hydrothermal bacterial biomineralization: potential …...ganese-oxidizing bacteria are also locally...

HYDROTHERMAL BACTERIAL BIOMINERALIZATION: POTENTIAL MODERN-DAY ANALOGUES FOR BANDED IRON-FORMATIONS

KURT 0 . KONHAUSER" Scl~ooi o f E~rr.t/~ Sr.i~rlcr.s, U~~iilel:rit>~ of Leei/,s, Lccrl.~, UK LS2 9JT

AIISTRACT: Preca~nbrian banded iro11-formations (BIFs) are chenlical sediments of hydrothermal origin and consist of Fe-1.ic11 minerals with alternating layers of chert. Because microorganisms potentially played a role in their precipitation, the study of bacterial-mineral interactions at modern hyclrothernial environments may pro\iicle small-scale analogues to those conditions under which they accumulated. Interestingly, lnicrobial populations currently gimwing at hot springs and deep-sea vents are commonly encrusted in iron and silicate minerals. lron biollliileralizatio~~ occurs either passively through interaction between the reactive sites of the cell ant1 dissol\ied cationic iron from the hydrother~nal fluid, or actively through chemolitl~otrophic iron-oxidation by bacteria such as Gnlliorlellir genera. Amorphous silica precipitates on indi\ridual bacteria through 11ydrogen bonding bctwcen hydroxy groups in the extracellular polymers and hydroxyl groups in dissolved silica, with some colollies becoming completely cemented together within a siliceous matrix LIV to several nlicronleters thick. Iron-silicates form due to reactions between dissol\ied silica and cell-bound iron. In these predo~niila~ltly nonspecific processes, bacterial cells simply catalyze reactions that are rendered possible by Llie supersaturated conditions created by the sudden physical and chemical changes induced through venting. Diagenetic reactions, sorile of which are also catalyzed by microorganisms growing in the sediment, can further alter the inineralogy of these primary precipitates, leading to the formation of secondary magnetite and siderite. I11 this way, all of the main mineralogical components of RIFs can be associated with microbial activity.

INTRODUCTION hydrothermal source (Sinionson, 1985; Klein and Beukes, 1992;

Banded iron-formations (BIFs) are some of the nlost abundant secli~nentary deposits of the Precambrian, with ~ilajor accu- nlulations ha\ling formed during the late Arche~un (2.7-2.5 Ga) and early Protcrozoic (2.5-1.8 Ga) (Isley, 1995). These thin- bedded or laminated rocks contain an anoiilalously high content of iron, and commonly, layers of chert (James, 1954). Most BIFs have sinlilar bulk conlpositions and mine~.al assemblages, althougl~ Archean BIFs are not as thick or laterally extensive as their Proterozoic coLunterparts (Gole and Klein, 1981). Iron and silica contents are typically ill the range of 20-35 wt. % and 40-50 wt. %, respecti\7ely (Jamcs, l983), while the mineralogy consists of quarfz, magnetite, hematite, \rarious carbonates, ferrous iron-silicates, and pyrite; some carbonaceous material is also present (McConchie, 1987: Klein and Beukes, 1992). The banding in BlFs generally coiilprises alternating layers of Fe- rich ~llinerals with layers of chert (and carbonate) that can he obser\ied at three different scales: macro, meso, and micro (Tre~ldall. 1965; Trendall ancl Blockley, 1970). Macroba~lds arc coarse alterations of contrasting Fe-oxide facies (magnetite, hematite) and shalc facies (referred to as S-macrobands, with chert, siderite, and iron-silicates) decimeters to tens of tileters in tliickness; mesobands are centinleter thick units; and ~nicrobands usually range from submillimeter lo 5 mm (Morris. 1993). The lateral continuity of mesobatlds is a characteristic feature of solile iron-formations, and in the Dales Gorge Member, Hamessley Group BIF, Western Australia, for example, they can be traced for niore than 130 knl tl~rough the basin (Ewers and R/Iorris, 198 1).

It is generally accepted that BIFs are cl~emical precipitates (Morris, 1993), with the iron primarily derived from a

Isley, 1995; Sirilonson and Hassler, 1996). This nlodel is supported by (i) estimates of higher surface heat flow and rapid hydrothermal cycling rates for the late Archean-early Proterozoic (T~~rcotte, 1980; J:lcobsen and Pimentel-Klose, 1988), (ii) rare earth elenlent (REE) distribution patterns in BIFs (Jacobscn and Pimentel-Klose, 1988; Derry and Jacobsen, 1990; Bau and Mdler, 1993), (iii) sulfur and neodymium isotopic con~positions of Archean and early Proterozoic iron-formations (Cameron, 1983; Jacobsen and Pimentel-Klose, 1988), and (iv) ~ r ~ ~ / ~ r ~ ~ ratios and 180 depletion in Archean and early Proterozoic carbonates (Veizer et al., 1982; Beulces et al., 1990). Interestingly, the transitio~l from hydrothermal-dominated ocean chemistry of the Archean to continental-do~llinated ~iloclern seawater was established by approximately 0.7-0.8 Ga, coi~~ciding with the last ~iiajor BIF deposits of the Rapitan Group, N.W.T., Canada (Klein and Beukes, 1992).

Although the origin of iron in BIFs has been clarified, n-tech- anisms for their depositioil are still under debate (see McConchie, 1987 and Morris, 1993, for detailed discussions). Interactions between ~ilicroorganistlls and dissolved ferrous iron and silica in the Precambrian ocean have been suggested as one plausible means for ~iliileral precipilation (Cloud, 1965, 1973; LaBerge, 1973; Baur et al., 1985; Birnbauni and Wireman, 1985; Holm, 1987, 1989; Nealson and Myers, 1990; Widdel et al., 1993; Ehrenreich and Widdel, 1994; Konhauser and Ferris, 1996). Cloud (1965, 1973) origi~lally proposed that primiti\re oxygen-releasitlg pllotosynthetic bacteria, arl~ich may have lacked suitably ad\ianced oxygen-mediating enzymes, attached any free O2 that was generated to co~lvenient oxygen acceptors in the inlnlediate vicinity of the bacterial population. The abundant ferro~rs imn in solution at that time woulcl have served as such an acceptor, thereby ~i~aintaining the reducing em~ironment necessary for their sur\rival. T h ~ s process of transferring oxygen

Copyright © 2012, Society for Sedimentary Geology (SEPM)

to nonbiological oxygen acceptors may have later evolved into a source of energy for cl~emolithotropl~ic microaerophilic bacteria such as Gnllio1~ellcr ,Ji.r-rllgi~lecr. Holm (1987, 1989) speculated that slow biological oxidation of hydrothermal solutio~ls by these microorganisms, in an ocean with limited free photosynthetic oxygen, was primarily responsible for the precipitatio~~ of ferrous to ferric iron over wide ocean basins. Alternatively, the cliscovery of purple bacteria that couple ferrous iron oxidatio~l to anoxygenic photosynthesis has been suggested as a Inealls for early BIF cleposition in the Archean, before free oxygen became available as an oxidant (Wicldel et al., 1993; Ehrenreich and Widdel, 1994).

Simultaneously, saturation of ocean waters with respect to silica due to the absence of silica-secreting microorganisms (e.g., diatoms, radiolarians) at that time provided ideal conditions for the precipitation of amorpl~ous silica (Sieves, 1992). The sudde~l cooling and pressure drop of hydrothennal effluent after venting on the ocean floor (Williams and Crerar, 1985), evaporation on contine~ltal shelves (Morris, 1993), tidal flats and restricted lagooils (Knoll, 1985), and coprecipitation with iron (Ewers, 1983) could each have lecl to silica supersaturation and precipitation. During diagenesis, these primary silica precipitates subsequently may have undergone a dissolutioit-reprecipitation pathway with a characteristic sequence of ~ni~leralogical transforinations progressing from amorphous silica to low-temperature cristobalite to quartz, the phase of lowest entropy and slowest nucleation and growth rates (Williams and Crerar, 1985). A biological role for amorphous silica precipitation has also been considered, since bacteria are known to create favorable geochemical conditions Tor silicification through their metabolic activities (Birnbaum and Wireman, 1985) and they provide highly reactive surfaces for silica ilucleation and mineral growth (Urrutia and Beveridge, 1993). The presence of Precambrian microfossils in carbo~laceous chests (Awrainik et al., 1983; Walsh and Lowe, 1985; Schopf and Packer, 1987; Schopf, 1993) and BIFs (Barghoorn and Tyler, 1965; Cloud, 1965; LaBerge, 1973), as well as the presence of silicified stromatolites (Lowe, 1980; Walter et al., 1980) would certainly seem to suggest that microorganisms were, in some maIiner, inextricably linked to silica precipitation.

Because Precambrian banded iron-forinations are chemical sediments of hydrother~nal origin, and inicroorganis~ns potentially played a role in their precipitation, I propose here that the st~rdy of bacterial-mineral i~lteractions at modern hydrothermal ein~ironments may provide small-scale analogues to those conditions under which they accumulated. This paper will show (i) that bacteria directly forin iron and silicate minerals in nod ern hydrothennal environments and (ii) that these biomineralization processes correlate well with conditions that nlay have existed in the Precambrian, resulting in banded iron-formation.

HYDROTHERMAL BIOMINERALIZATION

Bacteria are fomncl today tl~roughout the hyclrospl~cre, coin- monly inhabiting environments that cannot sustain other life forms. At hot springs, photosyntl~etic cyanobacteria (e.g.,

Sy~rechococc~rs sp.) and filamentous bacteria (e.g., Clzlo/.qflc.srrs sp.) frequently fonn well-defined laminated rnats adjacent to hyclrotherinal vents and in their o~rtflow channels (Walter et al., 1972; Walter, 1976; Doeinel and Brock, 1977). In Yellowstone National Park, USA, for example, these mats can reach several millimeters in thickness with both S~~~~echococc l r s and C h l o r ~ j l e x ~ ~ s cells growing to depths where sufficient light exists, below which anaerobic clecoinposition by heterotrophic bacteria takes place (Doemel and Broclc, 1977; Walter et al., 1992). 111 some high-sulfide springs, cyanobacterial growth is suppressed, and pure Clzlor.oJle,~lls populations are developed via anoxygenic photosyntl~esis (Castenholz, 1973). In groundwater from geotherii~al areas, such as those in Iceland, high ferrous iron concentrations encourage the growth of iron-oxidizing bacteria such as Gallio~lellcrfer.~.~rgirlerr (Holm, 1987).

In deep-sea hydrothermal environments, bacterial producti\~- itp is highest in the vicinity of warm vents, where reducing fluids mix with clownwelling cold, oxygenated seawater (Janilasch and Mottl, 1985). Bacterial commiu~~ities growing at these sites have concentrations that range from lo6 to lo9 cells per milliliter (Jannascli and Wirsen, 1981). Reduced sulfitr coinpounds appear to represent the major electron donors for aerobic microbial metabolism, although methane-, hydrogen-, iron-, and manganese-oxidizing bacteria are also locally important (Jannasch and Mottl, 1985). Methanogenic, acetogenic, and sulfur-recluc- ing bacteria co~nlno~lly occupy anoxic niches where C 0 2 and H2 are present (Baross et al., 1982; Harmsen et al., 1997).

These mixed bacterial communities interact diversely with their snrrounding aqueous environment. Passive interactions occur when ~nicroorganisms behave as a solid-phase sorbents for dissolved ions. This behavior steins from bacterial cells having high surface area to volume ratios (Beveridge, 1988) and exceptionally reactive surfaces comprising exposed anionic f~tnctional groups (Beveridge and Fyfe, 1985). Once bound to the cells, these metals reduce the activation energy barriers to heterogeneous ilucleation by providing sites where additional ions can be sorbed and autliigenic ~ninerals are formed (Ferris, 1997; Fortin et al., 1997; Konhauser, 1997, 1998). Conversely, some bacteria actively promote mineral formation by proclucing inetabolic end products (e.g., OH-, C02, H+, NH3, and H2S) that react with dissolved metal ions, leading to supersaturation (Thompson and Ferris, 1990; Fortin and Beveridge, 1997), 01

they produce enzyme-mediated changes in redox state (e.g., oxidation of ~ e ~ + to ~ e " ) (Ghiorse and Ehrlich, 1992). In this regard bacteria catalyze reactions that are kinetically slow, 01

even thermodynamically unfeasible prior to their intervention. and hence induce mineral growth uilcler potentially unfa\~orabk conditions (Ehrlich, 1996). The biogenically formed ininerals have crystal habits and chemical compositions siinilar to thost prod~rced by precipitation from inorganic solutions since the) are governed by the same equilibriuin principles that control mineralization of their inorganic counterparts.

In terrestrial hot springs, the presence of iron hydroxide, ir

134 KURT O. KONHAUSER

to nonbiological oxygen acceptors may have later evolved into asource of energy for chemolithotrophic microaerophilic bacteriasuch as Gallionella ferruginea. Holm (1987, 1989) speculatedthat slow biological oxidation of hydrothermal solutions bythese microorganisms, in an ocean with limited free photosynthetic oxygen, was primarily responsible for the precipitation offerrous to ferric iron over wide ocean basins. Alternatively, thediscovery of purple bacteria that couple ferrous iron oxidation toanoxygenic photosynthesis has been suggested as a means forearly BIF deposition in the Archean, before free oxygen becameavailable as an oxidant (Widdel et aI., 1993; Ehrenreich andWiddel, 1994).

Simultaneously, saturation of ocean waters with respect tosilica due to the absence of silica-secreting microorganisms(e.g., diatoms, radiolarians) at that time provided ideal conditions for the precipitation of amorphous silica (Siever, 1992).The sudden cooling and pressure drop of hydrothermal effluentafter venting on the ocean f100r (Williams and Crerar, 1985),evaporation on continental shelves (Morris, 1993), tidal flatsand restricted lagoons (Knoll, 1985), and coprecipitation withiron (Ewers, 1983) could each have led to silica supersaturationand precipitation. During diagenesis, these primary silica precipitates subsequently may have undergone a dissolution-reprecipitation pathway with a characteristic sequence ofmineralogical transformations progressing from amorphous silica to low-temperature cristobalite to quartz, the phase of lowestentropy and slowest nucleation and growth rates (Williams andCrerar, 1985). A biological role for amorphous silica precipitation has also been considered, since bacteria are known to create favorable geochemical conditions for silicification throughtheir metabolic activities (Birnbaum and Wireman, 1985) andthey provide highly reactive surfaces for silica nucleation andmineral growth (Urrutia and Beveridge, 1993). The presence ofPrecambrian microfossils in carbonaceous cherts (Awramik etaI., 1983; Walsh and Lowe, 1985; Schopf and Packer, 1987;Schopf, 1993) and BIFs (Barghoorn and Tyler, 1965; Cloud,1965; LaBerge, 1973), as well as the presence of silicified stromatolites (Lowe, 1980; Walter et aI., 1980) would certainlyseem to suggest that microorganisms were, in some manner,inextricably linked to silica precipitation.

