-depositional 18 O depletion

Ž .advanced from and not towards the main trend. Inorder to sustain the diagenetic alternative, one wouldhave to argue that the post-depositional alterationadvances at roughly a constant rate, regardless of theage of the rocks, and that all basins throughout the

Ž .world see the distribution of our samples haveundergone alteration strictly proportional to theirages. Considering the plethora of geologic histories,with samples from all continents but Antarctica, andfrom a multitude of sedimentary basins, this is not acredible proposition.

As already pointed out, the apexes of the second-order oscillations appear to have coincided with cold

Ž . 18periods Fig. 13 , perhaps indicating that the d Ooscillations reflect global cooling trends. Leavingaside the issue of the origin of these second-orderoscillations, let us concentrate on one of them, theOrdovicianrSilurian transition. The d18 O record of

Ž . 18this transition Fig. 15 shows the d O patternŽ .Carden, 1995; Azmy, 1996; Azmy et al., 1998where the positive d18 O excursions coincide, withinthe resolution of a biozone, with glacial sediments, a

pattern, except for the shift in the baseline, analo-gous to that of the Quaternary. Furthermore, thepronounced terminal Ordovician d18 O peak is clearly

Ž .a global feature within the acuminatus No. 21biozone, because this signal was detected in ChinaŽ .Carden, this project , the Baltic region, Canada and

ŽArgentina Brenchley et al., 1995; Marshall et al.,.1997 . This pattern can be easily understood in terms

of the primary signal, but is difficult to reconcilewith a diagenetic interpretation.

The third line of argumentation is based on themineralogy of the shells. Because of their exception-

Ž .ally well-preserved textures Fig. 1 , we believe thatlow-Mg calcitic shells were relatively resistant todiagenetic alteration. Moreover, molluscs that arestill preserved in their highly unstable original arago-nitic mineralogy also plot within the general trend

Ž .defined by the low-Mg calcitic shells Fig. 13 . Inorder to sustain the diagenetic alternative, it wouldagain be necessary to claim that all these shellsexchanged most of their oxygen, yet retained notonly their original low-Mg calcitic, but also originalaragonitic mineralogies.

The fourth line of argumentation is based on thechemical composition of the studied samples. As

Ž .already demonstrated Fig. 2 , a minimum of 2r3 ofthe 1370 samples studied by us for trace elementshave compositions directly comparable to modernlow-Mg calcitic shells. Diagenetic alterations usually

Ž . Ž .lead to decline in Sr Na and increase in Mn FeŽcontents Brand and Veizer, 1980; Veizer, 1983;

.Banner, 1995 . Based on such criteria, about 1r3 ofthe samples characterized by Sr contents of less thanabout 800 ppm may arguably be considered some-what altered. Nonetheless, trace element criteria forat least 2r3 of the samples are not consistent withthe proposition of diagenetic alteration.

The fifth line of argumentation is based on theisotopic patterns of the studied populations. Thepresent Phanerozoic dataset is the first of such quan-tity and quality where the samples have been studiedsimultaneously for d18 O, d13C, 87Srr86Sr, trace ele-ments, and optical parameters. In addition, some ofthe samples have been studied for d34S of struc-

Žturally bound sulfur Kampschulte and Strauss, 1996;.Strauss, this issue . The possibility of a large strati-

graphic mismatch for different isotopic systems istherefore mitigated. A crossplot of d18 O vs. 87Srr86Sr


Fig. 14. The d18 O vs. d13C crossplots for the samples listed in Figs. 9–12. Explanations: circles — BochumrOttawa data; upright crossesŽ . Ž— literature data LMC brachiopods, belemnites, oysters, foraminifera ; diagonal crosses — literature data LMC Quaternary temperate

.brachiopods . P 103r4: Permian, 103 BochumrOttawa samplesrfour literature samples. Note that the Tertiary field contains only pelagicforaminifera.


Fig. 15. The d18 O trend across the OrdovicianrSilurian transition. The Ordovician biozones, based on graptolites, are the following: 17 —Pleugograptus linearis, 18 — Dicelograptus complanatus, 19 — Dic. anceps, 20 — Climacograptus extraordinarius, 21 — Glyptograp-tus acuminatus. The Silurian graptolite biozones are the following: 1 — Parakidograptus acuminatus, 2 — AtaÕograptus ataÕus, 3 —Lagarograptus acinaces, 4 — Coronograptus cyphus, 5 — Monograptus triangulatus, 6 — Diplograptus magnus, 7 — Pribylograptusleptotheca, 8 — Monog. conÕolutus, 9 — Monog. sedgwickii, 10 — Monog. turriculatus, 11 — Monog. crispus, 12 — Monoclimacisgriestoniensis, 13 — Monoc. crenulata, 14 — Cyrtograptus centrifugus, 15 — Cy. murchisoni. Glacial episodes after Grahn and CaputoŽ .1992 . Patterns as in Fig. 6.

Ž .Fig. 16 shows a very high degree of correlation.From a purely theoretical perspective, one couldperhaps argue that this correlation is a result ofdiagenetic alteration. In that case, however, the Sr

Ž .isotopic trend Fig. 4 would have to be a secondaryfeature as well. In order to escape such a conclusion,it would be necessary to postulate that the 87Srr86Srof some 4000 samples remained almost perfectly

Ž .preserved Fig. 4 while their oxygens have beenheavily reset. This is not a credible proposition in

view of the repeatedly documented reproducibility ofthe Sr isotope trend, even taking into account thediffering waterrrock ratios for these two elements ina diagenetic system.

