was removed from UCR-20GaGeS-TAEA inthe same experiment.

Ion exchange with NH4� followed by cal-

cination makes it possible to remove extra-framework species at temperatures as low as100°C. For NH4

�-exchanged UCR-20GaGeS-TAEA, a thermogravimetric analysis showedthat the weight loss of 17.2% occurred between80° and 150°C, which is much less than thetemperature range needed for the direct calci-nation of the as-synthesized amine-containingsample (300° to 360°C). An x-ray powder dif-fraction shows that the sample remains highlycrystalline after the calcination of the NH4

�-exchanged sample at 180°C under argon atmo-sphere (fig. S2).

In addition to NH4�, these materials under-

go ion exchange with many mono- and divalentmetal cations. For example, upon exchangewith Cs� ions, the percentages of C, H, and Nin UCR-20GaGeS-TAEA were dramatically re-duced (20). Yet, like the original sample, theexchanged sample remains highly crystalline.The Cs�-exchanged UCR-20GaGeS-TAEAexhibits the type I isotherm characteristic of amicroporous solid (Fig. 4). This sample has ahigh Langmuir surface area of 807 m2/g and amicropore volume of 0.23 cm3/g, despite thepresence of much heavier elements (Cs, Ga, Ge,and S), as compared to the elements present inaluminosilicate zeolites. The median pore di-ameter calculated with the Horvath-Kawazoemethod is 9.5 Å, 14% larger than that for Mo-lecular Sieve Type 13X (8.2 Å) determined un-der the same experimental conditions.

These sulfides are also strongly photolu-minescent and can be excited with wave-lengths from 360 to 420 nm. The emissionmaximum occurs in the range from 460 to508 nm (Fig. 5). For example, UCR-20GaGeS-TAEA strongly luminesces at480 nm when excited at 370 nm. The gen-eral trend is that materials with heavierelements are excited and luminesce at alonger wavelength.

References and Notes1. M. E. Davis, Nature 417, 813 (2002).2. Ch. Baerlocher, W. M. Meier, D. H. Olson, Atlas of

Zeolite Framework Types (Elsevier, Amsterdam,2001).

3. A. Corma, M. J. Diaz-Cabanas, J. Martinez-Triguero, F.Rey, J. Rius, Nature 418, 514 (2002).

4. A. K. Cheetham, G. Ferey, T. Loiseau, Angew. Chem.Int. Ed. 38, 3268 (1999).

5. R. L. Bedard, S. T. Wilson, L. D. Vail, J. M. Bennett, E. M.Flanigen, in Zeolites: Facts, Figures, Future. Proceed-ings of the 8th International Zeolite Conference, P. A.Jacobs, R. A. van Santen, Eds. (Elsevier, Amsterdam,1989), pp. 375–387.

6. C. L. Cahill, J. B. Parise, J. Chem. Soc. Dalton Trans.2000, 1475 (2000).

7. S. Dhingra, M. G. Kanatzidis, Science 258, 1769(1992).

8. R. W. J. Scott, M. J. MacLachlan, G. A. Ozin, Curr. Opin.Solid State Mater. Sci. 4, 113 (1999).

9. H. Li, A. Laine, M. O’Keeffe, O. M. Yaghi, Science 283,1145 (1999).

10. H. Ahari, A. Lough, S. Petrov, G. A. Ozin, R. L. Bedard,J. Mater. Chem. 9, 1263 (1999).

11. O. M. Yaghi, Z. Sun, D. A. Richardson, T. L. Groy, J. Am.Chem. Soc. 116, 807 (1994).

12. K. Tan, A. Darovsky, J. B. Parise, J. Am. Chem. Soc.117, 7039 (1995).

13. C. L. Bowes et al., Chem. Mater. 8, 2147 (1996).14. E. M. Flanigen, in Introduction to Zeolite Science and

Practice, H. van Bekkum, E. M. Flanigen, J. C. Jansen,Eds. (Elsevier, New York, 1991), pp. 13–34.

15. P. Feng, X. Bu, G. D. Stucky, Nature 388, 735 (1997).16. X. Bu, P. Feng, G. D. Stucky, Science 278, 2080 (1997).17. X. Bu, N. Zheng, Y. Li, P. Feng, J. Am. Chem. Soc. 124,

12646 (2002).18. C. Wang, X. Bu, N. Zheng, P. Feng, J. Am. Chem. Soc.

124, 10268 (2002).19. An example of a typical synthesis condition with the

preparation of UCR-20GaGeS-TAEA is as follows.Gallium metal (82.7 mg), germanium oxide (109.1mg), sulfur (222.1 mg), and tris(2-aminoethyl)amine(2.1719 g) were mixed in a Teflon-lined stainlesssteel autoclave for �20 min. The vessel was thensealed and heated at 190°C for 6 days. The autoclavewas subsequently cooled to room temperature.Transparent, pale yellow crystals were obtained witha yield of �67%.

20. Materials and methods, results of the elemental anal-ysis, details of the thin-film structure, and x-ray

powder diffraction data are available on Science On-line. Crystallographic data (excluding structure fac-tors) have been deposited with the Cambridge Crys-tallographic Data Center (CCDC) as supplementarypublications CCDC 197105 through 197126 andCCDC 197207 through 197214.

21. M. O’Keeffe, M. Eddaoudi, H. Li, T. Reineke, O. M.Yaghi, J. Solid State Chem. 152, 3 (2002).

22. G. O. Brunner, W. M. Meier, Nature 337, 146 (1989).23. H. Li, J. Kim, T. L. Groy, M. O’Keeffe, O. M. Yaghi,

J. Am. Chem. Soc. 123, 4867 (2001).24. C. Wang et al., J. Am. Chem. Soc. 123, 11506 (2001).25. Supported by NSF (grant CHE-0213310) and the Air

Force Office of Scientific Research (for a grant topurchase an x-ray powder diffractometer). We thankH. Luo, S. Li, and Y. Yan for their assistance withadsorption measurements.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/298/5602/2366/DC1Materials and MethodsFigs. S1 and S2Table S1

23 September 2002; accepted 12 November 2002

Mass-Independent Sulfur ofInclusions in Diamond and

Sulfur Recycling on Early EarthJ. Farquhar,1 B. A. Wing,1 K. D. McKeegan,2 J. W. Harris,3

P. Cartigny,4 M. H. Thiemens5

Populations of sulfide inclusions in diamonds from the Orapa kimberlite pipein the Kaapvaal-Zimbabwe craton, Botswana, preservemass-independent sulfurisotope fractionations. The data indicate that material was transferred from theatmosphere to the mantle in the Archean. The data also imply that sulfur is notwell mixed in the diamond source regions, allowing for reconstruction of theArchean sulfur cycle and possibly offering insight into the nature of mantleconvection through time.