Because Precambrian banded iron-formations are chemicalsediments of hydrothermal origin, and microorganisms potentially played a role in their precipitation, I propose here that thestudy of bacterial-mineral interactions at modern hydrothermalenvironments may provide small-scale analogues to those conditions under which they accumulated. This paper will show (i)that bacteria directly form iron and silicate minerals in modernhydrothermal environments and (ii) that these biomineralizationprocesses correlate well with conditions that may have existedin the Precambrian, resulting in banded iron-formation.

HYDROTHERMAL BIOMINERALIZATION

Bacteria are found today throughout the hydrosphere, commonly inhabiting environments that cannot sustain other lifeforms. At hot springs, photosynthetic cyanobacteria (e.g.,

SynechococcliS sp.) and filamentous bacteria (e.g., Chloro/lexlissp.) frequently form well-defined laminated mats adjacent tohydrothermal vents and in their outflow channels (Walter et aI.,1972; Walter, 1976; Doemel and Brock, 1977). In YellowstoneNational Park, USA, for example, these mats can reach severalmillimeters in thickness with both Synechococclis andChloro.!lexlIs cells growing to depths where sufficient lightexists, below which anaerobic decomposition by heterotrophicbacteria takes place (Doemel and Brock, 1977; Walter et aI.,1992). In some high-sulfide springs, cyanobacterial growth issuppressed, and pure Chloro.!lexlIs populations are developedvia anoxygenic photosynthesis (Castenholz, 1973). In groundwater from geothermal areas, such as those in Iceland, high ferrous iron concentrations encourage the growth of iron-oxidizingbacteria such as Gnllionella ferrllginea (Holm, 1987).

In deep-sea hydrothermal environments, bacterial productivity is highest in the vicinity of warm vents, where reducing fluids mix with downwelling cold, oxygenated seawater (Jannaschand Mottl, 1985). Bacterial communities growing at these siteshave concentrations that range from 106 to 109 cells per milliliter (Jannasch and Wirsen, 1981). Reduced sulfur compoundsappear to represent the major electron donors for aerobic microbial metabolism, although methane-, hydrogen-, iron-, and manganese-oxidizing bacteria are also locally important (Jannaschand Mottl, 1985). Methanogenic, acetogenic, and sulfur-reducing bacteria commonly occupy anoxic niches where CO2 and H2

are present (Bm'oss et aI., 1982; Harmsen et aI., 1997).These mixed bacterial communities interact diversely with

their surrounding aqueous environment. Passive interactionsoccur when microorganisms behave as a solid-phase sorbentsfor dissolved ions. This behavior stems from bacterial cells having high surface area to volume ratios (Beveridge, 1988) andexceptionally reactive surfaces comprising exposed anionicfunctional groups (Beveridge and Fyfe, 1985). Once bound tothe cells, these metals reduce the activation energy barriers toheterogeneous nucleation by providing sites where additionalions can be sorbed and authigenic minerals are formed (Ferris,1997; Fortin et aI., 1997; Konhauser, 1997, 1998). Conversely,some bacteria actively promote mineral formation by producingmetabolic end products (e.g., OR, CO2, H+, NH3, and H2S) thatreact with dissolved metal ions, leading to supersaturation(Thompson and Ferris, 1990; Fortin and Beveridge, 1997), 01

they produce enzyme-mediated changes in redox state (e.g., oxidation of Fe2+ to Fe3+) (Ghiorse and Ehrlich, 1992). In thi,regard bacteria catalyze reactions that are kinetically slow, oreven thermodynamically unfeasible prior to their intervention.and hence induce mineral growth under potentially unfavorableconditions (Ehrlich, 1996). The biogenically formed mineral,have crystal habits and chemical compositions similar to thoseproduced by precipitation from inorganic solutions since theyare governed by the same equilibrium principles that controlmineralization of their inorganic counterparts.

fron Hydroxides

In terrestrial hot springs, the presence of iron hydroxide, ir

BlI-r: HYDROTIfERAfAL BACTERIAL BIOI~.II \"ERAL~~~TION 135

FIG. 1 .-Tratisliiissiol~ electron micrograph (TEM) of an epiiitliic bacterial cell fl-om a liot spring effluent channel at Lfsulibll, Iceland. TLlc bacterial ccll is surrounded by a dctlse, iron-rich capsule consisting of amorphous ferric liydrox- itle. Scalc bar = 140 nlii.

association with ~llicrobial mats, has been reported from geothermal areas in Iceland (Holm, 1987; Konhauser and Ferris, 1996) and Yellowstone National Park (Ferris et al., 1986). At Hlisatbttir, Iceland, samples of Fe(l1I) precipitate were taken from the bottom of a drilled well (80 111 in depth) used to pmnp low-temperature ( 1 2°C) geothermal water (Holm, 1987). Microscopic analysis revealed that the iron precipitate coilsisted inailily of livi~lg cells, stalks, and other remains of Grillionelln fel-t.lrgi11ea. On 1 ~ i l g of dry precipitate dispersed in 1 1111 of water, tlie total living cell count of Gcrllionellrr fe~.r.irgil,err was 3- x 10' cells (Holm, 1987). Transmission electron ~liicroscopp (TEM) and energy dispersi\~e X-ray spectroscopy (EDS) of ~uicrobial inats growing in liot spritlg outflow channels of Ljisul~bll, Iceland, inclicated a highly mineralized bactesial corn- munity (Konhauser and Ferris, 1996). Indi\~idual bacterial cells (Fig. 1) or microcolonies were frequently observed surrounded by an iron-rich capsule (a hydrated polysaccharide material that is naturally allionic and has great potential for scaveilgi~lg metals). I11 L$suh611, precipitatioll of iron phases was not restricted to cells with erlcapsulatillg material; the partial and complete encn~station of solne bacterial cells by fine-grained (1100 nm in diameter) spheroids composed of an~orphous ferric hydroxide were also observed. Other cells underwent intracelli~lar miner- alizatioll so that the cytoplas~ll was co~npletely replaced with amorpl~ous iroil hydroxide once the cell had lysed.

In marine 1iydrothe1.mal en\~iron~nents, Fe-hydroxides are commonly precipitated in the region i~ll~llediatelp surrounding the vent, tlie peripheral sediments, and the hydrothermal plume (Karl et al., 1988; Juniper ;tnd Tebo, 1995). Baross and Denzing (1985) provided evide~lce that at an active black smoker from the East Pacific Rise, iron was deposited in significantly greater amounts 011 rock surfaces co\~erecl wiih microbial mats than on uncolonized rocks. Electron l~iicroscopic examination of vestimentiferan tubes at two chimneys on the southern Juari dc Fuca Ridge (Fig. 2) revealed diverse unicellular and filamentous bac-

FIG. 2.-Eleciron microgniphs of vcstilnentifer;~~~ tul~es fro111 a site of Fe- oxidc encrustation on tlie southerti Juan dc Fuca Ridgc. (A) Scanning electrol~ micropral~h (SEM) of the tube surface ~vitli light colotiization by slicatlied baclc- ria. (B) TEM phoio sliowitig Fc pa~.ticle (11) accum~~lation on slieatli of bacler-ium (B) attached to tlie rube surface (T). (C) Extensive Fe particle accumulation in ~~nolganized cxtrnccllular slime slrn-ounding tlie bacterium. Reprinted \vith permission of tlie author ;~ntl CRC Press (Juniper alltl Tebo. 1995).

teria colonizing the tube silrfaces where they nccumulated iron hydroside (as Fe-rich spheres 6-8 pm in diameter) around their sheaths (Tunnicliffe and Fontaine, 1987). These spheres were slio\vn to eventually coalesce to form a coatinuous crust in both vertical and horizontal climensiotls. The sequeilce of formation suggests that iron was initially adsorbed to reactive organic sites 011 the cells, but that further iron accumulation may have resulted aittocatalytically (Tunnicliffe and Fontaine, 1987). Extensive deposits of biogenic ferric hydroxide have also been described from the surface sedilliellts of the ancient caldera of Santorini. Aegean Sea. Hanert (1973) and Holm (1987) both found Fe-encrusted stalks of Galliorzellrr~er.r.rrgi11ect o c c ~ u ~ i n g in sucl~ masses that the activity and importance of this bacter i~~~l l in iron sedimentation was unquestionable. I11 deep liydrotliennal areas. iron hydroxides were found to for111 on bacterial filaments growing i11 mud deposits at the Larson and Red Seamounts, near 21°N, East Pacific Rise (Alt, 1986, 1988) and 011 the Loilii Seamomnt, Hawaii (Karl et al.. 1988, 1989). Because the ~ n u d consisted essentially of bactesial filaments, and the hydrother- nlal waters in which these deposits formed were slightly acidic (pH 5 to 6) and low in dissolved 0 2 (spontaneous oxidation of ferro~ts to ferric iron wlould be very slo\v under these conditions), biologically catalyzed precipitation was clearly favored over abiotic oxidation processes (Juniper and Tebo, 1995).

The ~~licrobial precipitation of ferric hydroxide (e.g., ferrihydrite) call occur both passively and acti\iely. In the first instance, the oxidatioii and hydrolysis of cell-bound ferrous iron or the bindiag of ferric species [e.g., F~(oH)~' , Fe(OH)?'] ancl

BIFs: HYDROTHERMAL BACTERIAL BIOMINERALIZATION 135

FIG. I.-Transmission electron micrograph (TEM) of an epilithic bacterial

cell from a hot spring effluent channel at L)'suh61l, Iceland. Thc bacterial cell issurrounded by a dcnse, iron-rich capsule consisting of amorphous ferric hydroxide. Scale bar = 140 nm.

association with microbial mats, has been reported from geothennal areas in Iceland (Holm, 1987; Konhauser and Ferris,1996) and Yellowstone National Park (Ferris et a!., 1986). AtHusat6ttir, Iceland, samples of Fe(III) precipitate were takenfrom the bottom of a drilled well (80 m in depth) used to pumplow-temperature (12°C) geothermal water (Holm, 1987).Microscopic analysis revealed that the iron precipitate consistedmainly of living cells, stalks, and other remains of Gallionellaferruginea. On I mg of dry precipitate dispersed in I ml ofwater, the total living cell count of Gallionellaferruginea was 2x 108 cells (Holm, 1987). Transmission electron microscopy(TEM) and energy dispersive X-ray spectroscopy (EDS) ofmicrobial mats growing in hot spring outflow channels ofLysuh6ll, Iceland, indicated a highly mineralized bacterial community (Konhauser and Ferris, 1996). Individual bacterial cells(Fig. 1) or microcolonies were frequently observed surroundedby an iron-rich capsule (a hydrated polysaccharide material thatis naturally anionic and has great potential for scavenging metals). In Lysuh6ll, precipitation of iron phases was not restrictedto cells with encapsulating material; the partial and completeencrustation of some bacterial cells by fine-grained « 100 nm indiameter) spheroids composed of amorphous ferric hydroxidewere also observed. Other cells underwent intracellular mineralization so that the cytoplasm was completely replaced withamorphous iron hydroxide once the cell had lysed.

In marine hydrothermal environments, Fe-hydroxides arecommonly precipitated in the region immediately surroundingthe vent, the peripheral sediments, and the hydrothermal plume(Karl et a!., 1988; Juniper and Tebo, 1995). Baross and Deming(1985) provided evidence that at an active black smoker fromthe East Pacific Rise, iron was deposited in significantly greateramounts on rock surfaces covered with microbial mats than onuncolonized rocks. Electron microscopic examination of vestimentiferan tubes at two chimneys on the southern Juan de FucaRidge (Fig. 2) revealed diverse unicellular and filamentous bac-

FIG. 2.-Electron micrographs of vestimentiferan tubes from a site of Feoxide encrustation on the southern Juan de Fuca Ridge. CA) Scanning electronmicrograph (SEM) of the tube surface with light colonization by sheathed bacte

ria. (B) TEM photo showing Fe particle (p) accumulation on sheath of bacterium(B) attached to the tube surface (T). (C) Extensive Fe particle accumulation inunorganized extracellular slime surrounding the bacterium. Reprinted with permission of the author and CRC Press (Juniper and Tebo. 1995).

teria colonizing the tube surfaces where they accumulated ironhydroxide (as Fe-rich spheres 6-8 ~l1n in diameter) around theirsheaths (Tunnicliffe and Fontaine, 1987). These spheres wereshown to eventually coalesce to form a continuous crust in bothvertical and horizontal dimensions. The sequence of formationsuggests that iron was initially adsorbed to reactive organic siteson the cells, but that further iron accumulation may haveresulted autocatalytically (Tunnicliffe and Fontaine, 1987).Extensive deposits of biogenic ferric hydroxide have also beendescribed from the surface sediments of the ancient caldera ofSantorini, Aegean Sea. Hanert (1973) and Holm (1987) bothfound Fe-encrusted stalks of Gallionella ferruginea occurring insuch masses that the activity and importance of this bacteriumin iron sedimentation was unquestionable. In deep hydrothermalareas, iron hydroxides were found to form on bacterial filamentsgrowing in mud deposits at the Larson and Red Seamounts, near21°N, East Pacific Rise (Alt, 1986, 1988) and 011 the LoihiSeamount, Hawaii (Karl et a!., 1988, 1989). Because the mudconsisted essentially of bacterial filaments, and the hydrothermal waters in which these deposits formed were slightly acidic(pH 5 to 6) and low in dissolved O2 (spontaneous oxidation offerrous to ferric iron would be very slow under these conditions), biologically catalyzed precipitation was clearly favoredover abiotic oxidation processes (Juniper and Tebo, 1995).

The microbial precipitation of ferric hydroxide (e.g., ferrihydrite) can occur both passively and actively. In the firstinstance, the oxidation and hydrolysis of cell-bound ferrous ironor the binding of ferric species [e.g., Fe(OH)2+, Fe(OH)tJ and

136 KURT 0. KO/VH/l USER

cationic colloidal species [e.g., F ~ ( H ~ O ) ~ ( O H ) ~ ( O H ) ~ ' ] to negatively charged polyiners (MacRae and Edwards, 1972; Ferris el al., 1989) can result in ferric iron deposition. Alternatively, ferrous iron transported into an oxygenated environinent spontaneously reacts with dissolved oxygen (at circurnneutral pH) to precipitate rapidly as ferric hydroxide on available nucleation sites. Bacteria passively act as such sites, and over a short period of time the lnicrobial mats can become coinpletely encrusted in amorphous iron as abiological surface catalysis accelerates the rate of mineral precipitation (Ghiorse, 1984). It has even been suggested that under circumneutral pH any bactcriuin that produces acidic, extracellular polytners will nonspecifically adsorb positively charged Fe-hydroxides (Ghiorse, 1984). This is due to the point of zero charge (pH where the illinera1 has zero charge) of amorphous Fe-hydroxides (e.g., ferrihydrite) being in the range of 5.3 to 7.5 for natural samples (Schwertmann and Fechter, 1982), which thereby allows reactive organic sites to scavenge ferric iron from the surrounding waters. The growth of iron-depositing inicrobial populations in en\rironinents where they are constantly covered in amorphous iron hydroxide suggests that these sites must also provide the bacteria with favorable growth conditions. Quite possibly the continual supply of iron, as well as the trace metals and anions that adsorb or coprecipitate directly onto iron hydroxides (German et al., 1991), may serve as an ideal nutrient source in close proximity to the cells.