ŽFurthermore, all isotope pairs crosscorrelate posi-.tively or negatively at very high confidence levels

Ž .Fig. 17 . We are not aware of any diagenetic sce-nario that would have been capable of producingsuch interrelationships. They can be, however, un-derstood in terms of the primary signal that reflects


Fig. 16. Crossplot of mean 87Srr86Sr vs. d18 O values for 20 Ma intervals of the Phanerozoic. Mean values calculated from Figs. 4 and 13,respectively. The correlation coefficient is y0.57 for all points and y0.84 if the two Cenozoic points are excluded. Insert, based on theextension of the trend, suggests that seawater lies on a mixing line of two endmembers, primary magmatic water and meteoric water. The

Ž .composition of endmembers from Hoefs 1997 .

the operation of a unified and hierarchical exogenicŽ . Žsystem litho-, hydro-, atmo-, biosphere Veizer,

.1988 , with tectonics driving it on geological timescales via its control of biogeochemical cycles, inthis case the cycles of Sr, C, O and S. This proposi-

Ž .tion is supported by factor analysis Table 2 whichshows that the system is driven by three factors. Thefirst factor links oxygen and strontium isotopic evo-

Ž . 87 86lution Fig. 16 and the second one Srr Sr and34 Ž .d S Fig. 18 . These two factors explain up to 63%

of the total variance. The last factor links the carbonŽ .and sulfur isotopic evolution Fig. 19 . Because of

the high 87Srr86Sr loadings, the tentative interpreta-tion identifies the first two factors as tectonic, gener-ated by the relative importance of the hydrothermalflux within the oceanic crust to its complementary

Ž .meteoric component riverine flux . The loading ofsulfur and oxygen on two separate factors likelyreflects the fact that the 18 Or16O balance is tempera-

Žture-dependent, with a crossover at ;3008C e.g.,. 34 32Gregory, 1991 , while the Sr S is only marginally

dependent on this variable. The third factor is abiologically mediated redox linkage of sulfur and

Žcarbon cycles Garrels and Perry, 1974; Veizer et al.,.1980 . Considering that the ultimate goal of this

contribution is to tackle the primary vs. secondary

Fig. 17. Correlation coefficients for carbonate 87Srr86 Sr, d18 O,d13C and sulfate d34 S mean values during the Phanerozoic. TheSr, O and C isotope data from Figs. 4, 10 and 13, respectively.

34 Ž .The d S data from Strauss 1997, this issue . The means andtheir correlation coefficients have been calculated for 1 to 40 Maintervals. The shading encloses the 95% confidence level domain.


Table 2Factor analysis of 87Srr86 Sr, d13C, d18 O and d34 S for Phanerozoic seawater

Ž .The input data are 5 Ma averages from Figs. 4, 10 and 13 and from Strauss 1997, this issue , respectively. This resolution was selectedŽ .because it represents the realistic stratigraphic resolution of our database. Note, however, that other reasonable resolutions 1–40 Ma yield

Ž . Ž . 87 86similar outcomes cf. Fig. 17 . Calculated without and with in parentheses the anomalous Srr Sr data for the last 35 Ma.

Variable Factor 1 Factor 2 Factor 387 86 ( ) ( ) Ž .Srr Sr I0.83 I0.64 0.55 0.63 y0.06 y0.0518 ( ) Ž . Ž .d O 0.90 0.75 y0.04 y0.11 0.31 0.2834 Ž . ( ) ( )d S y0.15 y0.11 0.77 0.73 I0.37 I0.4113 Ž . Ž . ( )d C 0.17 0.21 y0.24 y0.21 0.63 0.65

Ž . Ž . Ž .Percent of variance explained 38.9 25.7 23.9 24.8 15.9 16.8

controversy for the d18 O record, we prefer to refrainfrom further discussion of the subject. At this stage,we only wish to reiterate that the isotope data are notconsistent with the diagenetic scenario.

The pattern in Fig. 16 shows that the latest ;35Ma deviate from the overall correlation, suggestingthat either the d18 O or the 87Srr86Sr of moderncycles are not in a steady state. Since this departureis visible only for Sr isotope data, it is likely that theSr cycle is the culprit in terms of its isotopic compo-sition. This may have a bearing on the controversyregarding the role of the Himalayas in the steep rise

87 86 Ž .in Srr Sr of the Cenozoic seawater Fig. 4 . If theunroofing of the Himalayas were indeed the cause,the rise in seawater 87Srr86Sr could have been ac-complished either by an enhanced riverine Sr fluxŽ .Raymo and Ruddiman, 1992 andror by its more

Fig. 18. Crossplot of mean 87Srr86 Sr and d34 S values for 20 Maintervals of the Phanerozoic. Mean values were calculated from

Ž Ž .Fig. 4 and Strauss 1997, this issue , respectively. Note that thed34 S data are those of evaporitic sulfates with their relatively poortemporal resolution.

Ž .radiogenic nature Edmond, 1992 . The isotopic rela-Ž .tionships described in this contribution Figs. 16–19

support the second alternative.In the above discussion, we reviewed five lines of

evidence for secondary vs. primary nature of theobserved d18 O secular trend. Our attempt to explainit as a secondary phenomenon failed, because it hadto rely in all categories on a set of implausible andinternally inconsistent special pleadings or could notbe reconciled at all. The primary alternative, on theother hand, can be reconciled with all experimentalobservations. Note that we do not wish to claim thatall samples are as well-preserved as those in Fig. 1.All screening techniques, singly or in concert, havetheir limitations and post-depositional alteration mayindeed be partly responsible for some of the 18 Odepleted ‘troughs’ in Fig. 13, such as the mid-

Fig. 19. Crossplot of mean d13C and d34 S values for 20 Maintervals of the Phanerozoic. Mean values were calculated from

Ž .Fig. 10 and Strauss 1997, this issue , respectively. Note also thatthis factor is linked to the tectonic factor 2 via its sulfur and not

Ž .the carbon cycle Table 2 .