An understanding of the nature of the sourcematerials for diamonds would provide impor-tant insights into large-scale geophysical pro-cesses. For example, elemental and isotopicdata have been used to argue that diamondsand their inclusions are relics of subductedcrustal materials (1–9), but alternate explana-tions such as mantle fractionation processesor relict primordial heterogeneity are plausi-ble (10–15). Here we report mass-indepen-dently fractionated {anomalous �33S � �33S– 1000 � [(1 � �34S /1000)0.515 – 1]} (16)sulfur isotope compositions for syngeneticsulfide inclusions in diamond from the Orapakimberlite pipe, Kaapvaal-Zimbabwe craton,

Botswana. We also discuss the implicationsof �33S as an almost perfect tracer of theexchange between Earth’s geochemical res-ervoirs because of its exclusive originthrough atmospheric photochemistry and itspreservation through subsequent mass-depen-dent fractionation processes.

The Orapa kimberlite pipe is located with-in the Magondi belt, a region of thick (150 to225 km) crust along the western margin ofthe Kaapvaal-Zimbabwe craton, which isconsidered to be the surface manifestation ofthe Proterozoic reactivation of the Kaapvaal-Zimbabwe craton (17–19). Diamonds fromOrapa are predominantly eclogite types andhave a wide range of �13C values [(–26 to –3per mil (‰)], �15N values (–10 to �6‰),nitrogen contents [(8 to 3450 parts per mil-lion (ppm)], and nitrogen aggregation states(a 0 to 95% degree of association) (1, 12, 20).Silicate and sulfide inclusions from these di-amonds have at least two distinct ages (1, 19):an Archean population of 2.9 Ga and a Pro-terozoic population of 1.0 billion years ago(Ga) (1, 19). The sulfide inclusions also have

1Earth System Science Interdisciplinary Center andDepartment of Geology, University of Maryland, Col-lege Park, MD 20742, USA. 2Department of Earth andSpace Sciences, University of California, Los Angeles,CA 90095, USA. 3Division of Earth Sciences, GregoryBuilding, University of Glasgow, Glasgow, G12 8QQ,UK. 4Laboratoire de Geochimie des Isotopes Stables,Universite Paris VII, Institut de Physique du Globe deParis, 75252 Paris Cedex 05, France. 5Department ofChemistry and Biochemistry, University of California,San Diego, La Jolla, CA 92093, USA.

R E P O R T S

www.sciencemag.org SCIENCE VOL 298 20 DECEMBER 2002 2369

a wide range of �34S (–11 to �9.5‰) thatmay relate to recycling of sedimentary mate-rials (2, 3, 11).

We extracted sulfide inclusions from 12Orapa diamonds for sulfur isotope analysis.These inclusions consist of finely exsolved mo-nosulfide solid solution with micron-scale Ni-rich and Cu-rich exsolution features. One of thehost diamonds (ORJF2) contained silicate in-clusions of eclogite-type (e-type) garnets andclinopyroxene. The other 11 diamonds con-tained only sulfide inclusions. The sulfide in-clusions (table S1) have low Ni contents andalso contain significant amounts of Cu and Co,suggesting an e-type affinity (21). Data forOrapa sulfides are comparable but have a nar-rower range than previous measurements (22).

The sulfur isotope compositions of 23 in-dividual inclusions (45 spot analyses) weremeasured by secondary ion mass spectrome-try with the University of California, LosAngeles Cameca IMS 1270 ion microprobe(table S2). Multiple Faraday cup detectorswere used for simultaneous measurement of32S–, 33S–, and 34S– ion beams (23). Two-sigma uncertainties are �0.12‰ for single�33S analyses and �1.4‰ for �34S spot anal-yses. Our �34S values yield a smaller range(–1.4 to 2.6‰) than has been previously ob-served for Orapa sulfides (2, 3). The smallerrange for both the �34S and the Ni content ofour data relative to that presented by priorstudies suggests that we have analyzed asubset of the Orapa sulfide population. Withone possible exception (ORJF2G4), �33S and�34S analyses of individual sulfide grainsfrom the same diamond are indistinguishablefrom one another within our analytical uncer-tainties. On the basis of homogeneous �33S,we grouped inclusions into populations thatare defined on the basis of the diamond inwhich they are found (Fig. 1). The �33Svalues of different inclusion populations ex-tend from –0.11 to 0.61‰, with �33S anom-alies resolvable at the 2� level from �33S �0 in 4 of the 12 diamonds.

The lack of resolvable �33S variabilityamong inclusions from a single diamond sug-gests that the source of sulfur for each popula-tion of inclusions (i.e., from an individual dia-mond) was well mixed and characteristic of theimmediate region in which the diamondformed. The variations in �33S among sulfidegrains from different diamonds imply that het-erogeneity does exist, but at a scale larger thanthat sampled by any single diamond. Thepresent data are insufficient to determinewhether �33S variability is coupled to otherchemical and isotopic heterogeneity (1, 12, 20).C and N in diamonds and Ni in diamond sulfideinclusions might have a different source thansulfur and may be uncoupled.

Classical thermodynamic, kinetic, and dif-fusion-controlled fractionation processes allproduce highly correlated relations between

�33S and �34S such that �33S typically variesby less than 0.01‰ for every 1‰ variation of�34S. Although minor variations of �33S canbe generated by different types of mass-de-pendent fractionation processes (24–27),these processes cannot account for the largermagnitude of the observed anomalous �33Sof sulfide inclusions and its lack of depen-dence on variations for �34S (Fig. 1). Thisrules out the possibility that mass-dependentfractionation processes in the mantle couldhave caused the �33S anomalies.

We do not favor interpreting the anoma-lous �33S values as primordial heterogene-ities inherited from diverse sources duringaccretion of Earth. Although nonzero �33Svalues have been observed in some rare com-ponents of meteorites (28–32), sequentialacid extracts and bulk extracts of sulfide sul-fur and sulfate sulfur from carbonaceouschondrites, enstatite chondrites, ordinarychondrites, iron meteorites, and ureilites haveindicated that �33S is homogeneous and has arange of values between –0.01 and �0.04‰(28, 29, 33, 34).

Recently, large positive and negative �33Svalues up to a few per mil have been observed

in Archean crustal and sedimentary rocks (16).The origin of these anomalous �33S values isattributed to atmospheric photochemistry in-volving sulfur dioxide in a primitive atmo-sphere with reduced oxygen and ozone andincreased ultraviolet transparency (16, 35).When the data for Archean �33S are plottedversus �34S (Fig. 2), they define an area thatincludes the observed data for sulfide inclusionsfrom Orapa diamonds. The �33S signature ob-served in some sulfide inclusions from Orapadiamonds is consistent with the recycling of thissurface sulfur reservoir.