The active process by which iron hydroxides fo rn stems from the tnetabolism of Fe(I1)-oxidizing bacteria that derive energy from the biological oxidation of Fe(I1) to Fe(I1I) for the fixation of carbon froin carbon dioxide dissolved in water (Ehrlich, 1990). At neutral pH, Fe(I1) oxidation by Galliorlella genera occurs under partially reduced conditions of 0.1-1 mg of O2 per liter (Ehrlich, 1990). The existence of these microaerophilic bacteria relies on their oxidizing efficiency relative to abiotic oxidation at low-oxygen fugacity. Because very little energy is generated in the oxidation of ferrous to ferric iron, these microaerophilic bacteria must oxidize large quantities of iron to grow (Lees et al., 1969). Consequently, even a small number of bacteria can be responsible for precipitating vast amounts of iron (Brock et al., 1984). Ferrous iron has also been obser\~ed to ~tnclergo ~nicrobial oxidation under anoxygenic conditions. Wicldel et al. (1993) and Ehrenreich and Widdel (1994) have described purple, nonsulfi~r bacteria that combined ferrous iron oxidation with C 0 2 fixation for cellular material, with light as the energy source. Ferrous iron oxidation took place on the outer nle~nbrane so that the insolitble ferric hydroxide remained outside of the cell (Ehrenreich and Widdel, 1994). Straub et al. (1996) also demonstrated that the biological oxidation of ferrous iron, in the absence of oxygen, was possible by light-independent, chemotrophic activity with nitrate as the electron acceptor.

Once the primary iron hydroxides are precipitated and incorporated into the sediment, several diagenetic reactions can subsequently alter their mineralogies, textures, and structures. The trailsforlnation of ferric hydroxide into hematite takes place through internal aggregation and rearrangement (Cornell et al., 1987), and in lnarine sediments, its formation is favored under

high temperatures and low organic matter content (Scllwertmann and Fitzpatrick, 1992). However, the observed presence of a mixed ferric hydroxide/hematite asselnblage on bacterial cells, in present-day biofilms, suggests that this process may also occur rapidly at surface conditions (Ferris et al., 1989; Sawicki et al., 1995). Dissiinilatory Fe(II1)-reducing bacteria have been shown to produce ~nagnetite under allaerobic conditions, as a by-product during the oxidation of organic compounds coupled with the reduction of poorly crystalline, Fe(II1) oxide (Lovley, 1991). For example, in hydrothern~al sediments located 25 kin from the East Pacific Rise, authigenic nlagnetite was recovered at the redoxcline between Fe-oxidizing to Fe- reducing conditions (Karlin et al., 1987). The fortnation of siderite in marine sediinents has also been shown to result from the ability of bacteria (e.g., Deslllfovib~io sp.) to release ferrous iron, via bacterial Fe(II1)-reduction, into waters with excess seawater bicarbonate (Coleman et al., 1993). In deep sediinent from Santorini, beneath the iron hydroxide layers, siderite was identified as a by-product of inicrobial diagenesis (Holm, 1987).

Alilo~l~llo~is Silica

The close association of ~nicroorganisms with amorphous silica has been reported from a variety of hot springs, including Yellowstone National Park (Walter et al., 1972; Walter, 1976; Ferris et al., 1986; Hiinnan and Lindstrom, 1996); Taupo Volcanic Zone, New Zealand (Jones et al., 1997a, b); Kenya Rift (Jones and Renaut, 1996); El Tatio, Chile (Jones and Renaut, 1997); and Geysir, Iceland (Schultze-Lam et al., 1995; Konhauser and Ferris, 1996). The role of rnicroorganisins in silica precipitation at hot springs generally has been considered a passive process (Walter et a]., 1972), in that rapid cooling at ambient air temperatmes, e\iaporation and/or steain loss, and a change in pH of the hydrothermal waters following discharge can all induce silicification of growing cells (Fournier, 1985). However, several recent studies using electron ~nicroscopy have shown that microorganisms provide favorable nucleation sites that serve as templates for silicification. In sediments from Yellowstone, Ferris et al. (1986) observed individual bacterial cells preserved in successive stages of amorphous silica encrustation; sonle cells were thoroughly e~nbedded in a lnineralized ~natrix. In New Zealand, bacterial filainents were sinlilarly encritsted by amorphous silica (Jones et al., 1997a, b). That ~nicroorganisins catalyze silica precipitation was confirllled through analyses of inicrobial mats growing around a hot spring vent in Geysir, Iceland, where filamentous phototrophic bacteria resembling Chlor.oflex~ls a ~ ~ ~ n n t i n c ~ l s , and the ~~nicellular cyanobacteriuln S ~ ~ r ~ e c h o c o c c ~ ~ s sp., underwent rapid silicification (Schultze-Lain et al., 1995; Konhauser and Ferris, 1996). The mineralization associated with the cells occurred as spheroidal grains (200-500 nin in diameter) both extracellularly on the sheaths of living bacteria and intracellularly within the cytoplasm presumably after the cells had lyzed. In many filaments, the silica crystallites appeared to have coalesced such that individual precipitates were no longer distinguishable (Fig. 3); entire colonies sonletiines became cenlented together in a


cationic colloidal species [e.g., Fe(H20)s(OH)s(OH)2+] to negatively charged polymers (MacRae and Edwards, 1972; Ferris etaI., 1989) can result in ferric iron deposition. Alternatively, ferrous iron transported into an oxygenated environment spontaneously reacts with dissolved oxygen (at circumneutral pH) toprecipitate rapidly as ferric hydroxide on available nucleationsites. Bacteria passively act as such sites, and over a short periodof time the microbial mats can become completely encrusted inamorphous iron as abiological surface catalysis accelerates therate of mineral precipitation (Ghiorse, 1984). It has even beensuggested that under circumneutral pH any bacterium that produces acidic, extracellular polymers will nonspecifically adsorbpositively charged Fe-hydroxides (Ghiorse, 1984). This is due tothe point of zero charge (pH where the mineral has zero charge)of amorphous Fe-hydroxides (e.g., ferrihydrite) being in therange of 5.3 to 7.5 for natural samples (Schwertmann andFechter, 1982), which thereby allows reactive organic sites toscavenge ferric iron from the surrounding waters. The growth ofiron-depositing microbial populations in environments wherethey are constantly covered in amorphous iron hydroxide suggests that these sites must also provide the bacteria with favorable growth conditions. Quite possibly the continual supply ofiron, as well as the trace metals and anions that adsorb or coprecipitate directly onto iron hydroxides (German et aI., 1991), mayserve as an ideal nutrient source in close proximity to the cells.

The active process by which iron hydroxides form stemsfrom the metabolism of Fe(II)-oxidizing bacteria that deriveenergy from the biological oxidation of Fe(II) to Fe(III) for thefixation of carbon from carbon dioxide dissolved in water(Ehrlich, 1990). At neutral pH, Fe(II) oxidation by Gallionellagenera occurs under partially reduced conditions of 0.1-1 mg ofO2 per liter (Ehrlich, 1990). The existence of thesemicroaerophilic bacteria relies on their oxidizing efficiency relative to abiotic oxidation at low-oxygen fugacity. Because verylittle energy is generated in the oxidation of ferrous to ferriciron, these microaerophilic bacteria must oxidize large quantities of iron to grow (Lees et aI., 1969). Consequently, even asmall number of bacteria can be responsible for precipitatingvast amounts of iron (Brock et aI., 1984). Ferrous iron has alsobeen observed to undergo microbial oxidation under anoxygenicconditions. Widdel et a1. (1993) and Ehrenreich and Widdel(1994) have described purple, nonsulfur bacteria that combinedferrous iron oxidation with CO2 fixation for cellular material,with light as the energy source. Ferrous iron oxidation tookplace on the outer membrane so that the insoluble ferric hydroxide remained outside of the cell (Ehrenreich and Widdel, 1994).Straub et a1. (1996) also demonstrated that the biological oxidation of ferrous iron, in the absence of oxygen, was possible bylight-independent, chemotrophic activity with nitrate as theelectron acceptor.

Once the primary iron hydroxides are precipitated and incorporated into the sediment, several diagenetic reactions can subsequently alter their mineralogies, textures, and structures. Thetransformation of ferric hydroxide into hematite takes placethrough internal aggregation and rearrangement (Cornell et aI.,1987), and in marine sediments, its formation is favored under

high temperatures and low organic matter content(Schwertmann and Fitzpatrick, 1992). However, the observedpresence of a mixed ferric hydroxide/hematite assemblage onbacterial cells, in present-day biofilms, suggests that thisprocess may also occur rapidly at surface conditions (Ferris etaI., 1989; Sawicki et aI., 1995). Dissimilatory Fe(III)-reducingbacteria have been shown to produce magnetite under anaerobicconditions, as a by-product during the oxidation of organic compounds coupled with the reduction of poorly crystalline, Fe(III)oxide (Lovley, 1991). For example, in hydrothermal sedimentslocated 25 km from the East Pacific Rise, authigenic magnetitewas recovered at the redoxcline between Fe-oxidizing to Fereducing conditions (Karlin et aI., 1987). The formation ofsiderite in marine sediments has also been shown to result fromthe ability of bacteria (e.g., Deslilfovibrio sp.) to release ferrousiron, via bacterial Fe(III)-reduction, into waters with excess seawater bicarbonate (Coleman et aI., 1993). In deep sediment fromSantorini, beneath the iron hydroxide layers, siderite was identified as a by-product of microbial diagenesis (Holm, 1987).

AIllOlpholiS Silica

The close association of microorganisms with amorphoussilica has been reported from a variety of hot springs, includingYellowstone National Park (Walter et aI., 1972; Walter, 1976;Ferris et aI., 1986; Hinman and Lindstrom, 1996); TaupoVolcanic Zone, New Zealand (Jones et aI., 1997a, b); Kenya Rift(Jones and Renaut, 1996); El Tatio, Chile (Jones and Renaut,1997); and Geysir, Iceland (Schultze-Lam et aI., 1995;Konhauser and Ferris, 1996). The role of microorganisms in silica precipitation at hot springs generally has been considered apassive process (Walter et aI., 1972), in that rapid cooling atambient air temperatures, evaporation and/or steam loss, and achange in pH of the hydrothermal waters following dischargecan all induce silicification of growing cells (Fournier, 1985).However, several recent studies using electron microscopy haveshown that microorganisms provide favorable nucleation sitesthat serve as templates for silicification. In sediments fromYellowstone, Ferris et a1. (1986) observed individual bacterialcells preserved in successive stages of amorphous silica encrustation; some cells were thoroughly embedded in a mineralizedmatrix. In New Zealand, bacterial filaments were similarlyencrusted by amorphous silica (Jones et aI., 1997a, b). Thatmicroorganisms catalyze silica precipitation was confirmedthrough analyses of microbial mats growing around a hot springvent in Geysir, Iceland, where filamentous phototrophic bacteriaresembling Chloroflexlls allmntiaclls, and the unicellularcyanobacterium SynechococCliS sp., underwent rapid silicification (Schultze-Lam et aI., 1995; Konhauser and Ferris, 1996).The mineralization associated with the cells occurred as spheroidal grains (200-500 nm in diameter) both extracellularly onthe sheaths of living bacteria and intracellularly within the cytoplasm presumably after the cells had lyzed. In many filaments,the silica crystallites appeared to have coalesced such that individual precipitates were no longer distinguishable (Fig. 3);entire colonies sometimes became cemented together in a

FIG. 3.-TEM image of a colony of filamentous bacteria (from Geysil; Iceland) with cpiccllular amorphous silica un outer sheath surfaces. Note tlie preserlce of a few tliscernible splicres 011 the outer edges on the ~iiinernlized matrix (al.rnw). Scale bar = I . 1 { ~ ~ i i .

siliceous matrix several micrometers thick (Kunhauser and Fessis, 1996). Eventually only the sheath and cell wall of the original organic framework remained recognizable. Thus, it appears that silicification begins when the ~nicroorganis~ns are living (i.e., they serve as nucleation sites) and continues for some time after their death. In the latter stages of silicification, the developiiient of amorplious silica cement between silicified microbial filaments is presuniably abiogenic (Jones et al.. 1997a). Silicification has also been shown to occur in tlieriiial springs in the Kenya Rift where water temperatures rcach above 30°C, temperatures too high to support cyanobacteria and most 3acteria. Instead, liyperthennopliilic bacteria were shown to sur- give in this extreme environment, preferentially fixing silica Into secreted rr~icrobial tnucus (Jones and Kenaut, 1996).

In a hydrothermal venting site with talus piles on the wall of 1 graben at 21°30'S, East Pacific Rise, hot water mixes with sea- uater in the talus piles producing sulfide and alnorphous silica leposits (Juniper and Fouquet, 1988). Upon magnification, the ;ilica crust revealed a filatnentous ~nicrotexture in which indi- iidual filaments were cemented together. Figure 4 shows a secion of srlch a filament heavily coated in coriccntric layers of leposited silica, with the hollow space likely representing the xiginal bacterial ccll (Juniper aud Foucluet, 1988).

Based on findings attained from studies on the silicification )f wood (Leo and Barghoorn, 1976; Sigleo, 1978), the associaion of silica with 1nicrobial s~~rfaces is probably established by ~ydrogen bonding of the hydroxyl groups in silicic acid Si(OH)4] to available hydroxy groups of the bacterial sheath Ferris et al., 1988). Continued reaclio~l of the bound silica with nore silicic acid, from the surrounding waters, results in polynerization and eventual dehydration, which leads to the formaion of silica crystallites (Iler, 1980). Results from artificial ossilization studies on inicroorganislns have confirmed that mcteria can bc coinpletely silicified (Oehler and Schopf, 197 1 ; :rancis et al., 1978; Westall et al., 1995), with the preferential

FIG. 4.-(A) Low-magnification SEM plio~o s h o w i ~ ~ g fila~nentous microtexture of liyd~~otliermal deposits (fro111 21°30'S. East Pacific Rise). ( B ) Section of a fila~~icnt heavily coated in silica. The hollow internal space likely represents the original fila~iient that was covered in concentric layers of depositecl silica. Reprinted ~\gitli permission of tlie author and Corlrrtlinrl ~\~lir~e~.nlogi.rr (Jtuniper 2nd Fouquet. 1988).

preservation of cell walls and sheaths (Oehler, 1976). Fenis et a]. (1988) have also shown the rapid development of silica crystallites on cellular debris following cell lysis. Evidently, lyzed ~nicroorganisms continue to serve as templates for silica deposition, with cellitlar degradation enhancing the mineralization process hy increasing the availability of functional groups capable of forining hydrogen bonds with dissolved silica (Knoll, 1985; Ferris et al., 1988). Ceileral features o l increasing silicification with time include thickening of silica crusts around tlie bacteria, increasing crystallization of the silica from loose aggregates into dense spheres, and eventually the formation of a tnainmillated external crust (Westall et al., 1995). Interestingly, pretreatment of cells with iron appeared to denature their autolytic enzymes, thus maintaining intact organic residues during silicification (Ferris et al., 1988).