Carboniferous. Nonetheless, except for these out-liers, the cumulative optical, geological, mineralogi-cal, chemical and isotopic evidence for the primarynature of the Phanerozoic secular d18 O trend, at leastin its major features, is compelling.

For the sake of a rounded discussion, we alsoconsider two objections frequently raised against theprimary nature of the d18 O secular trend.

Ž .1 The brachiopods define a secular trend thatoverlaps with that of the whole rocks, the latter

Ž .undoubtedly isotopically reset Land, 1995 .Ž .2 The studies of oceanic crust and of ophiolites

Že.g., Muehlenbachs and Clayton, 1976; Gregory and.Taylor, 1981 suggest that the magnitudes of high-

and low-temperature hydrothermal processes alongtheir depth profile are about equal and thus canceltheir opposing impact on the d18 O of circulatingwater. As a result, the seawater d18 O is, and hasbeen, buffered around its present day value of about

Ž .0‰ SMOW Gregory, 1991 .Concerning the first point, we already argued that

diagenetic stabilization of shelf carbonate facies ismostly related to transformation of metastable poly-

Ž .morphs aragonite, high-Mg calcite into stable onesŽ .low-Mg calcite and usually happens very early inthe post-depositional history of the rocks, at timeswhen diagenetic solutions still retain partial memory

Žof their seawater precursors Choquette and James,.1990; James and Choquette, 1990 . This often limits

Ž .the whole rock shift to about y2‰ Fig. 20 . Oncestabilized, the subsequent 18 O depletions are muchless pronounced. Such a shift can well be accommo-dated within the d18 O secular band, particularly sincethe major rock components, the micritic matrix andthe early, marine, cements, had initial d18 O composi-

Žtions around and above the top boundary e.g., Milli-.man, 1974; Bathurst, 1975; Tucker and Wright, 1990

of the band based on skeletal precipitates.Additional support for the claim that the bulk of

diagenetic stabilization of shelf carbonates usuallyceases at a rather early stage of diagenetic history isprovided by sulfur isotope data. Kampschulte and

Ž . 34Strauss 1998 studied the d S of trace sulfate frombrachiopods and their host rocks from this sameBochumrOttawa collection. The d34S of the shells,as well as of the host rocks, all plot on the Phanero-zoic d34S secular trend for evaporites. Such coher-ence can only be generated if the system ‘shutsdown’ at a very early diagenetic stage, at times whenpore waters still retain the memory of seawatercomposition.

Apart from the argument for the early diageneticoverprint, we wish to point out the misunderstand-ings that frequently arise from model images of

13 18 Ž .Fig. 20. The average formational d C and d O for the Silurian strata at Gotland. The brachiopod data are from Wenzel 1994 and theŽ .whole rock data from Jux and Steuber 1992 . Note that except for the Burgsvik Formation that contains a disproportionate share of late

meteoric cements, the d13C means are comparable, while the d18 O’s of whole rocks are depleted consistently by ;2‰ relative tobrachiopods. Courtesy of W. Buggisch and B. Wenzel.


diagenesis. The reigning models of bulk rock solu-Žtionrprecipitation, diffusionradvection type e.g.,

.Richter and DePaolo, 1987; Schrag, this issue haveonly limited applicability for diagenesis of multi-component carbonate assemblages. These models,like their groundwater flow system precursors envogue for the shelf facies in the 1960s and 1970s,predict well the chemical and isotopic gradients inpore waters, but not the chemistry of the diversesolid phases. The reason is that only a minusculeportion of the solids is required to generate the porewater gradients. Yet, it is not the millionth portion ofevery particle that reacts, at any instant, with thewater, but only the most soluble millionth particle.This is repeated with the next most soluble particle,etc. As a result, the stabilized bulk rock consists ofdomains that may have been completely, even re-peatedly, recrystallized, while neighbouring domains

Žcould have remained practically unscathed cf. Brand.and Veizer, 1980 . This is the reason for the excel-

lent textural, chemical and isotopic preservation ofsome components, such as the presently studied LMCshells.

Concerning the second argument, we concur thatŽthe oceanic ridge circulation Muehlenbachs and.Clayton, 1976; Gregory, 1991 is the, or one of

several, controlling factor of d18 O composition ofseawater. Accepting this, the Phanerozoic trend canbe reproduced, providing the magnitude of high-tem-perature alteration is kept consistently higher than

Ž .that of its low-temperature analogue Fig. 21 . Modi-fication of the ‘balanced’ model could therefore be asolution to the dilemma. Theoretically, the requiredimbalance could be created by either assigning agreater magnitude to the high-T flux or a lesser oneto the low-T flux. The former may result fromprocesses within the oceanic crust itself, such as

18 Žthose that generate the O depleted gabbros Lecuyer.and Reynard, 1996 , or at magma chambers of sub-

duction domains. Alternatively, the diminished low-Tflux may be a consequence of a decline in continen-tal weathering input.