The preservation of anomalous sulfur inthese inclusions allows us to place constraintson the coupling between Archean mantle,crust, and atmosphere. The mean �33S ofperidotite xenoliths from Kilbourne Hole istaken as an estimate for the mantle composi-tion (0.03 � 0.04‰ 2�). Monte Carlo resa-mpling of the mean �33S value for Archeansulfide (16) yields a �33S of 0.51 � 0.30‰for the average Archean crustal sulfide com-position. The experimental determination ofthe isotopic composition of elemental sulfurproduced by photolysis of sulfur dioxide with193-nm radiation has �33S � 65 � 4‰ (35).

Fig. 1. Plot of�33S and �34Sof inclusion populationsfrom individual diamondsfrom the Orapa kimberlitepipe (red points). Two-sigmamean uncertainty estimatesfor each inclusion popula-tion are presented assumingthat they are given by0.12‰ � n–1/2 and 1.4‰� n–1/2, respectively, wheren is the number of pointsanalyzed and 0.12 and1.4‰ are estimates of the2�uncertainty made on thebasis of long-term reproduc-ibility of the working stan-dard (CAR 123). The shadedregion centered at�33S� 0and extending from �0.12 to –0.12‰ is assumed to represent the mass-dependent field. Four points aredistinguished from this field and are interpreted to bemass-independent. For reasons that are not known, the�34S values of different inclusions define a smaller range than observed previously (2, 3).

Fig. 2. Plot of �33S and�34S of inclusion popula-tions from individual dia-monds from the Orapakimberlite pipe (red cir-cles), with the same as-pect ratio as Fig. 1 thatalso includes data for Ar-chean samples (blue tri-angles) and data for sam-ples younger than 2.45Ga (green diamonds) thatwere reported in (16) and(38) as well as additionalmass-dependent data (ta-ble S3). This plot allowscomparison of the sulfideinclusion data from Orapadiamonds with data from possible sulfur sources.

R E P O R T S

20 DECEMBER 2002 VOL 298 SCIENCE www.sciencemag.org2370

Because the �33S values of the anomalousinclusions are directly comparable to those ofthe average Archean crustal sulfide, our dataindicate that all of the sulfur in these inclu-sions may represent recycled but undilutedcrustal sulfur. Material balance also dictatesthat up to 1% of the sulfur in these inclusionswas processed through the Archean atmo-sphere. This estimate is a lower limit becausewe assumed that the maximum mass-inde-pendent fractionation that was observed inthe photolysis experiments (35) is applicableto the Archean atmosphere. Our detectionlimits of �0.12‰ would also allow for up to25% of the sulfur in the mass-dependent di-amonds to derive from a surface sulfide res-ervoir with a �33S of 0.51‰. The heteroge-neous �33S may reflect different proportionsof surface-derived sulfur, or it may reflectvariable �33S of the recycled component.

A conceptual model of the sulfur cycle in-dicated by the data (Fig. 3) starts with a volca-nic source releasing mass-dependent sulfur-bearing gases into the atmosphere. Photolysisof these gases produces a photochemical frac-tionation of sulfur between elemental and oxi-dized forms, each with a unique and anomalous�33S. The anomalous sulfur is transferred tosurface reservoirs through atmospheric deposi-tion and the elemental sulfur species areconverted to sulfide (possibly by microbial ac-tivity). After sedimentation, lithification, andmetamorphism, the anomalous sulfur is recy-

cled into the mantle. Evidence for these con-nections is only preserved, however, becausethe sulfur was encased in diamond and ulti-mately transported back to Earth’s surface.Mass balance in the sulfur cycle requires anequivalent flux of a 33S-depleted (�33S 0)component or a net decrease of �33S in one ormore of the exospheric sulfur reservoirs, orsome combination of the two. The sign of theanomalous �33S in our four diamonds is posi-tive, like both the elemental sulfur produced by193-nm photolysis and most of the measuredArchean sulfides (16, 35), and unlike the sulfateproduced in the photolysis experiments andobserved in Archean samples (16, 35). If oursamples are representative, the coincidence ofpositive �33S for sulfide inclusions and averageArchean sulfide suggests that the sulfur in theinclusions is predominantly derived from a sed-imentary sulfide component that was subductedto the source regions of the Orapa diamonds.Given that measurements of �33S from Arche-an barites are uniformly negative (16) and thatmeasurements of �33S from rocks with clearoceanic association also carry this signature,accumulation of 33S in the Archean oceanmight be sufficient to close the 33S massbalance.

Transfer of the negative �33S signaturefrom the oceanic sulfate reservoir, through sul-fate reduction during basalt alteration and sub-sequent subduction of the altered basalt, mayprovide a mechanism, however, that satisfies

mass balance constraints. Although a subducted33S-depleted basaltic component is not mandat-ed by our observations, the recognition of po-tential deep-mantle lithospheric graveyards (36,37) leads to the possibility that these regionscontain the missing 33S-depleted component.The presence of 33S-enriched mantle domainswith sedimentary affinities and 33S-depletedmantle domains with oceanic affinities remainsto be tested. These sulfur isotope data constitutea new type of isotope tracer that allows a re-construction of the Archean sulfur cycle, andthey indicate an interconnected global geo-chemical sulfur cycle extending from the atmo-sphere to the mantle on ancient Earth.

References and Notes1. S. B. Shirey et al., Science 297, 1683 (2002).2. M. Chaussidon, F. Albarede, S. M. F. Sheppard, Nature

330, 242 (1987).3. C. S. Eldridge, W. Compston, I. S. Williams, J. W.

Harris, J. W. Bristow, Nature 353, 649 (1991).4. C. B. Smith et al., Geochim. Cosmochim. Acta 55,

2579 (1991).5. R. L. Rudnick, C. S. Eldridge, G. P. Bulanova, Geology

21, 13 (1993).6. S. Aulbach, T. Stachel, K. S. Viljoen, G. P. Brey, J. W.

Harris, Contrib. Mineral. Petrol. 143, 56 (2002).7. T. Stachel, J. W. Harris, S. Aulbach, P. Deines, Contrib.

Mineral. Petrol. 142, 465 (2002).8. S. H. Richardson, S. B. Shirey, J. W. Harris, R. W.

Carlson, Earth Planet. Sci. Lett. 191, 257 (2001).9. N. V. Sobolev, E. M. Galimov, I. N. Ivanovskaia, E. S.

Efimova, Dokl. Akad. Nauk SSSR 249, 1217 (1979).10. P. Deines, J. W. Harris, J. J. Gurney, Geochim. Cosmo-

chim. Acta 51, 1227 (1987).11. J. J. Gurney, Nature 353, 601 (1991).12. P. Deines, J. W. Harris, J. J. Gurney, Geochim. Cosmo-

chim. Acta 57, 2781 (1993).13. P. Cartigny, J. W. Harris, M. Javoy, Science 280, 1421

(1998).14. P. Cartigny, J. W. Harris, D. Phillips, M. Girard, M.