BIFs: HYDROTHERMAl, BACTERIAL BIOMINERALIZATION 137

FIG. 3.-TEM image of a colony of filamentous bacteria (from Geysir,Iceland) with cpiccllular amorphous silica on outer sheath surfaces. Note the

presence of a few discernible spheres on the outer edges on the mineralizedmatrix (arrow). Scale bar = 1. I fun.

siliceous matrix several micrometers thick (Konhauser andFerris, 1996). Eventually only the sheath and cell wall of theoriginal organic framework remained recognizable. Thus, itappears that silicification begins when the microorganisms areliving (i.e., they serve as nucleation sites) and continues forsome time after their death. In the latter stages of silicification,the development of amorphous silica cement between silicifiedmicrobial filaments is presumably abiogenic (Jones et a!.,1997a). Silicification has also been shown to occur in thermal,prings in the Kenya Rift where water temperatures reach above:JOoC, temperatures too high to support cyanobacteria and mostJacteria. Instead, hyperthermophilic bacteria were shown to survive in this extreme environment, preferentially fixing silica)nto secreted microbial mucus (Jones and Renaut, 1996).

In a hydrothermal venting site with talus piles on the wall of1 graben at 21 °30'S, East Pacific Rise, hot water mixes with seaNater in the talus piles producing sulfide and amorphous silica:Ieposits (Juniper and Fouquet, 1988). Upon magnification, the;ilica crust revealed a filamentous microtexture in which indilidual filaments were cemented together. Figure 4 shows a secion of such a filament heavily coated in concentric layers ofleposited silica, with the hollow space likely representing the)riginal bacterial cell (Juniper and Fouquet, 1988).

Based on findings attained from studies on the silicification)f wood (Leo and Barghoorn, 1976; Sigleo, 1978), the associaion of silica with microbial surfaces is probably established bylydrogen bonding of the hydroxyl groups in silicic acidSi(OH)4] to available hydroxy groups of the bacterial sheathFerris et a!., 1988). Continued reaction of the bound silica withnore silicic acid, from the surrounding waters, results in polynerization and eventual dehydration, which leads to the formaion of silica crystallites (Iler, 1980). Results from artificialossilization studies on microorganisms have confirmed thatlacteria can be completely silicified (Oehler and Schopf, 1971;;rancis et aI., 1978; Westall et aI., 1995), with the preferential

FIG. 4.-(A) Low-magnification SEM photo showing filamentous microtexture of hydrothermal deposits (from 21 °30'5, East Pacific Rise). (B) Section

of a filament heavily coated in silica. The hollow internal space likely representsthe original filament that was covered in concentric layers of deposited silica.Reprinted with permission of the author and Canadian Mineralogist (Juniperand Fouquet, 19S5).

preservation of cell walls and sheaths (Oehler, 1976). Ferris etal. (1988) have also shown the rapid development of silica crystallites on cellular debris following cell lysis. Evidently, lyzedmicroorganisms continue to serve as templates for silica deposition, with cellular degradation enhancing the mineralizationprocess by increasing the availability of functional groups capable of forming hydrogen bonds with dissolved silica (Knoll,1985; Ferris et aI., 1988). General features of increasing silicification with time include thickening of silica crusts around thebacteria, increasing crystallization of the silica from looseaggregates into dense spheres, and eventually the formation of amammillated external crust (Westall et aI., 1995). Interestingly,pretreatment of cells with iron appeared to denature theirautolytic enzymes, thus maintaining intact organic residues during silicification (Ferris et aI., 1988).

Bacterial cells, associated with granular and spheroidal crystallites composecl excl~isively of iron and silica, have been examined froin acidic hot spring sediments froin Yellowstone National Park, USA (Fcrris et al., 1986) and from mats growing in hot spring outflow channels, Krisuvik, Icela~ld (Konhauser ancl Ferris, 1996). In the Yellowstolle samples, both intact and lyzecl cells were mineralized, wit11 grains typically smaller than 100 nm (Ferris et al., 1986); one cell that appeared to have undergone an extended episode of iron-silica growth and corn- paction had an interlocking inosaic that coinpletely encrusted the microorganism. In Krisuvik, bacterial cells were associated with diverse ~nilleral assemblages (Konhauser and Ferris, 1996). In soine cells several spheroidal grains (-200 11111) were shown to be embedded in dense Fe-rich capsular material. In other cells from the same sample site, larger amorphous precipitates (-500 nm), consisti~lg of iron and silica in approximately equal proportions, were evident directly on the cell wall (Fig. 5).

Iron-silica deposits have also been identified from v ~ I ~ O L I S -'

deep-sea hydrothennal sites. One active site at the Philosopher Vent on Explorer Ridge in the northwest Pacific had iron-silica deposits forming with a filarne~ltous moi-phology (Juniper aild Fouquet, 1988). The resemblance of these filainents to those produced by microorganisms, along with the presence of organic carbon (1.3%) and filamentous bacteria (from one vent sample), suggested that bacteria provided substrata for mineral nucleation. Transmitted light ~nicroscopy of these iron-silica deposits fi~rther identified intense iron accumulation on the fila- meIlts, up011 which silica precipitated (Fig. 6). Extensive accumulation of microbial-minerd flocs (dominated by iron and silica oxides) were also observed in association with ab~indant diffuse venting in new lava flows and around less extensive, new venting on older basalt in the CoAxial Segment of the Juan de Fuca Ridge (Juniper et a]., 1995). The presence of filame~lts (resembling Leptotlz~.is sp.) in nontroilite deposits (Fe-rich sinectite) fro111 seamounts in the East Pacific Rise (Alt, 1988),

a11d the distinct microtube-like ~norphology in nontronite deposits from white smoker chimncys in the Galapagos Rift and Marialla Trough (Kohler ct al., 1994), also suggest that bacteria are directly i~lvolved in catalyzing clay mi~leral formation.

The precipitation of iron-silicate phases closely follows the two-step mineralization process origillally described by Beveridge and Mul-l-ay (1976), in that iron is initially bound to anionic cellular sites, after which dissolved silica is added to the growing mineral via hydrogen bonding between hydroxyl groups. The Fe-rich sites 011 the cell surface presumably serve as precursors to the more complex surface precipitates, with the iron-silicates using some fraction of the precursor surface as a template for their own growth, in effect circurnvelltillg the need for direct nucleation (Steefel and Van Cappellen, 1990). This sequence of reactions appears to cori-espond with previous theories of hydrothermal nontronite formation that suggested these clays formed either by sorption of silica onto Fe-oxyhydroxides or directly through precipitation from hydrothennal fluids (Kohler et al., 1994).

Experimental work has also confirmed the ability of bacterial cells to form iron-silicate minerals. Silicate anion binding to bacteria was enhanced \vhen the cell walls were preloaded with Fe(IlI), indicating that the metal participated in a cationic bridging mecha~lism for silicate binding to the bacterial surface through the formation of ternary co~nplexes (e.g., wall-metalsilicate) (Urrutia and Beveridge, 1993, 1994). Growth of the precipitates then continued until co~nplex silicate structures were formed.

DISCUSSION

There are three similarities between present-day hydrothermal systems and Precambrian oceans that support this paper's contention that modern bacterial biornineralization may provide small-scale analogues to the conditions under which BlFs were

FIG. 5.-TEM inlage of a bacterial ccll from Krisuvik, Iceland, n~i th amorphous iron-silicate grains on the outer cell wall. Scale bar = 200 mil.

FIG. 6.-Transmittetl ligllt microscopy image of the filament-mineral association in iron-silica deposits in a chimney fragment from 12"50rN, East Pacific Rise. The transition from iron-rich (dark) to silica-rich (light) over a tlistance of 2 mm is sho\vn. Scale bar = 2 ~(111. Reprinted with permission of the author and Cnrzrrdinrl il/lirler.rrlogist (Juniper ant1 Fouquet, 1988).


fron Silicates

Bacterial cells, associated with granular and spheroidal crystallites composed exclusively of iron and silica, have beenexamined from acidic hot spring sediments from YellowstoneNational Park, USA (Ferris et aI., 1986) and from mats growingin hot spring outflow channels, Krisuvik, Iceland (Konhauserand Ferris, 1996). In the Yellowstone samples, both intact andIyzed cells were mineralized, with grains typically smaller than100 nm (Ferris et aI., 1986); one cell that appeared to haveundergone an extended episode of iron-silica growth and compaction had an interlocking mosaic that completely encrustedthe microorganism. In Krisuvik, bacterial cells were associatedwith diverse mineral assemblages (Konhauser and Ferris, 1996).In some cells several spheroidal grains (-200 nm) were shownto be embedded in dense Fe-rich capsular material. In other cellsfrom the same sample site, larger amorphous precipitates (-500nm), consisting of iron and silica in approximately equal proportions, were evident directly on the cell wall (Fig. 5).

Iron-silica deposits have also been identified from variousdeep-sea hydrothermal sites. One active site at the PhilosopherVent on Explorer Ridge in the northwest Pacific had iron-silicadeposits forming with a filamentous morphology (Juniper andFouquet, 1988). The resemblance of these filaments to thoseproduced by microorganisms, along with the presence oforganic carbon (1.3%) and filamentous bacteria (hom one ventsample), suggested that bacteria provided substrata for mineralnucleation. Transmitted light microscopy of these iron-silicadeposits further identified intense iron accumulation on the filaments, upon which silica precipitated (Fig. 6). Extensive accumulation of microbial-mineral flocs (dominated by iron andsilica oxides) were also observed in association with abundantdiffuse venting in new lava flows and around less extensive, newventing on older basalt in the CoAxial Segment of the Juan deFuca Ridge (Juniper et aI., 1995). The presence of filaments(resembling Leptothrix sp.) in nontronite deposits (Fe-richsmectite) from seamounts in the East Pacific Rise (Alt, 1988),

FIG. 5.-TEM image of a bacterial cell from Krisuvik. Iceland. with amorphous iron-silicate grains on the outer cell wall. Scale bar = 200 nm.

and the distinct microtube-like morphology in nontronitedeposits from white smoker chimneys in the Galapagos Rift andMariana Trough (Kohler et aI., 1994), also suggest that bacteriaare directly involved in catalyzing clay mineral formation.

The precipitation of iron-silicate phases closely follows thetwo-step mineralization process originally described byBeveridge and Murray (1976), in that iron is initially bound toanionic cellular sites, after which dissolved silica is added to thegrowing mineral via hydrogen bonding between hydroxylgroups. The Fe-rich sites on the cell surface presumably serve asprecursors to the more complex surface precipitates, with theiron-silicates using some fraction of the precursor surface as atemplate for their own growth, in effect circumventing the needfor direct nucleation (Steefel and Van Cappellen, 1990). Thissequence of reactions appears to correspond with previous theories of hydrothermal nontronite formation that suggested theseclays formed either by sorption of silica onto Fe-oxyhydroxidesor directly through precipitation from hydrothermal fluids(Kohler et a!., 1994).

Experimental work has also confirmed the ability of bacterial cells to form iron-silicate minerals. Silicate anion binding tobacteria was enhanced when the cell walls were preloaded withFe(III), indicating that the metal participated in a cationic bridging mechanism for silicate binding to the bacterial surfacethrough the formation of ternary complexes (e.g., wall-metalsilicate) (Urrutia and Beveridge, 1993, 1994). Growth of theprecipitates then continued until complex silicate structureswere formed.

DISCUSSION

There are three similarities between present-day hydrothermal systems and Precambrian oceans that support this paper'scontention that modern bacterial biomineralization may providesmall-scale analogues to the conditions under which BIFs were

FIG. 6.-Transmitted light microscopy image of the filament-mineral association in iron-silica deposits in a chimney fragment from 12°50'N, East PacificRise. The transition from iron-rich (dark) to silica-rich (light) over a distance of2 nUll is shown. Scale bar = 2 ~lm. Reprinted with permission of the author andCanadian Mineralogist (Juniper and Fouquet, 1988).

deposited. First, modern deep-sea vents zund hot springs are characterized by the release of wart11 to hot reducing solutions with high ele~nental concentrations (Corliss et al., 1979; He~lley, 1984). For example, effluent collected from vents along the Mid-Atlantic Ridge and the East Pacific Rise co~ltained dissolved silica and iron (averaging between the two sites) approximately 1,100 p p ~ n and 100 ppm, respectively (Campbell et al., 1988). These values are approximately 1,000 times and 50,000 times greater tlian concentrations for silica and iron, respec- tijely, ia background liiarine waters (Drever, 1988). Similarly, hot spring efiluents (from a variety of sites globally) have dissolved silica concentrations typically exceeding 200 ppm, with some geothermal wells having up to 1,100 ppm (Henley, 1984), while iron concentrations range from 0.8 ppm (Ljisul~bll. Iceland) (Konhauser and Ferris, 1996) to 2.4 ppln (Ma~nmotl~ Springs, Yellowstone) (Foumier, 1989). The conzbination of high dissolved iron and silica are unique to hydrothermal systems, and are analogous to conditions in much of tlze Precambrian, when ocean chemistry was controlled by the interaction of warln seawater with the basaltic oceanic crust (Veizer et al., 1982). In the Archean and early Proterozoic, hydrothermal effluent was transported from the deep sea into the oxic surface waters by u~lusual geostrophic or wind-driven upn~elling (Lowe, 1994) or disectly by hydrothermal plumes generated at mid- ocean ridge crests with substantially shallower depths than present (Isley, 1995). This led to surface seawater with high metal concentrations, perhaps up to 250 ppln iron (Veizer, 1983), and silica concentrations close to or at saturatiolz (-120 ppm) with respect to amol-phous silica (Siever, 1992).