A corollary to this argument is the claim thatalteration trends in ophiolites ‘prove’ the bufferingof d18 O of ancient seawater at near modern values.In reality, these trends prove only that the waterswere depleted in 18 O relative to the mineral assem-blages of the ophiolites. At this stage, considering

Fig. 21. Theoretical model of hydrothermal alteration required toproduce the entire d18 O secular trend in terms of changing seawater 18 Or16O ratios. The model calculated evolution of the ratioR between the 18 O flux consumed by the weathering of seafloorbasalts at low temperature and the 18 O flux produced by the hightemperature weathering that is needed to reproduce an increase ofabout 8‰ in the d18 O of seawater over the Phanerozoic. Thecurve was obtained by using a modified version of the ICM modelŽ .Godderis and François, 1996 , including the global alkalinity,´carbon, d18 O, d13C and 87Srr86 Sr budgets. The result indicatesthat the R should be about 40–50% in the course of the Phanero-zoic. Continental contribution to the oxygen cycle was also takeninto account.

the large scatter in the alteration patterns, we areneither in a position to demonstrate unequivocallythe equality of the low- and high-temperature fluxes,nor to differentiate between alteration halos gener-ated by water with d18 O of 0‰ instead of say y5‰SMOW.

We wish to emphasize that despite all the aboveargumentation, we do not suggest that the entireobserved d18 O secular trend would have to be ac-commodated by changes in isotopic composition of

Ž .ancient seawater cf. Azmy et al., 1998 . At thisstage, we only advocate a reassessment of discrepan-cies between theory and experimental data, an alter-native far preferable to the wholesale dismissal ofthe latter.

11. Conclusions

Optical, mineralogical, chemical and isotopicstudies of a large collection of newly sampled

ŽPhanerozoic low-Mg calcitic fossils brachiopods and.belemnites , complemented by geological considera-


tions, demonstrate that the bulk of the shells arewell-preserved and likely retained their primary iso-tope signals. The newly generated 87Srr86Sr curvebased on low-Mg fossils and conodonts resembles

Ž .the one by Burke et al. 1982 , but with greater detailand less dispersion.

The d13C curve for the Phanerozoic shows ageneral increase from about y1"1‰ to q4"2‰PDB during the Paleozoic, an abrupt drop of 2‰ atthe PermianrTriassic transition and oscillationsaround q2‰ in the course of the Mesozoic andCenozoic. Superimposed on this general trend arehigher-order oscillations, dominantly of global sig-nificance.

The d18 O shows a general increase from abouty8‰ to 0‰ PDB in the course of the Phanerozoic.Superimposed on this overall trend are second-orderoscillations of ;150 Ma duration, their apexes coin-cident with cold intervals.

The means of Phanerozoic 87Srr86Sr, d13C, d18 Oand d34S correlate at any resolution in excess ofsulfate1 Ma, suggesting that we are dealing with a unified

Ž .exogenic system litho-, hydro-, atmo-, biospheredriven by tectonic forces. Factor analysis shows thatthe above variables are controlled by three factors.The first factor is characterized by reciprocal load-ings of 87Srr86Sr and d18 O, the second by positiveloadings of 87Srr86Sr and d34S, and the third one byreciprocal loadings of d34S and d13C. We interpretthe first two factors as tectonic and the third one as abiologically mediated redox balance of carbon andsulfur cycles.

Acknowledgements

This study has been supported financially by theŽDeutsche Forschungsgemeinschaft particularly by

the endowment from the G.W. Leibniz Prize to Jan.Veizer and by the Natural Sciences and Engineering

Research Council of Canada. Financial support ofthe Canadian Institute for Advanced Research,through its ‘Earth System Evolution’ program, en-abled extensive exchange of ideas with copartici-pants. The establishment of the NSERCrNorandarCIAR ‘Earth System’ Industrial Research Chair atthe University of Ottawa partially freed J. Veizerfrom daily responsibilities. In addition, G.G. Hatch

provided financial support for the isotope laborato-ries. We thank U. Brand, W. Buggisch, S.J. Carpen-ter, H.S. Chafetz, E.L. Grossman, J.D. Hudson, H.-S.Mii, N.B. Popp, B. Wenzel and J. Zachos for helpwith acquisition of literature data, P. Wickham fortechnical assistance, and J. Hoefs, J.D. Hudson andL.P. Knauth for review of the manuscript.

References

Adlis, D.S., Grossman, E.L., Yancey, T.E., McLerran, D.R.,1988. Isotope stratigraphy and paleodepth changes of Pennsyl-vanian cyclical sedimentary deposits. Palaios 3, 487–506.

Ala, D., 1996. High resolution isotope stratigraphy of the LowerDevonian. MSc thesis, University of Ottawa, Ottawa, Canada.

Algeo, T.J., Wilkinson, B.H., 1988. Periodicity of mesoscalePhanerozoic sedimentary cycles and the role of Milankovitchorbital modulation. J. Geol. 96, 313–322.

Anders, M.H., Krueger, S.W., Sadler, P.M., 1987. A new look atsedimentation rates and the completeness of the stratigraphicrecord. J. Geol. 95, 1–14.

Anderson, T.F., Popp, B.N., Williams, A.C., Ho, L.-Z., Hudson,J.D., 1994. The stable isotopic record of fossils from the

Ž .Peterborough Member, Oxford Clay Formation Jurassic , UK:paleoenvironmental implications. J. Geol. Soc. London 151,125–138.

Arthur, M.A., Dean, W.E., Schlanger, S.O., 1985. Variations inthe global carbon cycle during the Cretaceous related toclimate, volcanism, and changes in atmospheric CO . Geo-2phys. Monogr. 32, 504–529.