Javoy, Chem. Geol. 147, 147 (1998).15. S. E. Haggerty, Science 285, 851 (1999).16. J. Farquhar, H. M. Bao, M. Thiemens, Science 289, 756

(2000).17. T. K. Nguuri et al., Geophys. Res. Lett. 28, 2501

(2001).18. D. E. James, M. J. Fouch, J. C. VanDecar, S. van der Lee,

Geophys. Res. Lett. 28, 2485 (2001).19. S. B. Shirey et al., Geophys. Res. Lett. 28, 2509

(2001).20. P. Cartigny, J. W. Harris, M. Javoy, in The 7th Inter-

national Kimberlite Conference Proceedings, The J. B.Dawson Volume, J. L. Gurney, J. J. Gurney, M. D.Pascor, S. H. Richardson, Eds. (Red Roof Design, CapeTown, South Africa, 1999), pp. 117–124.

21. G. P. Bulanova, W. L. Griffin, C. G. Ryan, O. Y. Shesta-kova, S. J. Barnes, Contrib. Mineral. Petrol. 124, 111(1996).

22. P. Deines, J. W. Harris, Geochim. Cosmochim. Acta 59,3173 (1995).

23. Materials and methods are available as supportingmaterial on Science Online.

24. J. R. Hulston, H. G. Thode, J. Geophys. Res. 70, 3475(1965).

25. Y. Matsuhisa, J. R. Goldsmith, R. N. Clayton, Geochim.Cosmochim. Acta 42, 173 (1978).

26. E. D. Young, A. Galy, H. Nagahara, Geochim. Cosmo-chim. Acta 66, 1095 (2002).

27. M. F. Miller, Geochim. Cosmochim. Acta 66, 1881(2002).

28. X. Gao, M. H. Thiemens, Geochim. Cosmochim. Acta55, 2671 (1991).

29. ���� , Geochim. Cosmochim. Acta 57, 3159(1993).

30. G. W. Cooper, M. H. Thiemens, T. L. Jackson, S. Chang,Science 277, 1072 (1997).

31. C. E. Rees, H. G. Thode, Geochim. Cosmochim. Acta41, 1679 (1977).

Fig. 3. Components of the geochemical sulfur cycle that are indicated by the data for sulfideinclusions from Orapa. Anomalous �33S of sulfide inclusions in diamond was produced when sulfurdioxide of volcanogenic origin was photodissociated in the Archean atmosphere by deep ultravioletradiation. This signature was transferred first to sedimentary sulfide and subsequently to themantle by subduction-related processes. Diamond formation in the mantle encapsulated thesulfides and preserved them until the point at which they were transported to Earth’s surface.Observations of negative �33S for Archean barite indicate that the oceanic reservoir had negative�33S. Hydrothermal reduction of oceanic sulfate with negative �33S would transfer this signatureto the hydrothermally altered products. Recycling of altered crust to the mantle may thereforeclose the sulfur balance required by the sulfide inclusion data. Sulfides in shales appear to be themost substantial repository for �33S-enriched sulfur (16).

R E P O R T S

www.sciencemag.org SCIENCE VOL 298 20 DECEMBER 2002 2371

32. J. R. Hulston, H. G. Thode, J. Geophys. Res. 70, 4435(1965).

33. X. Gao, M. H. Thiemens, Geochim. Cosmochim. Acta57, 3171 (1993).

34. J. Farquhar, T. L. Jackson, M. H. Thiemens, Geochim.Cosmochim. Acta 64, 1819 (2000).

35. J. Farquhar, J. Savarino, S. Airieau, M. H. Thiemens, J.Geophys. Res. Planets 106, 32829 (2001).

36. F. Albarede, R. D. van der Hilst, Eos 80, 535 (1999).37. R. D. van der Hilst, H. Karason, Science 283, 1885 (1999).

38. D. Heymann et al., Geochim. Cosmochim. Acta 62,173 (1998).

39. We acknowledge and thank DeBeers for providingdiamonds; M. Chaussidon for sulfur isotope stan-dards; A. D. Brandon for providing peridotite samplesfrom Kilbourne Hole; A. Paytan for making her dataavailable to us; NSF, NASA, and American ChemicalSociety for support to J.F. and B.A.W.; and the NASAAstrobiology program for supporting sulfur isotopestudies at UCLA and University of Maryland, CollegePark. The UCLA ion microprobe facility is partially

supported by a grant from the NSF Instrumentationand Facilities program. J.F. acknowledges editing andinsights of L.J. Tuit.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/298/5602/2369/DC1Materials and MethodsTables S1 to S3

20 September 2002; accepted 11 November 2002

Calibration of Sulfate Levels inthe Archean Ocean

Kirsten S. Habicht,1 Michael Gade,1 Bo Thamdrup,1 Peter Berg,2

Donald E. Canfield1*

The size of the marine sulfate reservoir has grown through Earth’s history,reflecting the accumulation of oxygen into the atmosphere. Sulfur isotopefractionation experiments on marine and freshwater sulfate reducers, to-gether with the isotope record, imply that oceanic Archean sulfate con-centrations were 200 M, which is less than one-hundredth of presentmarine sulfate levels and one-fifth of what was previously thought. Such lowsulfate concentrations were maintained by volcanic outgassing of SO2 gas,and severely suppressed sulfate reduction rates allowed for a carbon cycledominated by methanogenesis.

It is thought that the Archean Earth had lowatmospheric oxygen concentrations (1), lowoceanic sulfate concentrations (2), and ele-vated atmospheric concentrations of meth-ane, contributing to possible greenhousewarming of Earth’s surface (3). The biogeo-chemistries of these elements are linked, inthat low atmospheric oxygen levels suppressthe oxidative weathering of sulfides and thedelivery of sulfate to the oceans, contributingto the low sulfate concentrations (2). Lowsulfate levels could have inhibited sulfatereduction, enhancing methane production (2,4).

This reconstruction depends on our abilityto extract reliable sulfate concentration infor-mation from the isotope record of sulfide andsulfate through time. The isotope record revealssmall fractionations of generally 10 per mil(‰) between sulfates and sedimentary sulfidesbefore 2.5 to 2.7 billion years ago (Ga) (2). Thefew available pure culture studies suggest thatfractionations become suppressed at a sulfateconcentration around 1 mM (5, 6). Currentmodels link reduced fractionations at low sul-fate concentration to a limitation of sulfate ex-change across the cell membrane (6). In thiscase, most of the sulfate entering the cell be-comes reduced, and even with substantial inter-nal enzymatic fractionations, minimal net frac-

tionation is expressed. Sulfate limitation alsoreduces sulfate reduction rates, with half-satu-ration constants (km) values for marine strainsof 70 and 200 M (7, 8) and for freshwaterstrains, 5 to 30 M (7). If similar sulfate con-centrations limit both fractionation and sulfatereduction rate, then sulfate reducers shouldmaintain substantial fractionation at sulfate con-centrations considerably less than 1 mM.