Secondly, a consortium of ~nicroorg~misms are currently present in hot springs and deep-sea vents, some of which resemble those found as Precarnbrian microfossils, including cya~~obacte- ria, anoxygenic phototrophic bacteria, and iron-depositing bacteria. There are several morphological lines of evidence that suggest that cyanobacteria existed at 3.5 Ga (Awramik, 1992), including tlie presence, in rocks from the Warrawoona Group (-3.5 Ga), of trichome-like filanle~zts (Awramik et al., 1983); ~nicrofossils with widtlzs greater tlian most filamentous bacteria (Schopf, 1993); and coccoid forms e~iclosed in a wall-like structure (Sclzopf and Packer, 1987). The existelice of strolnatolites (presumed to have formed in shallow, near-slzore inari~ze settings) f~lrther implies tlze presence of cj~anobacteria, since the microorganisms that formed them were phototactic, photonu- totrophic, and produced mucus sheaths, all attributes found in present-day cyanobacteria (Aivramik, 1992). In addition to cyanobacteria, mats fornzed by sulfide-dependent, pliototropliic bacteria (e.g., Chlor-(~jlenrs sp.), and possibly sulfur-oxidizing bacteria such as Beggiritorr sp., may also lzave been locally abundant in the Precambrian (Doemel and Brock, 1977; Walter et al., 1972, 1992). Lastly, the presence of thread-like structures in chert fro111 the Gunflint 11-011 For~natio~l, Ontario (aged - 2.0 Ga) that resemble filamentous iron-depositing bacteria such as C~zrzotht.is sp., Sl~hrrei.otil~ls sp., and Lel~tothr-i.1- ssp., additionally suggcsts the presence of bacteria that \Yere involved in the oxidation of ferrous to ferric iron (Barglzoorn and Tyler, 1965; Cloud, 1965).

The resenlblance of a wide consortia of early Precanzbria~z

microorganisms to tlzose existing toda)? is not surprising. During that period in Earth's history, when nlajor BIFs were accumu- luting, microorganisms \yere already Ilourishing. with nearly all prokaryotic phyla llaving by that time e\wl\~ed and diversified (Awramik, 1992). This view is supported by the presence of fossilized microbial remains, stromatoliles, chemical fossils, and biogenic ~ni~ierals reported from early Arclzean rocks (Lowe, 1980; Awramik et al., 1983; Hayes et al., 1983; Ohmoto et nI., 1993), sonle at least 3.8 Ga (Mojzsis et al., 1996). In fac~, the construction of strolnatolites by 3.5 Ga implies that these microbes were morphologically well-developed by that time (A\vra~nik, 1992: McNa~nara and A\vmmik, 1992).

The third similarity reflects the cornmon occurrence of iron and silicate minerals around modern hydrothermal vents and their presence as primary mineralized constituents of BIFs. Iron hydroxides and amorphous silica are tlze predon-tinant ~ninesal phases tlzat precipitate around lokv-temperature, sulfide-deficient, deep-sea vents, where they can form centi~neter-thick layers (by diffuse venting through underlying basalts or sulfides), oxide muds, and chilnney structures (Juniper and Tebo, 1995). At hot spri~igs, iron and silicate biominerals are also observed, sonletinles displaying crude ~nicroba~lcling (K.O. Konhauser, in prep., 2000) reli~iniscent of the very finely laminated iron-chert beds in, for example, tlze Ha~nersley BIF (Morris, 1993).

As discussed above, bacteria contribute to the precipitation of authigenic minerals in modern hydrothermal systems. Herein lies the premise of this paper. If microorganisms are present today in hydrothermal environments, and if they precipitate iron and silicate minerals, then it seenls reasonable to argue that Precambrian bacterial populations, also growing in hydrotlzer- ma1 waters, coulci have formed similar biominerals, i.e., the pri- lnary minerals ia BIFs. 1 speculate that previous models for BIF deposition (discussed below) have underestimated the role of ~nicroorganisnis in the deposition of BIFs. 111 this new Ilppotlze- sis, bacteria, growing in the Precambrian oceans. lnap have served as nucleation sites that catalyzed primary iron and silica mineralization.

In tlie early Precambrian, biological activity was probably limited to the flat ~nargins of the microcontinents. \vhere microorganisms grew as benthic mats or in the open ocean above areas of dynamic upwelling of deep, nutrien-rich ~vaters or hydrothermal plumes (Lowe, 1994). The planktonic microorganisms presumably flourished when ferrous iron ancl nutrients were available, allowing them to for111 iron hydroxides, but they declined in numbers (no mineralization period) when these components became depleted (Cloud, 1973; Baur et al., 1985). It is unclear what proportion of tlze ferric iron was initially produced in the Arclzean via plzotoclzemical processes (Caisns- Smith, 1978; Braterman et a]., 1983; Fran~ois , 1986), anoxygenic pl~otosyntl~esis (Widdel et a]., 1993; Elzrenreich and Widclel, 1994), photodissociation of water vapor in the atmosphere, generating free oxygen to react with rerrous iron (Walker et al., 1983), or oxygenic photosyntl~esis (Cloud, 1965, and many others). In a large part it will depend on when cyanobacteria came into existence. If they existed as early as 3.5 Ga (Awramilc, 1993), then these niic~.oorganis~i~s ~vould certainly have had a major impact on the initial procluction of oxygen

BIFs: HYDROTHERMAL BACTER1AL BIOMINERAUZATlON 139

deposited. First, modern deep-sea vents and hot springs arecharacterized by the release of warm to hot reducing solutionswith high elemental concentrations (Corliss et aI., 1979; Henley,1984). For example, effluent collected from vents along theMid-Atlantic Ridge and the East Pacific Rise contained dissolved silica and iron (averaging between the two sites) approximately 1,100 ppm and JOO ppm, respectively (Campbell et aI.,1988). These values are approximately 1,000 times and 50,000times greater than concentrations for silica and iron, respectively, in background marine waters (Drever, 1988). Similarly,hot spring eft1uents (from a variety of sites globally) have dissolved silica concentrations typically exceeding 200 ppm, withsome geothermal wells having up to 1, 100 ppm (Henley, 1984),while iron concentrations range from 0.8 ppm (Lysuh611,Iceland) (Konhauser and Ferris, 1996) to 2.4 ppm (MammothSprings, Yellowstone) (Fournier, 1989). The combination ofhigh dissolved iron and silica are unique to hydrothermal systems, and are analogous to conditions in much of thePrecambrian, when ocean chemistry was controlled by the interaction of warm seawater with the basaltic oceanic crust (Veizeret aI., 1982). In the Archean and early Proterozoic, hydrothermaleffluent was transported from the deep sea into the oxic surfacewaters by unusual geostrophic or wind-driven upwelling (Lowe,1994) or directly by hydrothermal plumes generated at midocean ridge crests with substantially shallower depths than present (Isley, 1995). This led to surface seawater with high metalconcentrations, perhaps up to 250 ppm iron (Veizer, 1983), andsilica concentrations close to or at saturation (-120 ppm) withrespect to amorphous silica (Siever, 1992).

Secondly, a consortium of microorganisms are currently present in hot springs and deep-sea vents, some of which resemblethose found as Precambrian microfossils, including cyanobacteria, anoxygenic phototrophic bacteria, and iron-depositing bacteria. There are several morphological lines of evidence thatsuggest that cyanobacteria existed at 3.5 Ga (Awramik, 1992),including the presence, in rocks from the Warrawoona Group(-3.5 Ga), of trichome-like filaments (Awramik et aI., 1983);microfossils with widths greater than most filamentous bacteria(Schopf, 1993); and coccoid forms enclosed in a wall-like structure (Schopf and Packer, 1987). The existence of stromatolites(presumed to have formed in shallow, near-shore marine settings) further implies the presence of cyanobacteria, since themicroorganisms that formed them were phototactic, photoautotrophic, and produced mucus sheaths, all attributes found inpresent-day cyanobacteria (Awramik, 1992). In addition tocyanobacteria, mats formed by sulfide-dependent, phototrophicbacteria (e.g., Chloroflexlis sp.), and possibly sulfur-oxidizingbacteria such as Beggiaton sp., may also have been locally abundant in the Precambrian (Doemel and Brock, 1977; Walter et '11.,1972, 1992). Lastly, the presence of thread-like structures inchert from the Gunflint Iron Formation, Ontario (aged - 2.0 Ga)that resemble filamentous iron-depositing bacteria such asCrenothrix sp., Sphaerotillis sp., and Leptothrix sp., additionallysuggests the presence of bacteria that were involved in the oxidation of ferrous to ferric iron (Barghoorn and Tyler, 1965;Cloud, 1965).

The resemblance of a wide consortia of early Precambrian

microorganisms to those existing today is not surprising. Duringthat period in Earth's history, when major BIFs were accumulating, microorganisms were already flourishing, with nearly allprokaryotic phyla having by that time evolved and diversified(Awramik, 1992). This view is supported by the presence of fossilized microbial remains, stromatolites, chemical fossils, andbiogenic minerals reported from early Archean rocks (Lowe,1980; Awramik et al., 1983; Hayes et aI., 1983; Ohmoto et aI.,1993), some at least 3.8 Ga (Mojzsis et '11., 1996). In fact, theconstruction of stromatolites by 3.5 Ga implies that thesemicrobes were morphologically well-developed by that time(Awramik, 1992; McNamara and Awramik, 1992).

The third similarity reflects the common occurrence of ironand silicate minerals around modern hydrothennal vents andtheir presence as primary mineralized constituents of BIFs. Ironhydroxides and amorphous silica are the predominant mineralphases that precipitate around low-temperature, sulfide-deficient, deep-sea vents, where they can form centimeter-thick layers (by diffuse venting through underlying basalts or sulfides),oxide muds, and chimney structures (Juniper and Tebo, 1995).At hot springs, iron and silicate biominerals are also observed,sometimes displaying crude microbanding (K.O. Konhauser, inprep., 2000) reminiscent of the very finely laminated iron-chertbeds in, for example, the Hamersley BIF (Morris, 1993).

As discussed above, bacteria contribute to the precipitationof authigenic minerals in modern hydrothermal systems. Hereinlies the premise of this paper. If microorganisms are presenttoday in hydrothermal environments, and if they precipitate ironand silicate minerals, then it seems reasonable to argue thatPrecambrian bacterial populations, also growing in hydrothermal waters, could have formed similar biominerals, i.e., the primary minerals in BIFs. I speculate that previous models for BIFdeposition (discussed below) have underestimated the role ofmicroorganisms in the deposition of BIFs. In this new hypothesis, bacteria, growing in the Precambrian oceans, may haveserved as nucleation sites that catalyzed primary iron and silicamineralization.

In the early Precambrian, biological activity was probablylimited to the flat margins of the microcontinents, wheremicroorganisms grew as benthic mats or in the open oceanabove areas of dynamic upwelling of deep, nutrient-rich watersor hydrothermal plumes (Lowe, 1994). The planktonic microorganisms presumably flourished when ferrous iron and nutrientswere available, allowing them to form iron hydroxides, but theydeclined in numbers (no mineralization period) when thesecomponents became depleted (Cloud, 1973; Baur et aI., 1985).It is unclear what proportion of the ferric iron was initially produced in the Archean via photochemical processes (CairnsSmith, 1978; Bratennan et aI., 1983; Franyois, 1986),anoxygenic photosynthesis (Widdel et aI., 1993; Ehrenreich andWiddel, 1994), photodissociation of water vapor in the atmosphere, generating free oxygen to react with ferrous iron (Walkeret aI., 1983), or oxygenic photosynthesis (Cloud, 1965, andmany others). In a large part it will depend on when cyanobacteria came into existence. If they existed as early as 3.5 Ga(Awramik, 1992), then these microorganisms would certainlyhave had a major impact on the initial production of oxygen

140 KURT 0. KOIVHAUSER

(Cloud, 1965, 1973). Because the initial O2 partial pressures may have been extremely low (Walker et al., 1983), bacteria would not have bee11 able to gain energy from oxidation of Fe(I1) to Fe(II1) with O2 as an electron acceptor, using instead energy sources such as sunlight to oxidize Fe (Ehrlich, 1990). As 0 2 partial pressures increased beyond some threshold concentration, oxidation by microaerophilic bacteria such as Gallio~zellcr sp. would have become possible and perhaps even favored (under low p02) compared to inorganic oxidation of Fe(l1) (Holm, 1987).

The end result of any of these modes of biolnineralization would be the precipitation of ferric hydroxide in the water column, its sedimentation, and the systelnatic acculnulation of iron-rich sediment on the ocean floor (Walker et al., 1983). Along with ferric iron, the remains of the lnicroorganisrn would be deposited. This is suggested by the connnon presence of spheroidal structures in BIFs, which have been interpreted as relict microfossils (Cloud and Licari, 1968; LaBerge, 1973). Once deposited, the ferric hydroxide eventually would have consolidated as (i) l~einatite, where oxidizing conditions were maintained in the sediment or organic inaterial was limited; (ii) as magnetite, where partial Fe-reduction occurred in the sediments by microbial organic matter mineralization; or (iii) as sideritelferrous silicate when substantial Fe-reduction occurred (Morris, 1993). Magnetite and siderite are both niajor components in iron-oxide facies and S-rnacrobands, respectively (Ewers and Morris, 1981), and their presence suggests that a significant fraction of BIFs were diagenetically modified under suboxic to anoxic conditions (Walker et al., 1983). That the latter diagenetic reactions were microbially catalyzed is suggested by isotopically light 613c values in BIF carbonate minerals (Pessy et al., 1973; Baur et al., 1985), which implies that isotopically light organic carbon in the bottom sediments was oxidized to bicarbonate with the concomitant reduction of some electron acceptor, presumably ferric iron (Walker, 1984). This oxidation may also explain the absence of organic inatter in BIFs. The by-product of Fe-reduction could be either magnetite, cominonly evident as overgrowths 011 hematite (Morris, 1980), or siderite, which is present in the Gunflint Iron Fonnation, Canada, for example, as uniform, spherical grains (approximately 30-40 Lun in diameter) with organic-rich cores (Goodwin, 1956). This correlates well with the suggestion that during the hundreds of inillions of years when ferric iron was being deposited and organic matter accumulated in the sediment, that a simple respiratory chain to ferric iron developed and eventually served as a precursor to more complex respiratory chains associated with extant organisms (Nealson and Myers, 1990).

The fundamental differences between lnineralizatioll in present-day hydrothermal environments and the Precambrian oceans are (i) the size of the deposits and (ii) the relationship between the microorganisms and the chemocline (where anoxic, ferrous-rich waters meet oxygenated surface waters). In the first instance, the oxygenated conditions on the ocean floor today preclude the formation of large-scale deposits in modern hydrothermal environments. Mineral precipitation occurs peripheral to the vent, where anoxic hydrothermal fluids are

rnixed with oxygenated deep-sea waters (Juniper and Tebo, 1995). I11 the presence of oxygen and at pH-8 (the half-life of fenous iron being only be a few minutes, Millero et al., 19871, fel-rous iron is rapidly oxidized to ferric iron and precipitated as ferric hydroxide (Schwertmann and Fitzpatrick, 1992). Even if the hydrothermal effluent is hot (giving low oxygen solubility) and acidic, and the surrounding ambient ocean bottom waters are low in dissolved 02, ferric hydroxides still accumulate in relative proximity to the vents, with a gradual depletion outwards (Juniper and Tebo, 1995). Thus, high concentrations of ferrous iron are not found in modern marine waters. Similarly, in hot spring effluents there is an allnost immediate precipitation of ferric hydroxide near the geyser. In contrast, the anoxic bottonl waters of the Precambrian oceans allowed for iron to dissipate greater distances away froin the vent, and subsequently precipitate out of solution in the deep oceans during the Archean or in huge basinal areas (e.g., 100,000 km2 for the Hamersley Group, Trendall and Blockley, 1970) d ~ ~ r i n g the Proterozoic (Beukes and Klein, 1992; Simonson and Hassler, 1996). Precipitation occurred over hundreds of millions of years as the oceans became increasingly oxygenated, and it was only when the rate of photosynthetic oxygen production exceeded the rate of sug- ply of ferrous iron that it becanle substantially depleted in marine waters and free oxygen accu~nulated (Walker et al., 1983). Based on the oxidation state of paleosols, the nature of uranium ores, the presence of red-beds, trace metal content in black shales, and the evolution of eukaryotes, this may not have occurred until after 2.2 Ga (Holland, 1994).