Azmy, K., 1996. Isotopic composition of Silurian brachiopods:implications for coeval seawater PhD Thesis, University ofOttawa, Ottawa, Canada.

Azmy, K., Veizer, J., Bassett, M.G., Copper, P., 1998. Oxygenand carbon isotopic composition of Silurian brachiopods: im-plications for seawater isotopic composition and glaciation.Geol. Soc. Am. Bull. 110, 1499–1512.

Azmy, K., Veizer, J., Wenzel, B., Bassett, M.G., Copper, P.,1999. Silurian strontium isotope stratigraphy. Geol. Soc. Am.Bull. 111, 475–483.

Baertschi, P., 1957. Messung und Deutung relativer Haufig-¨keitsvariationen vom O18 und C13 in Karbonatgesteinen undMineralien. Schweiz. Mineral. Petrogr. Mitt. 37, 73–152.

Banner, J.L., 1995. Application of the trace element and isotopegeochemistry of strontium to studies of carbonate diagenesis.Sedimentology 42, 805–824.

Banner, J.L., Hanson, G.N., 1990. Calculation of simultaneousisotopic and trace element variations during water–rock inter-action with application to carbonate diagenesis. Geochim.Cosmochim. Acta 54, 3123–3138.

Banner, J.L., Kaufman, J., 1994. The isotopic record of oceanchemistry and diagenesis preserved in non-luminescent bra-chiopods from Mississippian carbonate rocks, Illinois andMissouri. Geol. Soc. Am. Bull. 106, 1074–1082.


Barrera, E., Baldauf, J., Lohmann, K.C., 1993. Strontium isotopeand benthic foraminifer stable isotope results from Oligocenesediment at site 803. Proc. ODP, Sci. Res. 130, 269–279.

Bathurst, R.G.C., 1975. Carbonate Sediments and their Diagene-sis. Elsevier, Amsterdam.

Berner, R.A., 1994. GEOCARB II: a revised model of atmo-spheric CO over Phanerozoic time. Am. J. Sci. 294, 56–91.2

Berner, R.A., Lasaga, A.C., Garrels, R.M., 1983. The carbonate–silicate geochemical cycle and its effect on atmospheric car-bon dioxide over the past 100 million years. Am. J. Sci. 283,641–683.

Bertram, C.J., Elderfield, H., Aldridge, R.J., Morris, S.C., 1992.87Srr86Sr, 143Ndr144 Nd and REEs in Silurian phosphaticfossils. Earth Planet. Sci. Lett. 113, 239–249.

Bowring, S.A., Erwin, D.H., 1998. A new look at evolutionaryrates in deep time: uniting paleontology and high-precisiongeochronology. GSA Today 8, 1–8.

Brand, U., 1989a. Aragonite–calcite transformation based onPennsylvanian molluscs. Geol. Soc. Am. Bull. 101, 377–390.

Brand, U., 1989b. Biogeochemistry of late Paleozoic North Amer-ican brachiopods and secular variation of seawater composi-tion. Biogeochemistry 7, 159–193.

Brand, U., 1991. Strontium isotope diagenesis of biogenic arago-nite and low-Mg calcite. Geol. Soc. Am. Bull. 101, 377–390.

Brand, U., Legrand-Blain, M., 1993. Paleoecology and biogeo-chemistry of brachiopods from the Devonian–Carboniferousboundary interval of the Griotte Formation, La Serre, Mon-tagne Noire, France. Ann. Soc. Geol. Belg. 115, 497–505.´

Brand, U., Veizer, J., 1980. Chemical diagenesis of a multicompo-nent carbonate system: 1. Trace elements. J. Sediment. Petrol.50, 1219–1236.

Brand, U., Veizer, J., 1981. Chemical diagenesis of a multicompo-nent carbonate system: 2. Stable isotopes. J. Sediment. Petrol.50, 987–997.

Brass, G.W., 1976. The variation of the marine 87Srr86 Sr ratioduring Phanerozoic time: interpretation using a flux model.Geochim. Cosmochim. Acta 40, 721–730.

Brass, G.W., Southam, J.R., Peterson, W.N., 1982. Warm salinebottom water in the ancient ocean. Nature 296, 620–623.

Brenchley, P.J., Carden, G.A.F., Marshall, J.D., 1995. Environ-mental changes associated with the ‘first strike’ of the lateOrdovician mass extinction. Mod. Geol. 20, 69–82.

Bruckschen, P., Veizer, J., 1997. Oxygen and carbon isotopiccomposition of Dinantian brachiopods: paleoenvironmentalimplications for the Lower Carboniferous of western Europe.Paleogeogr., Paleoclimatol., Paleoecol. 132, 243–264.

Bruckschen, P., Bruhn, F., Meijer, J., Stephan, A., Veizer, J.,1995a. Diagenetic alteration of calcitic fossil shells: proton

Ž .microprobe PIXE as a trace element tool. Nucl. Instrum.Methods Phys. Res. B 104, 427–431.

Bruckschen, P., Bruhn, F., Veizer, J., Buhl, D., 1995b. 87Srr86 Srisotope evolution of the Lower Carboniferous seawater: Di-nantian of western Europe. Sediment. Geol. 100, 63–81.

Bruckschen, P., Oesmann, S., Veizer, J., 1999. Isotope stratigra-phy of the European Carboniferous: proxy signals for oceanchemistry, climate and tectonics. Chem. Geol. 161, 127–163.