In continuous culture, we explored thefractionations at millimolar and submillimo-lar sulfate concentrations by Archaeoglobusfulgidus grown on lactate at its optimalgrowth for temperature of 80°C. A. fulgidus isan archaeon and was chosen to representpossible early sulfate reducers from hydro-thermal settings. We also examined natural

populations of sulfate reducers from a coastalmarine sediment (natural sulfate concentra-tion, 20 mM) and a freshwater lake sediment(natural sulfate concentration, 300 M).Freshwater sulfate reducers are especiallyadapted to low sulfate concentrations (9) andcould reflect the behavior of possible earlylow sulfate–adapted organisms, whereas ma-rine sulfate reducers are adapted to high sea-water salinities. In the natural population ex-periments, sediment was incubated at 17°C ina rapidly recirculating flow-through plug re-actor (10) with lactate (1 mM) as the organicsubstrate (11).

All three different microbial populationsproduced high fractionations (11) of up to32‰ with 200 M or greater sulfate (Fig. 1).The average fractionation for sulfate between200 and 1000 M was 22.6 � 10.3‰, whichis similar to the average for pure bacterialcultures (6) (18 � 10‰) and natural popula-tions (6) (28 � 6‰) of sulfate reducers uti-lizing 20 mM or greater sulfate. By contrast,fractionations were consistently less than 6‰(an average of 0.7 � 5.2‰) with sulfateconcentrations less than 50 M. Thus, sulfatesubstantially limited fractionation up to aconcentration somewhere between 50 Mand around 200 M. This is also the concen-tration range where sulfate limits rates ofsulfate reduction (8, 9).

The isotopic composition of sedimentarysulfides will, in addition to the bacterial frac-tionation, depend on the extent to which sul-fides form in a zone of sulfate depletion (6,

1Danish Center for Earth System Science and Instituteof Biology, University of Southern Denmark, Campus-vej 55, DK-5230, Odense M, Denmark. 2Departmentof Environmental Sciences, Clark Hall, University ofVirginia, VA 22903, USA.

*To whom correspondence should be addressed.

Fig. 1. Isotope fraction-ation as a function ofsulfate concentration forfreshwater (diamonds)and marine (squares)natural populations ofsulfate reducers and forthe hyperthermophile A.fulgidus (triangles). Forthe freshwater and ma-rine populations, hori-zontal bars plot therange of sulfate con-centrations within thereactor, with the higherconcentration enteringthe reactor, and the lowconcentration exitingthe reactor. The sym-bols are positioned onthe bars at the average concentration in the reactor.

R E P O R T S

20 DECEMBER 2002 VOL 298 SCIENCE www.sciencemag.org2372

Supporting Material (2 tables and methods)

Materials and MethodsSulfide inclusions were liberated from diamond by fracturing the diamonds in apercussion mortar. The inclusions were mounted in epoxy along with a workingreference sulfide and polished for SIMS and EPMA analysis. Sulfur isotope data forsulfide inclusions were collected over the course of two analytical sessions (September2001 and March 2002) at UCLA using a Cameca IMS 1270 with a Cs+ ion beam. Theprimary beam was mass-filtered and accelerated to 20 keV impact energy. Typicaloperating conditions included a 20-25 micron spot and ~2 nA current. A low energy,normal incidence, electron flood gun was used for charge compensation. Low energy (0-25 eV initial kinetic energy) negative secondary ions were analyzed with a massresolution of 4000 for all isotopes. Ion beams for 32S, 33S, and 34S were collectedsimultaneously in faraday cups. Integration times were 10 seconds on all ion beams and10 to 15 cycles were collected per analysis. Standards were included on the samplemounts and standard analyses were undertaken between every two or three sampleanalyses. Standards included &DQ Q�'LDEOR�troilite (CDT), CAR 123 and Balmat pyrite,and Anderson pyrrhotite(38). Uncertainties for the isotopic analyses were estimated fromthe external reproducibility of measurements of the working standard (CAR 123 pyrite)made in all analytical sessions. Sulfide inclusions were examined in polished grainmounts from all samples with backscattered electron (BSE) imaging supplemented byqualitative energy dispersive (EDS) analysis using the JEOL JXA-8900 electronmicroprobe at the University of Maryland. EDS analysis revealed the presence ofprimary Fe and S, significant Ni and Cu, and minor Co. X-ray maps showed that Ni andCu were concentrated in sub-µm Ni-rich and Cu-rich domains. Quantitative chemicalanalyses of sulfide inclusions were measured with the electron microprobe usingsynthetic and natural mineral standards and a Phi-Rho-Z data correction scheme.Accelerating voltage was 20 kV, beam current was 50 nA, nominal spot size was ~3 µm,and count times were ~30 s. Reported compositions represent an average of 3-5 ‘spot’analyses per grain.

Table S1 Elemental compositions of sulfide inclusions in Orapa diamonds.S (Wt%) Ni (Wt%) Fe (Wt%) Co (Wt%) Cu (Wt%) Total