The second key difference between present-day hydrothermal en\lironments and those under which BIFs were deposited is that today, extant microorganisms, growing as siiats in hot spring effluent or on vent chilnneys and in surrounding marine sediments, are foi~nd at the chemocline. In the Precambrian, however, biolnineralization was rernoved from the site of hydrothermal emissions, allowing the "biogenic" deposits to form thick accun~ulatiolls of great lateral extent. The disparity is due to radically different conditions on the early Earth. The Archean (at the time 3.5-3.2 Ga) was characterized by high rates of mantle outgassing, limited chemical weathering, elevated atmospheric C 0 2 levels and surface temperatures of perhaps 30-50°C that prevented the accumulation of polar ice, cold-water sinking, and deep ocean stirring (Lowe, 1994). The Archean oceans were thus strongly stratified with a shallow, wind-mixed surface layer containing dissolved C 0 2 and small amounts of 02, which were derived from oxygenic photosynthesis in areas of high planktonic productivity (Cloud, 1965, 1973; Schidlowski, 1976), and from pl~otocl~einical processes that dissociate water vapor (Walker et al., 1983). The deeper, reducing waters were enriched in ferrous iron (Klein and Beukes, 1989; Beukes et al., 1990; Beukes and Klein, 1992).

The distribution of Archean iron-formations suggests that these relatively small deposits (compared to Proterozoic BIFs) were ubiquitous in deep water, not overwhelmed by coarse clastic debris (Lowe, 1994), and presulnably represented local accumulations close to marine vents (Simonson and Hassles, 1996). The relative absence of shallow-water BIFs indicates that the upper, wind-mixed layer of the ocean contained little dissolved


(Cloud, 1965, 1973). Because the initial O2 partial pressuresmay have been extremely low (Walker et aI., 1983), bacteriawould not have been able to gain energy from oxidation ofFe(II) to Fe(III) with O2 as an electron acceptor, using insteadenergy sources such as sunlight to oxidize Fe (Ehrlich, 1990).As O2 partial pressures increased beyond some threshold concentration, oxidation by microaerophilic bacteria such asGallionella sp. would have become possible and perhaps evenfavored (under low p02) compared to inorganic oxidation ofFe(II) (Holm, 1987).

The end result of any of these modes of biomineralizationwould be the precipitation of ferric hydroxide in the water column, its sedimentation, and the systematic accumulation ofiron-rich sediment on the ocean floor (Walker et aI., 1983).Along with ferric iron, the remains of the microorganism wouldbe deposited. This is suggested by the common presence ofspheroidal structures in BIFs, which have been interpreted asrelict microfossils (Cloud and Licari, 1968; LaBerge, 1973).Once deposited, the ferric hydroxide eventually would haveconsolidated as (i) hematite, where oxidizing conditions weremaintained in the sediment or organic material was limited; (ii)as magnetite, where partial Fe-reduction occurred in the sediments by microbial organic matter mineralization; or (iii) assiderite/ferrous silicate when substantial Fe-reduction occurred(Morris, ]993). Magnetite and siderite are both major components in iron-oxide facies and S-macrobands, respectively(Ewers and Morris, 198]), and their presence suggests that asignificant fraction of BIFs were diagenetically modified undersuboxic to anoxic conditions (Walker et at., 1983). That the latter diagenetic reactions were microbially catalyzed is suggestedby isotopically light (PC values in BIF carbonate minerals(Perry et aI., 1973; Baur et aI., 1985), which implies that isotopically light organic carbon in the bottom sediments was oxidized to bicarbonate with the concomitant reduction of someelectron acceptor, presumably ferric iron (Walker, 1984). Thisoxidation may also explain the absence of organic matter inBIFs. The by-product of Fe-reduction could be either magnetite,commonly evident as overgrowths on hematite (Morris, 1980),or siderite, which is present in the Gunflint Iron Formation,Canada, for example, as uniform, spherical grains (approximately 30-40 ~l1n in diameter) with organic-rich cores(Goodwin, 1956). This correlates well with the suggestion thatduring the hundreds of millions of years when ferric iron wasbeing deposited and organic matter accumulated in the sediment, that a simple respiratory chain to ferric iron developed andeventually served as a precursor to more complex respiratorychains associated with extant organisms (Nealson and Myers,]990).

The fundamental differences between mineralization in present-day hydrothermal environments and the Precambrianoceans are (i) the size of the deposits and (ii) the relationshipbetween the microorganisms and the chemocline (where anoxic,ferrous-rich waters meet oxygenated surface waters). In the firstinstance, the oxygenated conditions on the ocean floor todaypreclude the formation of large-scale deposits in modernhydrothermal environments. Mineral precipitation occursperipheral to the vent, where anoxic hydrothermal fluids are

mixed with oxygenated deep-sea waters (Juniper and Tebo,1995). In the presence of oxygen and at pH-8 (the half-life offerrous iron being only be a few minutes, Millero et aI., 1987),ferrous iron is rapidly oxidized to ferric iron and precipitated asferric hydroxide (Schwertmann and Fitzpatrick, ]992). Even ifthe hydrothermal effluent is hot (giving low oxygen solubility)and acidic, and the surrounding ambient ocean bottom watersare low in dissolved 02, ferric hydroxides still accumulate in relative proximity to the vents, with a gradual depletion outwards(Juniper and Tebo, 1995). Thus, high concentrations of ferrousiron are not found in modern marine waters. Similarly, in hotspring effluents there is an almost immediate precipitation offerric hydroxide near the geyser. In contrast, the anoxic bottomwaters of the Precambrian oceans allowed for iron to dissipategreater distances away from the vent, and subsequently precipitate out of solution in the deep oceans during the Archean or inhuge basinal areas (e.g., 100,000 km2 for the Hamersley Group,Trendall and Blockley, 1970) during the Proterozoic (Beukesand Klein, 1992; Simonson and Hassler, 1996). Precipitationoccurred over hundreds of millions of years as the oceansbecame increasingly oxygenated, and it was only when the rateof photosynthetic oxygen production exceeded the rate of supply of ferrous iron that it became substantially depleted inmarine waters and free oxygen accumulated (Walker et aI.,1983). Based on the oxidation state of paleosols, the nature ofuranium ores, the presence of red-beds, trace metal content inblack shales, and the evolution of eukaryotes, this may not haveoccurred until after 2.2 Ga (Holland, 1994).

The second key difference between present-day hydrothermal environments and those under which BIFs were deposited isthat today, extant microorganisms, growing as mats in hot springeffluent or on vent chimneys and in surrounding marine sediments, are found at the chemocline. In the Precambrian, however, biomineralization was removed from the site ofhydrothermal emissions, allowing the "biogenic" deposits toform thick accumulations of great lateral extent. The disparity isdue to radically different conditions on the early Earth. TheArchean (at the time 3.5-3.2 Ga) was characterized by highrates of mantle outgassing, limited chemical weathering, elevated atmospheric CO2 levels and surface temperatures of perhaps 30-50°C that prevented the accumulation of polar ice,cold-water sinking, and deep ocean stirring (Lowe, 1994). TheArchean oceans were thus strongly stratified with a shallow,wind-mixed surface layer containing dissolved CO2 and smallamounts of 02, which were derived from oxygenic photosynthesis in areas of high planktonic productivity (Cloud, ]965,1973; Schidlowski, 1976), and from photochemical processesthat dissociate water vapor (Walker et aI., 1983). The deeper,reducing waters were enriched in ferrous iron (Klein andBeukes, 1989; Beukes et aI., 1990; Beukes and Klein, 1992).

The distribution of Archean iron-formations suggests thatthese relatively small deposits (compared to Proterozoic BIFs)were ubiquitous in deep water, not overwhelmed by coarse clastic debris (Lowe, 1994), and presumably represented local accumulations close to marine vents (Simonson and Hassler, 1996).The relative absence of shallow-water BIFs indicates that theupper, wind-mixed layer of the ocean contained little dissolved

BIFFs: H S D R O T H E R I ~ ~ ~ ~ L BACTERIAL BIOI\.IIIVEKALIL/\TI~~\'

FIG. 7.-Schematic tliagram of the Arcliean environment, modified from Lowe (1994). The Archean oceans were strongly stratified with a shallo\v surface layer containing ~ ~ i i a l l amounts of oxygen, \vIlile tlie tlecper reducil~g \\titers were cnriclied in ferrous iron. Benthic mats grew on the shallow sediments of the microcontincntal shelf (S), \vliile bacterial pl;~nkton gre\IJ in the open ocean above arras of dynamic up\velling or hydmthernlal plumes. At the cliemocline (C), ferric hytlroxide precipitation was catalyzed by plankton tlirougll binding of cationic iron onto their cells; the source of the ferric iron was a combination of pllotocliemical processes. anoxygenic or oxygenic photosyntl~esis. Thus, all iron \var removed from the deep \traters before it spread onto the shelf. Upon death, these iron-rich cells iccumulated in tlie bottom sedimenra iB), \\'liere diagcnetic reactions associated \\,it11 burial let1 to the transformation of ferric hydroxide into hematite, magnetite. or siderite. Iron deposition may have bcen episotlic clue to fluctuating nutrient supplies for the overlying plankton. Precipitation of silica occurl-ed on the ocean floor clue to pressure/tempel.at~~~-e changes after venting. At the chemocline, soiile silica {nay also Ilave reacted \\,it11 cell-bound iron to forin clay-like 11hases. The remai~iing dissolved silic:~ was transpo~tctl onto the shelf, \\~liel-e silica supe~.saturation (via evaporalion) allowed microorganisms to induce silicificatio~l, leading to the forn~ation of silicified stromatolites.

iron, quite possibly due to iron being removed from upwelling deep waters before it spread onto adjacent continental shelves (Lowe, 1994). This is reflected in Archean shallow water deposits that are depleted in iron (Beukes and Klein, 1992). Thus, it seems likely that in the Archean major BIFs \vould have formed above oceanic hydrothermal vents or in areas of ~tpwelling, where ferrous iron and dissolved silica were brought into contact with oxygenated water at the chemocline (Beukes and Klein, 1992), yet below the level of the shallow contine~ltal shelf (Lowe, 1994). Plankton growing in the sltrface waters of the open ocean could have provided reactive sites to catalyze iron ~nineralization (Fig. 7). Quite possibly some of the cell-bound iron may also have reacted with dissolved silica, forming iron- silicate phases reminiscent of those forined today (Siever and Woodford, 1973; Siever, 1992). Such reactions \vould have kept dissolved silica concentrations in the open ocean below satusa- tion with respect to amo~-pllous silica (Knoll, 1985), although amorphous silica could have precipitated on the ocean floor due to the supersaturated conditions created by the sudden physical and cl~emicnl changes induced through venting. As iron \vas removed from the upwelled waters, the remaining Si-rich waters may have been transportecl onto the flat margins of the microcontinents where supersaturated conditions (via evaporntion) would have allowed benthic microorganisms to induce silicification. That microorganisms may have facilitated [his rorm of min- cralization is evident froill the associations of silicified

~nicrofossils with ancient microbial mats and organic-rich mud. as represented by silicified strolnatolites and carbonaceous cherts (Knoll, 1985). The lack of extant bacteria that actively control silica precipitation (i.e., in the n~anller of diatoms), and the natural and experimental studies that show silicate nlinerals commonly associated with microbial cells (wl~en subject to high dissolved silica) suggest that it is possible to invoke a passive biological inter1)retation for Precambrian cherts, and hence BIFs.

The end of the Archean witnessed significant environmental changes, beginning with the for~nation of several large blocks of continental crust about 3.3-3.1 Ga, increased pllysical and chemical weathering, removal of C 0 2 from the atinosphere, and the subsequent deposition of large carbonate platforms by 2.6 Ga (Lowe, 1994). The shallow sea hosted enormous bacterial mats, which both consumed nlore C 0 2 through photosyntliesis and released O2 to the atmosphere. The depletion of atmospheric C 0 2 by weathering, biological activity, and sediinenta- tion on these new continental bloclcs was accompnnied by the Huronian glaciation 2.4-2.1 Ga ago, \vhich led to the lormation of polar shelf ice, cold water sinking, and inevitably the mixing of ocean waters (Lowe, 1994). As the stratified ocean began to break clown, ferrous iron and silica were actively transported onto the continental slopes and shallow shelves, leading to thick, laterally extensive BIF deposits (Simonson and Hassler, 1996), in waters ranging from 10-700 m (Islcy, 1995). The lateral grading of BlFs into stromatolites or oolites, along with sedimentary

BIFs: HYDROTHERi'vIAL BACTERIAL BIOMINERALIZATION 141

.. .. ..

sea level

.. .. ..FIG. 7.-Schematic diagram of the Archean environment, modified from Lowe (1994). The Archean oceans were strongly stratified with a shallow surface layer con

taining small amounts of oxygen, while the deeper reducing waters were enriched in ferrous iron. Benthic mats grew on the shallow sediments of the microcontincntalshelf (S), while bacterial plankton grew in the open ocean above areas of dynamic upwelling or hydrothermal plumes. At the chemocline (C), ferric hydroxide precipitation was catalyzed by plankton through binding of cationic iron onto their cells; the source of the ferric iron was a combination of photochemical processes. anoxygenicor oxygenic photosynthesis. Thus, all iron was removed from the deep waters before it spread onto the shelf. Upon death, these iron-rich cells accumulated in the bottomsediments (B), where diagcnetic reactions associated with burial led to the transformation of ferric hydroxide into hematite, magnetite, or siderite. Iron deposition mayhave been episodic due to fluctuating nutrient supplies for the overlying plankton. Precipitation of silica occurred on the ocean floor due to pressure/temperature changesafter venting. At the chemocline, some silica may also have reacted with cell-bound iron to form clay-like phases. The remaining dissolved silica was transpo11ed onto theshelf, where silica supersaturation (via evaporation) allowed microorganisms to induce silicification, leading to the formation of silicified stromatolites.

iron, quite possibly due to iron being removed from upwellingdeep waters before it spread onto adjacent continental shelves(Lowe, 1994). This is reflected in Archean shallow waterdeposits that are depleted in iron (Beukes and Klein, 1992).Thus, it seems likely that in the Archean major BIFs would haveformed above oceanic hydrothermal vents or in areas ofupwelling, where ferrous iron and dissolved silica were broughtinto contact with oxygenated water at the chemocline (Beukesand Klein, 1992), yet below the level of the shallow continentalshelf (Lowe, 1994). Plankton growing in the surface waters ofthe open ocean could have provided reactive sites to catalyze ironmineralization (Fig. 7). Quite possibly some of the cell-boundiron may also have reacted with dissolved silica, forming ironsilicate phases reminiscent of those formed today (Siever andWoodford, 1973; Siever, 1992). Such reactions would have keptdissolved silica concentrations in the open ocean below saturation with respect to amorphous silica (Knoll, 1985), althoughamorphous silica could have precipitated on the ocean floor dueto the supersaturated conditions created by the sudden physicaland chemical changes induced through venting. As iron wasremoved from the upwelled waters, the remaining Si-rich watersmay have been transported onto the flat margins of the microcontinents where supersaturated conditions (via evaporation)would have allowed benthic microorganisms to induce silicification. That microorganisms may have facilitated this form of mineralization is evident from the associations of silicified

microfossils with ancient microbial mats and organic-rich mud,as represented by silicified stromatolites and carbonaceous cherts(Knoll, 1985). The lack of extant bacteria that actively controlsilica precipitation (i.e., in the manner of diatoms), and the natural and experimental studies that show silicate minerals commonly associated with microbial cells (when subject to highdissolved silica) suggest that it is possible to invoke a passivebiological interpretation for Precambrian cherts, and hence BIFs.