Bruhn, F., 1995. Kombinierte Spurenelement-Mikroanalysen und

Kathodolumineszenz-Untersuchungen: Entwicklung einerŽ .Messmethodik fur die Bochumer Protonenmikrosonde PIXE¨

und Fallstudien aus der Sedimentologie. RNDr Thesis, Ruhr–Universitat, Bochum, Germany.¨

Bruhn, F., Bruckschen, P., Richter, D.K., Meijer, J., Stephan, A.,Veizer, J., 1995. Diagenetic history of sedimentary carbonates:constraints from combined cathodoluminescence and trace ele-ment analyses by micro-PIXE. Nucl. Instrum. Methods Phys.Res. B 104, 409–414.

Bruhn, F., Korte, C., Meijer, J., Stephan, A., Veizer, J., 1997.Trace element concentrations in conodonts measured by theBochum proton microprobe. Nucl. Instrum. Methods Phys.Res. B 130, 630–640.

Burke, W.H., Denison, R.E., Hetherington, E.A., Koepnick, R.B.,Nelson, H.F., Otto, J.B., 1982. Variation of seawater 87Srr86 Srthroughout Phanerozoic time. Geology 10, 516–519.

Carden, G.A.F., 1995. Stable isotopic changes across the Ordovi-cian–Silurian boundary. PhD Thesis, Univ. of Liverpool, Liv-erpool, England.

Ž .Carls, P., 1988. The Devonian of Celtiberia Spain and Devonianpaleogeography of SW Europe. In: McMillan, M.J., Embry,

Ž .A.F., Glass, D.J. Eds. , Devonian of the World. Mem. Can.Soc. Petrol. Geol., Vol. 14, pp. 421–466.

Carpenter, S.J., Lohmann, K.C., 1995. d18 O and d13C values ofmodern brachiopods. Geochim. Cosmochim. Acta 59, 3749–3764.

Ž 13 18 .Chaisi, S.D., Schmitz, B., 1995. Stable d C, d O and stron-Ž87 86 .tium Srr Sr isotopes through the Paleocene at Gebel

Aweina, eastern Tethyan region. Paleogeogr., Paleoclimatol.,Paleoecol. 116, 103–129.

Choquette, P.W., James, N.P., 1990. Limestones — the burialdiagenetic environment. In: McIlreath, I.A., Morrow, D.W.Ž .Eds. , Diagenesis. Geosci. Can. Repr. Ser., Vol. 4, pp. 75–112.

Clayton, R.M., Degens, E.T., 1959. Use of C isotope analyses fordifferentiating freshwater and marine sediments. AAPG Bull.42, 890–897.

Coleman, M.L., Walsh, J.N., Benmore, R.A., 1989. Determinationof both chemical and stable isotope composition in milligram-size carbonate samples. Sediment. Geol. 65, 233–238.

Compston, W., 1960. The carbon isotopic composition of certainmarine invertebrates and coals from the Australian Permian.Geochim. Cosmochim. Acta 18, 1–22.

Copper, P., 1995. Five new genera of Late Ordovician–earlySilurian brachiopods from Anticosti Island, eastern Canada. J.Paleontol. 69, 846–862.

Cummins, D.I., Elderfield, H., 1994. The strontium isotopic com-position of Brigantianrlate Dinantian seawater. Chem. Geol.118, 255–270.

Dash, E.J., Biscaye, P.E., 1971. Isotope composition of strontiumin Cretaceous–Recent pelagic foraminifera. Earth Planet. Sci.Lett. 11, 201–204.

Degens, E.T., Epstein, S., 1962. Relationship between 18 Or16Oratios in coexisting carbonates, cherts and diatomites. Am.Assoc. Petrol. Geol. Bull. 46, 534–542.

Delaney, M.L., Popp, B.N., Lepzelter, C.G., Anderson, T.F.,1989. Lithium-to-calcium ratios in modern Cenozoic, and


Paleozoic articulate brachiopod shells. Paleoceanography 4,681–691.

Denison, R.E., Koepnick, R.B., Burke, W.H., Hetherington, E.A.,Fletcher, A., 1994. Construction of the Mississipian, Pennsyl-vanian and Permian seawater 87Srr86 Sr curve. Chem. Geol.Isot. Geosci. Sect. 112, 145–167.

DePaolo, D.J., Ingram, B.L., 1985. High-resolution stratigraphywith strontium isotopes. Science 227, 938–941.

Derry, L.A., Kaufman, A.J., Jacobsen, S.B., 1992. Sedimentarycycling and environmental change in the Late Proterozoic:evidence from stable and radiogenic isotopes. Geochim. Cos-mochim. Acta 56, 1317–1329.

Dia, A.N., Cohen, A.S., O’Nions, R.K., Shackleton, N.J., 1992.Seawater Sr isotope variation over the past 300 kyr andinfluence of global climate cycles. Nature 356, 786–788.

Diener, A., 1991. Variationen im Verhaltnis von 87Srr86 Sr in¨Brachiopodenschalen aus dem Devon der Eifel. Diplom Ar-beit, Ruhr-Universitat, Bochum, Germany.¨

Diener, A., Ebneth, S., Veizer, J., Buhl, D., 1996. Strontiumisotope stratigraphy of the middle Devonian: brachiopods andconodonts. Geochim. Cosmochim. Acta 60, 639–652.

Dingus, L., 1984. Effects of stratigraphic completeness on inter-pretations of extinction rates across the Cretaceous–Tertiaryboundary. Paleobiology 10, 420–438.