ORJF1g1 39.05 1.90 58.75 0.20 0.02 99.93ORJF1g1 39.23 1.61 59.09 0.17 0.03 100.13ORJF1g1 39.00 0.98 59.45 0.12 0.04 99.59ORJF1g2 38.93 0.67 60.00 0.09 0.35 100.03ORJF1g2 38.72 1.88 57.73 0.17 0.98 99.48ORJF1g2 39.11 0.54 60.31 0.09 0.23 100.28ORJF1g2 39.10 0.92 59.54 0.10 0.37 100.02ORJF2g1 39.01 0.59 60.37 0.10 0.09 100.16ORJF2g1 38.05 1.30 54.94 0.25 5.31 99.85ORJF2g1 39.12 0.99 59.70 0.21 0.27 100.29ORJF2g1 38.75 0.46 60.36 0.10 0.42 100.09ORJF2g1 38.86 0.68 60.70 0.16 0.02 100.42ORJF2g1 39.34 0.35 60.50 0.08 0.03 100.31ORJF2g2 38.60 0.86 54.34 0.19 5.43 99.42ORJF2g2 38.78 1.03 59.79 0.24 0.03 99.87ORJF2g2 38.75 1.46 59.76 0.26 0.04 100.27ORJF2g2 38.72 0.64 60.70 0.11 0.01 100.18ORJF2g2 38.52 0.99 56.80 0.19 3.54 100.04ORJF2g3 38.88 0.42 60.30 0.07 0.04 99.71ORJF2g3 38.73 1.31 59.77 0.27 0.01 100.09ORJF2g3 38.87 0.86 60.47 0.14 0.04 100.38ORJF2G4 39.04 0.43 60.88 0.07 0.03 100.44ORJF2g4 39.36 0.88 59.57 0.18 0.04 100.02ORJF2G4 38.70 3.35 57.34 0.70 0.06 100.15ORJF2g4 39.38 1.46 58.90 0.28 0.02 100.05ORJF2G4 38.97 0.80 60.59 0.20 0.02 100.57ORJF2g4 39.38 1.23 59.04 0.23 0.01 99.89ORJF2G4 38.50 3.39 57.70 0.74 0.03 100.37ORJF2g4 39.39 1.62 58.88 0.28 0.02 100.19ORJF3g2 38.53 5.16 54.50 0.19 1.98 100.36ORJF3g2 38.57 6.56 53.15 0.29 1.53 100.10ORJF3g2 38.40 6.14 52.44 0.25 2.46 99.69ORJF4g2 37.79 6.97 51.43 0.53 3.27 99.99ORJF4g2 39.71 0.29 60.41 0.06 0.17 100.64ORJF5g1 38.41 3.20 57.72 0.24 0.62 100.18ORJF5g1 37.80 7.30 53.71 0.48 0.82 100.12ORJF5g2 38.50 2.46 58.47 0.20 0.32 99.96ORJF5g2 38.45 2.44 58.36 0.19 0.51 99.95ORJF5g2 38.32 3.30 57.13 0.25 0.93 99.93ORJF5g2 38.72 2.75 58.27 0.22 0.29 100.25ORJF7g1 38.99 1.57 58.39 0.22 0.94 100.12ORJF7g1 38.86 3.21 56.69 0.43 0.99 100.18ORJF7g1 39.02 2.15 58.19 0.29 0.83 100.49ORJF7g1 38.84 2.78 57.15 0.34 1.03 100.14ORJF8g1 39.48 0.93 59.33 0.20 0.45 100.39ORJF8g1 39.51 1.16 59.20 0.27 0.45 100.59

ORJF8g1 39.42 1.01 59.11 0.18 0.64 100.36ORJF8g1 39.27 1.27 59.26 0.30 0.40 100.50ORJF9g2 38.69 1.15 58.82 0.21 0.03 98.90ORJF9g2 39.12 2.92 57.72 0.40 0.08 100.24ORJF9g2 39.59 1.02 59.51 0.15 0.03 100.30ORJF9g2 39.53 1.11 59.41 0.17 0.04 100.27ORJF10g1 39.17 3.39 57.80 0.35 0.28 100.98ORJF10g1 39.19 3.23 57.59 0.35 0.20 100.56ORJF10g1 39.20 2.69 58.50 0.28 0.20 100.87ORJF10g1 39.02 3.22 57.18 0.36 0.42 100.20ORJF10g2 39.10 3.25 57.80 0.34 0.27 100.76ORJF10g2 39.18 2.31 58.35 0.22 0.23 100.30ORJF10g2 39.14 2.60 57.86 0.27 0.16 100.03ORJF10g2 38.80 3.42 57.30 0.33 0.18 100.03ORJF10g3 38.91 2.56 58.25 0.28 0.23 100.23ORJF10g3 38.97 2.89 57.63 0.29 0.39 100.17ORJF10g3 38.97 3.73 56.86 0.36 0.38 100.30ORJF11bg1 39.49 1.70 58.72 0.23 0.44 100.58ORJF11bg1 39.42 1.87 58.49 0.25 0.37 100.39ORJF11bg1 39.37 2.15 58.00 0.39 0.50 100.41ORJF11bg1 39.57 2.01 58.35 0.34 0.41 100.68ORJF11cg1 39.35 1.76 58.28 0.32 0.28 99.99ORJF11cg1 39.34 1.97 58.52 0.32 0.55 100.70ORJF11cg1 39.46 1.87 58.94 0.24 0.34 100.85ORJF11cg1 39.35 1.97 58.80 0.29 0.52 100.93ORJF11dg1 39.25 1.60 58.58 0.22 0.63 100.28ORJF11dg1 39.24 2.45 57.67 0.43 0.50 100.28ORJF11dg1 39.16 1.88 58.48 0.23 0.49 100.24ORJF11dg1 39.21 2.11 58.09 0.34 0.53 100.28ORJF11dg2 39.31 2.08 58.07 0.23 0.41 100.10ORJF11dg2 39.36 1.30 59.04 0.18 0.38 100.26ORJF11dg2 38.95 2.20 57.88 0.23 0.46 99.72ORJF11dg2 39.43 1.80 58.22 0.29 0.57 100.31ORJF12g1 38.68 5.71 54.60 0.59 1.17 100.75ORJF12g1 39.13 3.05 57.02 0.34 0.89 100.44ORJF12g1 38.74 4.04 56.01 0.30 1.08 100.17ORJF12g1 39.08 3.12 57.61 0.32 0.64 100.77ORJF12g2 38.46 5.36 54.19 0.50 1.64 100.15ORJF12g2 38.80 3.45 57.05 0.34 0.54 100.18ORJF12g2 39.35 1.52 58.35 0.15 0.92 100.29ORJF12g2 39.20 1.98 58.53 0.22 0.16 100.09

Table S2: Sulfur Isotope Compositions of inclusions in Orapa diamonds and standard materials.

Analysis Number/Sample Name mineral δ34S (‰) δ33S (‰) ∆33S (‰)

STANDARDS:013009CDTG1S1 Troi 0.6 0.1 -0.17013009CDTG1S2 Troi -0.9 -0.4 0.01013009CDTG1S3 Troi -0.3 -0.2 -0.03013009CDTG1S4 Troi 0.8 0.5 0.14013009CDTG1S5 Troi -0.2 -0.2 -0.15020328CDTG1S1 Troi 0.0 -0.2 -0.13020328CDTG1S2 Troi 0.4 0.2 -0.06020328CDTG1S3 Troi -0.2 -0.2 -0.12020328CDTG1S4 Troi 0.5 0.3 0.00020328CDTG1S5 Troi 0.1 0.0 -0.07020328CDTG1S6 Troi -0.8 -0.4 0.03020329CDTG1S8 Troi -0.2 -0.2 -0.10020329CDTG1S9 Troi 0.2 -0.1 -0.15AVERAGE: 0.0 -0.1 -0.06

±1.0 (2σ) ±0.18 (2σ)