The end of the Archean witnessed significant environmentalchanges, beginning with the formation of several large blocks ofcontinental crust about 3.3-3.1 Ga, increased physical andchemical weathering, removal of CO2 from the atmosphere, andthe subsequent deposition of large carbonate platforms by 2.6Ga (Lowe, 1994). The shallow sea hosted enormous bacterialmats, which both consumed more CO2 through photosynthesisand released O2 to the atmosphere. The depletion of atmospheric CO2 by weathering, biological activity, and sedimentation on these new continental blocks was accompanied by theHuronian glaciation 2.4-2.1 Ga ago, which led to the formationof polar shelf ice, cold water sinking, and inevitably the mixingof ocean waters (Lowe, 1994). As the stratified ocean began tobreak down, ferrous iron and silica were actively transportedonto the continental slopes and shallow shelves, leading to thick,laterally extensive BIF deposits (Simonson and Hassler, 1996),in waters ranging from 10-700 m (Isley, 1995). The lateral grading of BIFs into stromatolites or oolites, along with sedimentary

evaporation

a plankton sea level

3 + 2 +

Fe ---9 Fe ---+ cell .e-- Fe +---- Fe +--- 0 sudace

sediment

FIG. 8.-Schematic dic~grarii of BIT; deposition in lhe Proterozoic, modified from a genetic model of tlie Haincrsley Group (-2.5 Ga) froill Ivlorris (1993). When the stratified ocenn began to bl-enk do~vn, fel-rous iron aiitl silica were transported onto tlie continental shelf. Up\velling also aupplied nutrients to tlie depository. triggering a l~arallel gt.owtIi of microol.ganisn~s. These bacterial commuiiities subsequelltly served as passive nucleation sites for ferric il.ofi tleposition (as iroil-rich mesobands) by catalyzing reactiolis rendered possible by sul~el-saturated conditio~is tle\,eloped at tlie chemocline (see inset). Within these iron oxide ~nesobands, the alternatilig Fe-rich alitl Si-rich microbn~ids signified seasonal and climatic changes, will1 iron precipitated \\,hell ul~welling currents uecliarged thc tlel>ository with Fc(l1) ant1 nutrients. ant1 a~liorplious silica precipitatcil ilue to evaporation on tlie shelf whcii iron was depleted.

struclures consistent with wave agitation ancl sllallo\v water deposition (Gole and Klein, 1981; Simonson, 1985), also con- fir~lls tliat aleod ode positional depths had decreased. This transition may have been imposed by a transgressive sequence that inhibited siliciclastic input and shifted primary organic matter and carbonate production shoreward (Klein and Beukes, 1989; Beukes and Klein, 1992). The chemocline thus moved from the open oceans onto the continental shelves, thereby I'acilitating BIF cleposition in these relatively shallow waters (Schissel and Aro, 1992; Simonson and Hassler, 1 996).

According to a genetic nlodel for the banded-iron formation of the 1Hamersle)i Group (-2.5 Ga), Morris and Horwitz ( 1 983) and Morris (1 993) proposed that during periods of vigorous ancl sustained mid-ocean riclge activity, \vhich allowed high levels of Fe(I1) lo reach the continental shelf, the combined effects of pl~oto-oxidation (which possibly i~lclt~ded a~toxygenic photosynthesis, Widdel et al., 1993: Ehrenreich and Widclel, 1994) and oxygenic pl~otosyntllesis in the Preca~nbrian oceans caused ferric iron deposition and the formation of iron-rich mcsobancls (Fig. 8). Upwelling also supplied nutrients to the depository, triggering a parallel growth of microorganisms. If, as moclern

hydrotliennal systems suggest, bacteria have an important role as ~iucleatio~l sites for the precipitation of ~ni~lerals (Juniper ancl Tebo, 1995; Konhauser and Ferris, 1996), then it is possible that in the Precambrian, these bacterial communities also catalyzed the mineralization process. Within iron oxide mesobands, the altemaling Fe-rich and Si-rich microbailds fi~rther signified seasonal and clinlatic changes. Iron precipitated when up\vellitlg currents recharged the depository \\lit11 Fe(I1) and ~u~trients, while atnorphous silica precipitated clue to evaporation on the shelf when iron was depleted. Si-levels likely reflected background surface ocean concentrations from a co~liposite of sources i~lcludilig continental weatliering and hydrothermal input. Superimposed on this background deposition were sustained periods of volcanicity which resulted in the Si-rich macrobands gradually returning to the normal Fe-oxide macrobands as volcanic input waned. Once the chemical stralification i11 the oceans broke down corupletely, they became completely mixed and depleted in iron, as is indicated by the general absence of major iron-formations after 1.8 Ga (Beukes and Klein, 1992).

142 KURT 0. KONI-lAUSER

evaporation

Photo-oxidation

sea level

Photosynthesis

o2 +

.-- Fe3 +

--. cell .-- Fesurface

tFe(OH) + (SiO )

3 2

t

3 +

Fe2 +

Fe

sediment

FICi. 8.-Schematic diagram of BIF deposition in the Proterozoic, modified from a genetic model of the Hmnersley Group (-2.5 Ga) from Morris (1993). Whenthe stratified ocean began to break down, ferrous iron and silica were transported onto the continental shelf. Upwelling also supplied nutrients to the depository, triggering a parallel growth of microorganisms. These bacterial communities subsequently served as passive nucleation sites for ferric iron deposition (as iron-rich

mesobands) by catalyzing reactions rendered possible by supersaturated conditions developed at the chemocline (sec inset). Within these iron oxide mesobancls, thealternating Fe-rich and Si-rich microbands signified seasonal and climatic changes, with iron precipitated when upwelling currents recharged the depository with

FeW) and nutrients. and amorphous silica precipitated due to evaporation on the shelf when iron was depleted.

structures consistent with wave agitation and shallow waterdeposition (Gole and Klein, 1981; Simonson, 1985), also confinns that paleodepositional depths had decreased. This transition may have been imposed by a transgressive sequence thatinhibited siliciclastic input and shifted primary organic matterand carbonate production shoreward (Klein and Beukes, ]989;Beukes and Klein, 1992). The chemocline thus moved from theopen oceans onto the continental shelves, thereby facilitatingElF deposition in these relatively shallow waters (Schissel andAro, 1992; Simonson and Hassler, J996).

According to a genetic model for the banded-iron formationof the Hamersley Group (-2.5 Ga), Morris and Horwitz (1983)and Morris (1993) proposed that during periods of vigorous andsustained mid-ocean ridge activity, which allowed high levels ofFe(II) to reach the continental shelf, the combined effects ofphoto-oxidation (which possibly included anoxygenic photosynthesis, Widdel et aI., 1993; Ehrenreich and Widdel, 1994)and oxygenic photosynthesis in the Precambrian oceans causedferric iron deposition and the formation of iron-rich mesobands(Fig. 8). Upwelling also supplied nutrients to the depository,triggering a parallel growth of microorganisms. If, as modern

hydrothermal systems suggest, bacteria have an important roleas nucleation sites for the precipitation of minerals (Juniper andTebo, 1995; Konhauser and Ferris, 1996), then it is possible thatin the Precambrian, these bacterial communities also catalyzedthe mineralization process. Within iron oxide mesobands, thealternating Fe-rich and Si-rich microbands further signified seasonal and climatic changes. Iron precipitated when upwellingcurrents recharged the depository with Fe(II) and nutrients,while amorphous silica precipitated due to evaporation on theshelf when iron was depleted. Si-levels likely reflected background surface ocean concentrations from a composite ofsources including continental weathering and hydrothermalinput. Superimposed on this background deposition were sustained periods of volcanicity which resulted in the Si-rich macrobands gradually returning to the normal Fe-oxide macrobandsas volcanic input waned. Once the chemical stratification in theoceans broke down completely, they became completely mixedand depleted in iron, as is indicated by the general absence ofmajor iron-formations after J.8 Ga (Beukes and Klein, J992).

Funding for parts 01' this worlc was provided by a Royal Society Research grant (574005.G501). I thank Simon Bottrell, Joe Cann, Sarah Gleeson, Rob Raiswell, Sue Bowler, and two anonymous reviewers for their critical cornineilts during the preparation of this manuscript.

REFERENCES

ALT, J. C., 1986, A 1iydrotlic1-liiallbncter~al iron deposit from ;i seamount near 2I0N, East Pacific Rise (ahs.): Gcological Society of Alncrica Abstracts, v. 18, p. 526.

ALT, J . C., 1988, Hydrotherrn:~l oxide and nontronite deposits on seamou~~ts in the eastern Pacific: Marine Geology, \'. 81: 11. 227-239.

A\\JRARIIK, S. M., SCCIOPF, J . \V., AND \\'ALTER: RI. R., 1983, Filamentous fossil bacteria from the i\rcllean of \frcstern Australis: Precambrian Research, v. 20. 12. 357-374.

A\VI~ARIIK, S. M., 1992, The oldest I-ecortls of photosynthesis: Photosynthesis Research, 1,. 33, p. 75-89.

BARGI-IOORN, E. S., AND TYLER, S. A,. 1965, Micl-oorg~uiisms frotin tlie Gunflint cllert: Sciencc, v. 147. p. 563-577.

Bi\l<oss, J. A,, AND D E ~ I I N G , J . \V., 1985. The role of bacteria in the ecology of l~lnck-smolter en~iron~nents, in Joner, k4. I,., etl., Hytlrothertnal vcnts of the Eastern Pacific: An overvie\\': Bulletil~ of the Biological Society of Msliingtot~, v. 6, p. 355-371.

BAROSS, J . A,, LILLEY, M. D.. AND GORDON, L. 1.. 1982, 1s the CH4, ]-I2 and CO velrting from subrliarilie hydrotl~ermal systems ~prodnced by thrrmophilic bacteria?: Naturc, \,. 298, p. 366-368.

BAU, M.. I\ND MOLLER: P., 1993, Rare earth elerl~elit systematics of the chcliii- cally l~recipitatetl coriiponent i l l Enl.ly Prccambrian iron fotortnations rund thc evolution of the terrestrial atmosphere-hycluos],here-litlios~~lierc system: Geocllimicn et Cosmochimica Acta, v. 57, p. 2239-2249.

BAUR, h'l. E., HAYES, J. >I., STUDLEY, S. A,, AND W~\L.TER, M. R.. 1985. Millimeter-scale \lariations of stable isotope abundances in c;~rbonates from banded iron-formations in tlie Haluerslep Grouli of \Vcstern Australia: Economic Geology, v. 80, 11. 270-282.

B~ultes . N. .I., AND KLEIN, C., 1002, Models for irorl-formatiol~ depusition. ill

Scliopf. J. IV., n ~ ~ d Klein. C.. eds., Tlie Proterozoic Biosphere: Cambridge. Carvihridge Uni~ersity Press, 13. 147- 15 1.

BEuKES. N. J . , KLEIN, C., K A U F ~ I A N , A. J., A N D HAYES, J . hf., 1990. Carbonate ~~etrograpliy. kerogen distribution, ant1 carbon and oxygen isotope variations in an carly Proterozoic tra~~sition TI-om limestone to iron-formation depositioll. Transvaal Supergroup, South Africa: Economic Geology, v. 85. 12. 663-690.

BEVERIDGE, T. J., 1988, The l?nctrl.ial surface: Gener:ll consideratio~rs lo\vards dcsign and function: Canatlian Journal of hlic~.obiology, v. 34.13. 363.372.

BE\/ERIDGE, T. J., AND M U R R ~ Y , R. G. E.. 1976, Ul~take ant1 retention of metals by cell walls of Bacillus subtilis: Journal of Bacteriology, \,. 127, p. 1502- 1518.

BE~ERIDGE, T. J., AND FYPE, \V. S., 1985, Metal fixation by bacterial cell walls: Canadian Journt~l of Eartli Sciences. v. 22, 11. 1893-1 898.

But~n:\uhr, S. J.. A N D \ \ 7 ~ ~ ? ~ ~ ~ ~ ~ . J. \V., 1985, Sulfate-reducing bnctcria and silica solubility: A possible mechaaism for evaporite diagcnesis and silica precipitatio~i in ba~ldetl iron forliiatio~ls: Cnriatlian .Iourn;~l of Earth Scie~vces. v. 22, 11. 1904- 1909.

BRATER~IAN, P. S., C-\IRNS-SMITH, A. G., AND SLOI'ER, R. W.; 1983. Plioto-osi- dation of hydrated ~e"-Significance for handed iron formations: Nature. v. 303. 11. 163-164.

BROCK. T. D., SXIITI-~, TI. \V., auu M;\DIGI\N, R4 . T., 1984, Biology of Microorg:wisms (4'" Edition): Et~gle\vood Cliffs, Prentice-Hall. 847 11.

CAIRNS-S~IITH, A. G.. 1978, Prccambrian solution pllotocl~emistry, inverse seg- regari011, and banded iron formntioiis: Nature. \I. 276, 11. 807-808.

CALIBRCIN, E. M., 1983. Genesis of Proterozoic iron-iorniatiot~: Sulpl~ur isotope c\'itlcnce: Geocliiniica el Cosmochimica Acta, v. 47, p. 1069-1074.