Douglas, R.G., Savin, S.M., 1971. Isotopic analysis of planktonicforaminifera from the Cenozoic of northwest Pacific, Leg 6.Init. Rep. Deep See Drill. Proj. 6, 1123–1127.

Douglas, R.G., Savin, S.M., 1973. Oxygen and carbon isotopeanalyses of Cretaceous and Tertiary foraminifera from CentralNorth Pacific. Init. Rep. Deep Drill. Proj. 17, 591–605.

Ebneth, S., 1991. Variationen im Verhaltnis von 87Srr86 Sr in¨Conodonten aus dem Devon der Eifel. Diplom Arbeit, Ruhr-Universitat, Bochum, Germany.¨

Ebneth, S., Diener, A., Buhl, D., Veizer, J., 1997. Strontiumisotope systematics of conodonts: Middle Devonian, EifelMountains, Germany. Paleogeogr., Paleoclimatol., Paleoecol.132, 79–96.

Edmond, J.M., 1992. Himalayan tectonics, weathering processesand the strontium isotope record in marine limestones. Science258, 1594–1597.

Elderfield, H., 1986. Strontium isotope stratigraphy. Paleogeogr.,Paleoclimatol., Paleoecol. 57, 71–90.

Farrell, J.W., Clemens, S.C., Gromet, L.P., 1995. Improvedchronostratigraphic reference curve of late Neogene seawater87Srr86Sr. Geology 23, 403–406.

Faure, G., 1986. Principles of Isotope Geology. Wiley, New York.Fauville, A., 1995. Der Nachweis der analytischen und strati-

graphischen Reproduzierbarkeit von Strontiumisotopenvaria-tionen hoherer Ordnung. Diplom Arbeit, Ruhr-Universitat,¨ ¨Bochum, Germany.

Frakes, L.A., Francis, J.E., Syktus, J.I., 1992. Climate Modes ofthe Phanerozoic: The History of the Earth’s Climate over thePast 600 Million Years. Cambridge Univ. Press, Cambridge.

Garrels, R.M., Perry, E.A., 1974. Cycling of carbon, sulfur andŽ .oxygen throughout geologic time. In: Goldberg, E.D. Ed. ,

The Sea. Wiley-Interscience, New York, pp. 303–336.Godderis, Y., François, L.M., 1996. Balancing the Cenozoic´

carbon and alkalinity cycles: constraints from isotopic records.Geophys. Res. Lett. 23, 3743–3746.

Golks, C., 1995. Strontiumisotopentrends hoherer Ordnung, glob-¨ales Signal oder lokale Trends? Diplom Arbeit, Ruhr-Uni-versitat, Bochum, Germany.¨

Grahn, Y.G., Caputo, M.V., 1992. Early Silurian glaciation inBrazil. Paleogeogr., Paleoclimatol., Paleoecol. 99, 9–15.

Gregory, R.T., 1991. Oxygen isotope history of seawater revis-ited: composition of seawater. In: Taylor, H.P., Jr., O’Neil,

Ž .J.R., Kaplan, I.R. Eds. , Stable Isotope Geochemistry: aTribute to Samuel Epstein. Geochem. Soc. Spec. Publ., Vol. 3,pp. 65–76.

Gregory, R.T., Taylor, H., 1981. An oxygen isotope profile in asection of Cretaceous oceanic crust, Samail Ophiolite, Oman:

18 Ž .evidence for d O buffering of the oceans by deep 5 kmseawater hydrothermal circulation at mid-ocean ridges. J. Geo-phys. Res. 86, 2737–2755.

Grossman, E.L., Zhang, C., Yancey, T.E., 1991. Stable-isotopestratigraphy of brachiopods from Pennsylvanian shales inTexas. Geol. Soc. Am. Bull. 103, 953–965.

Grossman, E.L., Mii, H.-S., Yancey, T.E., 1993. Stable isotopesin late Pennsylvanian brachiopods from the United States:implications for Carboniferous paleoceanography. Geol. Soc.Am. Bull. 105, 1284–1392.

Grossman, E.L., Mii, H.-S., Zhang, C., Yancey, T.E., 1996.Chemical variation in Pennsylvanian brachiopod shells —effects of diagenesis, taxonomy, microstructure and paleoenvi-ronment. J. Sediment. Res. 66, 1011–1022.

Harland, W.B., Armstrong, R.L., Cox, A.V., Craig, L.E., Smith,A.G., Smith, D.G., 1990. A Geologic Time Scale. CambridgeUniv. Press, Cambridge.

Hayes, J.M., Strauss, H., Kaufman, A.J., this issue. The abun-dance of 13C in marine organic matter and isotopic fractiona-tion in the global biochemical cycle of carbon during the past800 Ma. Chem. Geol.

Hess, J., Bender, M.L., Schilling, J.-G., 1986. Evolution of theratio of Strontium-87 to Strontium-86 in seawater from Creta-ceous to present. Science 231, 979–984.

Hess, J.L., Scott, D., Bender, M.L., Kennett, J.P., Schilling, J.-G.,1989. The Oligocene marine microfossil record: age assess-ments using strontium isotopes. Paleoceanography 41, 655–679.

Hodell, D.A., Woodruff, F., 1994. Variations in strontium isotopicratio of the seawater during the Miocene: stratigraphic andgeochemical implications. Paleoceanography 9, 405–426.

Hodell, D.A., Mueller, P.A., McKenzie, J.A., Mead, G.A., 1989.Strontium isotope stratigraphy and geochemistry of the lateNeogene ocean. Earth Planet. Sci. Lett. 92, 165–178.