013009BALG2S1 Pyrite 15.0 7.5 -0.09013009BALG2S2 Pyrite 15.2 7.7 0.02013009BALG2S3 Pyrite 15.0 7.7 0.05020328BALMATG2S1 Pyrite 15.1 7.8 -0.07020328BALMATG2S2 Pyrite 15.4 8.1 0.10020328BALMATG2S3 Pyrite 15.0 7.9 0.06020328BALMATG2S4 Pyrite 15.1 7.7 -0.11020329BALMATG2S5 Pyrite 15.4 7.8 -0.04020329BALMATG2S6 Pyrite 15.1 7.7 0.04AVERAGE: 15.2 7.8 -0.01

±0.3 (2σ) ±0.15 (2σ)

013009CAR123G1S1 Pyrite 0.2 0.2 0.07013009CAR123G1S2 Pyrite 0.1 0.0 -0.03013009CAR123G1S1 Pyrite 1.6 0.8 -0.02013009CAR123G3S1 Pyrite 1.0 0.5 0.04013009CAR123G1S2 Pyrite 2.4 1.2 -0.06013009CAR123G1S3 Pyrite 2.4 1.2 0.01013009CAR123G1S4 Pyrite 3.1 1.6 0.05013009CAR123S5 Pyrite 0.6 0.3 -0.01020328CAR123G1S1 Pyrite 0.8 0.5 0.10020328CAR123G1S2 Pyrite 1.3 0.7 0.05020328CAR123G1S3 Pyrite 1.3 0.6 -0.06020328CAR123G1S4 Pyrite 0.5 0.3 0.01020328CAR123G1S5 Pyrite 1.0 0.7 0.15020328CAR123G1S5 Pyrite 1.6 0.9 0.07020328CAR123G1S6 Pyrite 1.6 0.9 0.02020329CAR123G2S1 Pyrite 1.2 0.6 -0.01

020329CAR123G3S1 Pyrite 1.0 0.5 -0.07020329CAR123G3S2 Pyrite 1.4 0.8 0.07020329CAR123G3S3 Pyrite 1.0 0.5 -0.02020329CAR123G3S5 Pyrite 1.4 0.6 -0.12020329CAR123G3S6 Pyrite 1.4 0.7 0.00020329CAR123G3S7 Pyrite 1.6 0.9 0.08020329CAR123G3S8 Pyrite 0.7 0.3 0.01020329CAR123G3S9 Pyrite 0.9 0.4 -0.04020329CAR123G1S7 Pyrite 2.0 0.9 -0.06AVERAGE: 1.3 0.7 0.01

±1.4(2σ) ±0.12 (2σ)

01093009ANS1 Pyrrh 1.7 1.0 0.12020328ANPOG1S3 Pyrrh 1.5 0.7 -0.02020328ANPOG1S1 Pyrrh 1.4 0.7 0.03020328ANPOG1S2 Pyrrh 1.4 0.7 0.02020329ANPOG2S1 Pyrrh 1.3 0.6 -0.09020329ANPOG1S7 Pyrrh 1.5 0.7 -0.07AVERAGE: 1.4 0.7 0.00

±0.3(2σ) ±0.15(2σ)

01093009LTBS1 Pyrrh -0.6 -0.2 0.06

SULFIDE INCLUSIONS:013009ORJF1G1S1VYSMALL Pyrrh 1.1 0.6 0.08013009ORJF1G2S1 Pyrrh 1.4 0.6 -0.10013009ORJF1G2S2 Pyrrh 0.8 0.5 0.07AVERAGE: 1.1 0.6 0.02

013009ORJF2G1S1 Pyrrh 2.2 1.7 0.55013009ORJF2G2S1 Pyrrh 1.7 1.5 0.64013009ORJF2G2S2 Pyrrh 2.6 1.9 0.63020328ORJF2G2S3 Pyrrh 0.0 0.6 0.59013009ORJF2G3S1 Pyrrh 2.1 1.6 0.50013009ORJF2G3S2 Pyrrh 1.1 1.1 0.49020329ORJF2G4S1 Pyrrh 1.0 0.8 0.28020329ORJF2G4S2 Pyrrh 0.0 0.4 0.43AVERAGE: 1.3 1.2 0.51

013009ORJF3G1S1 Pyrrh 1.1 0.6 0.08020328ORJF3G1S2 Pyrrh -0.7 0.0 0.38020329ORJF3G1S3 Pyrrh 0.6 0.2 -0.09013009ORJF3G2S1 Pyrrh 2.1 1.5 0.41013009ORJF3G2S2 Pyrrh -0.1 0.2 0.27020328ORJF3G2S3 Pyrrh -0.6 0.0 0.33020328ORJF3G2S4 Pyrrh 0.4 0.5 0.26020329ORJF3G2S4 Pyrrh -0.2 0.1 0.21AVERAGE: 0.3 0.4 0.23

013009ORJF4G1S1 Pyrrh 0.7 0.5 0.15013009ORJF4G2S1 Pyrrh 0.5 0.3 0.03AVERAGE: 0.6 0.4 0.09

013009ORJF5G1S1 Pyrrh 1.3 0.6 -0.06013009ORJF5G2S1 Pyrrh 1.6 0.8 0.03020328ORJF5G2S3 Pyrrh -1.4 -0.7 0.07AVERAGE: 0.5 0.3 0.01

013009ORJF6G1S1 Pyrrh 1.6 0.8 0.05013009ORJF6G2S1 Pyrrh 2.3 1.3 0.10013009ORJF6G2S1 Pyrrh 1.2 0.7 0.09AVERAGE: 1.7 0.9 0.08

020329ORJF7G1S1 Pyrrh 1.4 1.1 0.36020329ORJF7G1S2 Pyrrh 1.4 1.0 0.26AVERAGE: 1.4 1.0 0.31

020329ORJF8G1S1 Pyrrh 1.9 0.8 -0.17020329ORJF8G1S2 Pyrrh 1.8 0.9 -0.06020329ORJF8G1S3 Pyrrh 1.9 0.9 -0.06AVERAGE: 1.8 0.8 -0.09

020329ORJF9G1S1 Pyrrh 1.6 1.4 0.59020328ORJF9AG2S1 Pyrrh 0.6 1.0 0.63020329ORJF9G2S2 Pyrrh 1.5 1.4 0.61AVERAGE: 1.2 1.2 0.61

020328ORJF10G1S1 Pyrrh 0.4 0.3 0.08020329ORJF10G1S1 Pyrrh 1.3 0.5 -0.12020329ORJF10G1S2 Pyrrh 1.3 0.6 -0.11020329ORJF10G2S2 Pyrrh 0.9 0.3 -0.11AVERAGE: 1.0 0.4 -0.06