Hodell, D.A., Mead, G.A., Mueller, P.A., 1990. Variation in theŽ .strontium isotopic composition of seawater 8 Ma to recent :

implications for chemical weathering states and dissolvedfluxes to the oceans. Chem. Geol. 80, 291–307.

Hodell, D.A., Mueller, P.A., Garrido, J.R., 1991. Variations in thestrontium isotopic composition of seawater during the Neo-gene. Geology 19, 24–27.

Hoefs, J., 1997. Stable Isotope Geochemistry. Springer-Verlag,Berlin.


Hudson, J.D., Anderson, T.F., 1989. Ocean temperatures andisotopic compositions through time. R. Soc. Edinburgh Trans.Earth Sci. 80, 183–192.

Jacobsen, S., Kaufman, A.J., this issue. The Sr, C and O evolutionof Neoproterozoic seawater. Chem. Geol.

James, N.P., Choquette, P.W., 1990. Limestones — the meteoricdiagenetic environment. In: McIlreath, I.A., Morrow, D.W.Ž .Eds. , Diagenesis. Geosci. Can. Repr. Ser., Vol. 4, pp. 35–73.

James, N.P., Bone, Y., Kyser, T.K., 1997. Brachiopod d18 Ovalues do reflect ambient oceanography: Lacepede Shelf,southern Australia. Geology 25, 551–554.

Jasper, T., 1999. Strontium-, Kohlenstoff- und Sauerstoff-iso-topische Entwicklung des Meerwassers: Perm. RNDr. Thesis,Ruhr-Universitat, Bochum, Germany.¨

Jones, C.E., 1992. Strontium isotopes in Jurassic and Early Creta-ceous seawater. PhD Thesis, University of Oxford, Oxford,England.

Jones, C.J., Jenkyns, H.C., Coe, A.L., Hesselbo, S., 1994a. Stron-tium isotopic variations in the Jurassic and Cretaceous seawa-ter. Geochim. Cosmochim. Acta 58, 3061–3074.

Jones, C.E., Jenkyns, H.C., Hesselbo, S.P., 1994b. Strontiumisotopes in the early Jurassic seawater. Geochim. Cosmochim.Acta 58, 1285–1305.

Jux, U., Steuber, T., 1992. C and C — Isotopenverhaltnisse¨carb orgin der silurischen Schichtenfolge Gotlands als Hinweis aufMeeresspiegelschwankungen und Krustenbewegungen.Palaontol. Monatsh. 7, 385–413.

Kampschulte, A., Strauss, H., 1996. The sulfur isotopic composi-tion of Jurassic and Cretaceous seawater as deduced fromtrace sulfate in belemnites. J. Conf. Abstr. 1, 303.

Kampschulte, A., Strauss, H., 1998. The isotopic composition oftrace sulphates in Paleozoic biogenic carbonates: implicationsfor coeval seawater and geochemical cycles. Min. Mag. 62A,744–745.

Karhu, J., Epstein, S., 1986. The implication of oxygen isotoperecords in coexisting cherts and phosphates. Geochim. Cos-mochim. Acta 50, 1745–1756.

Kaufman, A.J., Knoll, A.H., Awramik, S.M., 1992. Biostrati-graphic and chemostratigraphic correlation of Neoproterozoicsedimentary successions: upper Tindir Group, northwesternCanada, as a test case. Geology 20, 181–185.

Killingley, J.S., 1983. Effects of diagenetic recrystallization on18 Or16O values of deep-sea sediments. Nature 301, 594–597.

Knauth, L.P., Epstein, S., 1976. Hydrogen and oxygen isotoperatios in nodular and bedded cherts. Geochim. Cosmochim.Acta 40, 1095–1108.

Knoll, A.H., Hayes, J.M., Kaufman, A.J., Swett, K., Lambert,L.B., 1986. Secular variation in carbon isotope ratios fromupper Proterozoic successions of Svalbard and east Greenland.Nature 321, 832–838.

Koepnick, R.B., Burke, W.H., Denison, R.E., Heatherington, E.A.,Nelson, H.F., Otto, J.B., Waite, L.E., 1985. Construction ofthe seawater 87Srr86 Sr curve for the Cenozoic and Creta-ceous: supporting data. Chem. Geol. Isot. Geosci. Sect. 58,55–81.

Kolodny, Y., Epstein, S., 1976. Stable isotope geochemistry ofdeep sea cherts. Geochim. Cosmochim. Acta 40, 1195–1209.

Korte, C., 1999. 87Srr86Sr, d18 O, und d13C — Evolution destriassischen Meerwassers: Geochemische und StratigraphischeUntersuchungen an Conodonten und Brachiopoden. RNDr.Thesis, Ruhr Universitat, Bochum, Germany.¨

Korvin, G., 1992. Fractals in flatland: a romance of -2 dimen-Ž .sions. In: Korvin, G. Ed. , Fractal Models in the Earth

Sciences. Elsevier, Amsterdam, pp. 87–118.Kroopnick, P.M., 1985. The distribution of 13C of CO in the2

world oceans. Deep Sea Research 3, 57–84.Kump, L.R., Arthur, M.A., this issue. Inte

87Sr/86Sr, Î´13C and Î´18O evolution of Phanerozoic seawater · 2016. 12. 2. · volume Hayes...

Documents

Transcript of 87Sr/86Sr, Î´13C and Î´18O evolution of Phanerozoic seawater · 2016. 12. 2. · volume Hayes...