020329ORJF11BG1S1 Pyrrh -0.7 -0.4 -0.02020329ORJF11CG1S1 Pyrrh -0.9 -0.6 -0.11020329ORJF11DG1S1 Pyrrh -0.5 -0.3 0.01020329ORJF11DG1S2 Pyrrh -0.5 -0.4 -0.17AVERAGE: -0.7 -0.4 -0.07

020329ORJF12G1S1 Pyrrh 0.0 -0.1 -0.08020329ORJF12G2S1 Pyrrh 0.3 0.0 -0.15AVERAGE: 0.2 0.0 -0.11

PERIDOTITE XENOLITH acid volatile fraction(KILBOURNE HOLE)

KIL72PC3 -0.03 -0.01 0.01KIL72PC3 -1.36 -0.63 0.07

KIL72PC11 -2.49 -1.27 0.01KIL70 -1.43 -0.71 0.03KIL72 -2.71 -1.34 0.06AVERAGE: -1.60 -0.79 0.03

Sample numbers for SIMS analyses are given in the following format YYDDMMSAMPLEG#S@where # is the grain number and @ is the spot number. Isotopic analyses of Kilbourne Hole sampleswere undertaken using techniques outlined in (32).

Table S3: Mass-dependent data used to plot Figure 2. Sample name δ33S δ34S ∆33S seawater barite 10.50 20.34 0.121N 10.36 20.01 0.151S 10.48 20.33 0.112S 10.53 20.49 0.071N 11.00 21.33 0.111S 63-7 10.31 19.88 0.17572A 1-3 10.60 20.42 0.18572A 4-3 10.94 21.23 0.11572A 8-3 11.04 21.27 0.19572A 12-3 10.97 21.26 0.12572A 16-3 11.22 21.75 0.12572D 2-2 11.35 22.00 0.13572D 7-2 11.11 21.61 0.09572D 12-2 11.73 22.69 0.15572D 18-2 11.32 21.91 0.14572D 20-2 11.65 22.56 0.14574A 2-3 11.03 21.24 0.19574A 5-3 10.96 21.17 0.16574A 7-3 10.91 21.07 0.16574C 5-4 10.88 21.11 0.11574C 11-3 10.72 20.77 0.12574C 13-4 dup 10.96 21.22 0.13574C 17-1 10.93 21.11 0.16574C 17-5 10.91 21.07 0.16574C 19-2 10.81 20.91 0.14574C 20-2 11.00 20.95 0.31574C 22-3 10.49 20.27 0.15574C 24-4 10.71 20.73 0.13574C 27-2 10.94 21.21 0.12574C 31-1 10.86 21.02 0.13574C 33-2 11.17 21.67 0.12574C 35-2 11.37 22.04 0.12575B 4-3 10.95 21.17 0.14575B 4-6 11.12 21.46 0.17575B 4-6 10.93 21.20 0.11575B 5-3 125 11.49 22.20 0.16575B 5-6 11.02 21.30 0.15575B 6-5 10.99 21.21 0.17575B 7-2 10.81 20.94 0.13575B 9-2 11.19 21.67 0.13575B 14-4 10.99 21.30 0.12366 10-2 11.66 22.66 0.11366 12-2 10.55 20.19 0.25366 14-2 11.94 23.16 0.12366 16-2 11.88 23.03 0.13366 20-2 11.63 22.49 0.16

366 22-2 11.10 21.30 0.23366 24-3 11.30 21.86 0.15366 34-2 10.32 20.01 0.11366 36-2 10.10 19.44 0.19366 38-2 9.22 17.72 0.18366 42-2 8.95 17.36 0.09366 43-2 9.16 17.78 0.09366 45-2 9.29 18.04 0.09305 7-2 10.69 20.71 0.12305 8-2 10.85 20.95 0.16305 9-2 10.91 20.97 0.21305-9-6 11.11 21.42 0.18305 10-2 dup 11.08 21.36 0.18305 10-4 11.18 21.51 0.21305 10-6 11.21 21.71 0.14305 11-2 13-16 9.55 18.38 0.17305 12-2 27.31 8.91 17.23 0.12305 13-2 15-18 8.65 16.64 0.16305 14-2 63-67 9.06 17.49 0.14577 5-4 85-87 11.14 21.59 0.13577 7-4 80 11.15 21.65 0.11577 7-6 10.93 21.06 0.18577 8-4 70-80 9.76 18.92 0.11577 9-4 8-15 9.06 17.57 0.09577 9-4 8 9.04 17.53 0.09577 9-6cc 8.55 16.41 0.18577 10-6 5-13 8.68 16.77 0.13577 11-1 8-12 8.87 17.11 0.13577 11-1dup 8.93 17.25 0.13577 11 3 96-100 8.98 17.37 0.12577-11-6 110-11 9.31 17.83 0.21577 12-1 61-69 9.39 18.02 0.20577 12-1 101-10 9.84 19.07 0.11577 12-1 101-10 9.54 18.40 0.15577 12-1 104-10 9.84 19.10 0.09577 12-3 125-12 9.91 19.21 0.10577 12-4 70-76 9.76 18.94 0.10(SMALL) 577 12-4 132-14 9.85 19.06 0.13577 12-5 50-58 9.82 19.01 0.12577 12-5 80-87 9.92 19.25 0.09577 12-5 96-105 9.76 18.95 0.09577 12-6 60-70 10.13 19.66 0.10death valley 3.99 7.76 0.00GOR 14-6 9.84 19.25 -0.03GBB-16-6 9.11 17.76 0.00SWA-9 9.46 18.41 0.02Loupe-Grey 1.81 3.27 0.13applemart 8.17 15.88 0.03Desert Rose sahara 8.48 16.48 0.02UCSD drawer Pyrite -4.17 -8.31 0.12

Balboa Park 4.30 8.36 0.01aus 6-5 8.24 15.99 0.04biogenic H2S 10699 -5.31 -10.67 0.20biogenic H2S 102899 -1.66 -3.58 0.19applemart 2 3.13 6.09 0.00loope -1 0.36 0.67 0.02loope 2 0.53 0.98 0.03SP-2 T1Z 1.22 mg 2.20 4.29 -0.01290957 cambrian sulfate 18.04 35.17 0.08bnp27 0.91 1.72 0.03bnp 28 -3.40 -6.93 0.17bnp 25-1 3.88 7.45 0.05kec 81-db1 -0.95 -1.84 0.00kec 81-12 0.16 0.30 0.00kec 81-4 2.66 4.93 0.13kec 81-3 -0.38 -0.80 0.04kec 81-21 -0.43 -0.95 0.05ASW -0.12 -0.41 0.09feed sulfate 3.75 7.21 0.04Shade Mountain .1 15.94 31.09 0.04Data for marine barite sample collected by A. Paytan and D. Campbell. One sample in plot but not included in table is from (39).