C,\~IPBLLL, A. C., P..\LRIER. PI. R.. KLINI<MARIRIEI<. G. P . BO\\TRS. T. S.. E U ~ I O N D , J. M.. LA\VI<ENCE, .I. R . , CASEY, J . F.. 'THOAIPSON, G., HU~II ' I IRIS. S.. RONA, P.,

ANII KARSON. .I. A,, 1988, Cl~emiatry of hot spri~igs on tllc blid-Atlarltic Ridge: Nature. v. 335. p. 514-519.

CASI.ENI-IOLZ, R . W.: 1973, The posslble pliotospntl~etic nse of sulfide by tlie filamentous l~l~ototrnpliic bacteria of hot springs: Limnology ant1 Ocea~~ograpliy. 1,. 18, 11. 863-876.

CLOUII, P. E., Jr., 1965, Significance of the Gunflint (Prec:tmbriaiv) ~nicroflora: Science. \,. 148, p. 27-35.

CLOIID, P., 1973, Paleoecologicel sigllii'icance of tlie banded il-on-formiitio~i: Econcllnic Geology; v. 68. p. 1 135-1 143.

C~or io . P.. AND ldlCA~l, G. R.. 1968, Microhiotas of the banded iron for~nc~tions: Nation;~l Acatlem)~ of Scie~ices P~.oceedings (U~litctl States), v. 61. 11. 779-786.

Cor.~hlaN, bf. L,., HEDRICK, D. B., L.OVLE~: D. R.. WHITE, n. C., AND I'YE. I<., 1993, Retluction of Fc(II1) in setliments by sulpliate-reducing bacteria: N;lturc. \,. 361. p. 436-435.

Conr-~ss, J. B., DYBIOND. J., GORDON, L. I., EDRIOND, J. M., VCIN HEWEN, R. P.. BAI.I.AI?~I, R n., GREEN, K., Wll.Ll,\hl~, D., BAINL~I%II~GE. A,, CKANE, K.. ,\Nil

VAN PINDEI., T. H.. 1970, S~bli~ilririe tlier~~ial springs on the GnlAl~ngos Rift: Science, v. 203, p. 1073-1083.

CORNEI-L. R. hi.. GI~\;ANOI.I, R. , A N D SCIIINI)ILEH, P. \V., 1987, Efftct of silicate species on tlie tr:msformation of fcrriliydrite inlo gocthite and liematitc in alkillinc 111edi;i: Clays and Clay Minerals, 1'. 35. 11. 21-28.

Derclt~, L. A, , I \ND J,\CODSEN, S . B.. 1990, Tlie chemical cvolutiorl of Prccanihrian seawater: Evidence Tl-orn REEs in bal~decl iron formalic~~ia: Geochilnica et Cosmocliimica Acta, v. 54. p. 2965-2977.

DOERIEI.. \V. N.. ANLI BROCK, T. D.. 1977, Structure, growth. ant1 dcco~nposition of laminated algal-bacterial mats in alhnlinc hot springs: Applied ancl Er~\'ironmental Microbiology, \,. 34, 11. 433-452.

DRE\IER. .I. 1.. 1988, Tlie Geochelnistrp of Natural Waters (2"" Edition): Engle\\iood Cliffs, Prentice Hall, 437 11.

Erut~ ICI-!, H. L.. 1990. Gcomicrohiology (2"" Edition): Nc\%, h r k . h.l~~~-cel Dckker. 646 11.

EI+I<I.ICH, H. L., 1996, How n~icrobes ir~flue~lce mineral gro\\"li and tlissol~~tion: Chemical Geology, \I. 132, 11. 5-9.

EHRENREICH. A,, AND WIDDEI., F,? 1994, Al~acrohic oxidation of fe r1 .o~~ iroli by pu~ple bacteria. a new type of phototrophic melabolism: Applied and Ell\~ironmental Ivficrobiology, v. 60. p. 45 17-4526.

E\\TRs, W. E., 1983, Chemical Eactol.s in the dcpnsition and dinsenehis ol' banded iron-formation, in Trel~tlal. A. E, and Morris, R. C.. cds., Bn~~tlcd Iron- Fomiation: Facts and Problems: Amsterdam, Else~~ier. 11,. 491 -5 1 ?.

E\\JCRS, \V. E.. AND MOIIRIS, R . C.. 1981, Stu~lies 0f t11e Dale\ Gorge Mcniber ol' the Brockman Iron Formation. \Vestem Australia: Econolnic Geology, 11. 76. p. 1929-1953.

FERRIS. F. G., 19%'. For~vlatioli of autl~igcnic ~iiinerals by bacteria, in McIntos11, I . M., and Groat, L. A,, etls., Biological-Mineralogical Interac~iv~~s: Ottawa. hlineralogical Association of Canatici Sliol-t Course Series 25. 11. 187-208.

Fcrt~<rs. F. G.. Bsvenrrx~. T. J.. A N D F Y ~ , \rr. S., 1986, 1ro11-silica crystallite nucleation by bacteria in a geothermsl sediment: Nature, v. 320, p. 609-61 I.

FERRIS, F. G., FYFE. W. S., AN[) BEVEL'\II)GE, T. 1.. 1988. Mek\llic ion bindill$ by Bacillus subtilis: Implications toor the fossilization of microorganismy: Geology, \'. 16, 11. 149-152.

Fri t~rs . F. G.. 'T,\z,\I<I, K., .mn Fure. \\I. S.. 1989. 11.on o\ides in acid mine drainage en~iron~iier~ts and tlleir association \\lit11 bacteria: Chcmical Geology. \I. 74, [I. 32 1-330.

FOI~TIN. D., ,\ND BE~EKII IGE, T. J.. 1997. Microbial solfate retluctior~ u'ithin sul- fidic mine tailings: Formation of diagenctic Fe sulfides: Geomicl-obiology Journal. \'. 14,p. 1-21.

FonTlN, D., FERRIS, F, G., AND BEVERIDGE. T. J., 1997, Surface-metliatcd mineral development by bacteria, in Banticld, J . F,. ant1 Nenlson, I<. H.. eds.. Geomicrobiology: 1ntel.actions Bet~iaeen Ivlicrobes a~id Minerals: \V:lshington, D.C., hfinernlogical Society of America. p, 161 -180.

FOlir<Nle~, R. O., 1985. Tlie behavior of silica in 11ydrotherm:~l solutions, ill Berger. B. R., and Betlike. P. M.. eds.. Geology ant1 Gcochcmistrp of Epitliernial Systems: Kevie\\zs in Economic Geology, v. 2. 11. 45-61,

I:OLIRNIER, R. (I., 1989. Geochemistry ;mtl dyrvnmics of the Yello\vsto~~e National Park l~ydrotl~er~nt~l system: i\nnunl Rcvicw of Earth and Pl;inctary Sciences. 1;. 17, 11. 13-53.

FRANCIS. C . , ~ ' I - \RGuI~Is . L.. i \ N I l B,\ROHO(~RN. E. S., 1078, 011 1 1 1 ~ exp~~illlcntal silicification of microorganisms. TI. On the tiiilc of appearance of eu1tar)totic

f]]rs: HYDROTHERMAL BACTERIAL BIOMINtRAUZATJON 143

ACKNOWLEDGlvlENTS

Funding for parts of this work was provided by a RoyalSociety Research grant (574005.G501). I thank Simon Bottrell,Joe Cann, Sarah Gleeson, Rob Raiswell, Sue Bowler, and twoanonymous reviewers for their critical comments during thepreparation of this manuscript.

REFERENCES

ALT, 1. c., 1986, A hydrothermal/baeterial iron deposit from a seamount near2 ION, East Paeifie Rise (abs.): Geological Society of America Abstracts, v. 18,p.526.

ALT, J. c., 1988, Hydrothermal oxide and nontronite deposits on seamounts inthe eastern Pacific: Marine Geology, v. 81, p. 227-239.

AWRAMtK, S. M., SCHOPF, J. W., AND WALTER, M. R., 1983, Filamentons fossilbacteria from the Archean of\Vestern Austra1ia: Precambrian Research, v. 20.p. 357-374.

AWRAMIK, S. M., 1992, The oldest records of photosynthesis: PhotosynthesisResearch, v. 33, p. 75-89.

BARGHOORN, E. S., AND TYLER, S. A., 1965, Microorganisms from the Gunflintchert: Science, v. 147, p. 563-577.

BAtWSS, J. A., AND DgIING, 1. W., 1985, The role of bacteria in the ecology ofblack-smoker environments, in Jones, M. L., ed., Hydrothermal vents of theEastern Pacific: An overview: Bulletin of the Biological Society ofWashington, v. 6, p. 355-371.

BAROSS, J. A, LtLLEY, M. D., AND GORDON, L. 1.,1982, Is the CH4, H2 and COventing from submarine hydrothennal systems produced by thermophilic bacteria?: Nature, v. 298, p. 366-368.

BAU, M., AND MiiLLER, P., 1993, Rare earth element systematics of the chemically precipitated component in Enrly Precambrian iron formations and theevolution of the terrestrial atmosphere-hydrosphere-lithosphere system:Geocl1imica et Cosmochimiea Acta. v. 57. p. 2239-2249.

BAUR, Iv1. E., HAYES, J. 'vI., STUDLEY, S. A, AND WAtTER, M. R., 1985.Millimeter-scale variations of stable isotope abundances in carbonates frombanded iron-formations in the Hamersley Group of \Vestern Australia:Economic Geology, v. 80, p. 270-282.

BEUKES. N. J., AND KLEtN, c., ]992, Models for iron-formation deposltion, inSchopf. 1. W., and Klein, c., eds., The Proterozoic Biosphere: Cambridge.Cambridge University Press, p. 147-151.

BEUKES, N. 1., KLEIN, c., Kt\Uf~IAN, A 1., AND HAYES, J. M., 1990. Carbonatepetrography. kerogen distribution, and carbon and oxygen isotope variationsin an carll' Proterozoic transition fromlimeslOne to iron-formation deposition.Transvaal Supergroup, Soutil Africa: Economic Geology, v. 85, p. 663-690.

BEVERIDGE, T. 1., 1988, The bacterial surface: General considerations towardsdesign and function: Canadian Journal of Microbiology, v. 34. p. 363-372.

BEVERIDGE, T. J., AND MURRAY, R. G. E., 1976. Uptake and retention of metalsby cell walls of Bacillus subtilis: Journal of Bacteriology, v. 127, p. 15021518.

BEVERIDGE, T. J., AND FYFE, W. S., 1985, Metal fixation by bacterial cell walls:Canadian Journal of Earth Sciences, v. 22, p. 1893-]898.

BIRNB,\Uilt, S. 1.. AND WIREMAN. 1. W., 1985, Sulfate-reducing bacteria and silica solubility: A possible mechanism for evaporite diagenesis and silica precipitation in banded iron formations: Canadian Journal of Earth Sciences,v. 22, p. 1904-1909.

BRATERMAN, P. S., C~IRNS-SMITH, A. G., AND SLOPER, R. W., 1983, Photo-oxidation of hydrated Fe2+-Significance for banded iron formations: Nature,v. 303, p. 163-164.

BROCK. T. D., SMtTH, D. W., A'iD IvL\DIG;\N, Iv!. T. 1984, Biology ofMicroorganisms (4'1\ Edition): Englewood Cliffs, Prentice-Hall. 847 p.

CAIRNS-SMITH, A G.. 1978, Precambrian solution photochemistry, inverse segregation, and banded iron formations: Nature, v. 276, p. 807-808.

CAMERON, E. .\1., 1983, Genesis of Proterozoic iron-formation: Sulphur isotopecvidcnce: Geochimica et Cosmochimica Acta. v. 47, p. 1069-1074.

CAMPBELL, A. c.. PALMER, JVI. R.. KUNKHAiltilIER, G. P, BOWERS, T. S.. EDMOND,J. M .. LAWRENCE, .T. R., CASEY, J. F.. THOMPSON, G., HUMPIIRIS, S., RONA, P.,

AND KARSON, 1. A., 1988, Chemistry of hot springs on the Mid-Atlantic Ridge:Nature. v. :B5, p. 514-519.

CASTENHOLZ, R. W., 1973, The possible photosynthetic use of sulfide by tile filamentous phototrophic bacteria of hot springs: Limnology and Oceanography,v. 18, p. 863-876.

CLOUD, P. E., Jr., 1965, Signlileance of the Gunflint (Precambrian) microflora:Science. v. 148, p. 27-35.

CLOUD, P., 1973, Paleoecological significance of the banded iron-formation:Economic Geology, v. 68, p. 1135-11,13.

CLOUD, P.. AND LICARI, G. R., 1968, Microbiotas of the banded iron formations:National Academy of Sciences Proceedings (United States), Y. 61. p. 779-7R6.

COLEMA'i, M. L., HEDRtCK, D. B., LOVLEY, D. R., WHITE, D. c., AND PYE. K.,1993, Reduction of Fe(I1l) in sediments by sulphate-reducing bacteria:Nature. v. 361, p. 436-438.

CORLISS, J. B., DYMOND, J., GORDON, L. 1., EDtvlOND, J. M., VON HERZEN, R. P.,BALLARD, R. D., GREEN, K., \VII.LlA~IS, D., BAINBRIDGE, A., CRANE, K., ANDVAN ANDEL, T. H., 1979, Submarine thermal springs on the Galapagos Rift:Science, v. 203, p. 1073-1083.

CORNELL, R. M., GIOVANOLl, R., AND SClnNDLER, P. W., 1987, Effect of silicatespecies on the transformation of ferrihydrite into goethite and hematite inalkalim; media: Clays and Clay Minerals, v. 35, p. 21-28.

DERRY, L. A.. AND JACOBSEN, S. B., 1990, The chemical evolution ofPrccamhri<lJ1 seawater: Evidence frOll1 REEs in banded iron fonnatiuns:Geochimica et Cosmochimica Acta, v. 54. p. 2965-2977.

DOEMEL, W. N., AND BROCK, T. D., 1977, Structure, growth. and decomjlositionof laminated algal-bacterial mats in alkaline hot springs: Applied andEnvironmental Microbiology, v. 34, p. 433-452.

DREVER, 1. I., 1988, The Geocllemistry of Natural Waters (2'''1 Edition):Englewood Cliffs, Prentice Hall, 437 p.

El1RI ICH, H. L.. 1990. Geomicrobiology (2lld Edition): New York. MarcelDekker, 646 p.

EHRLICH, H. L., 1996, How microbes influence mineral growth and dissolution:Chemical Geology, v. 132, p. 5-9.

EHRENREICH, A, AND WIDDEL, F., 1994, Anaerobic oxidation of ferrous iron bypurple bacteria, a new type of phototrophic metabolism: Applied andEnvironmental ]'vlicrobiology, v. 60, p. 4517-4526.

EWERS, \V. E., 1983, Chemical factors in the dcposition and diagenesis orbanded iron-formation, in Trend'll, A. F., and Morris, R. C, eds., Banded IronFormation: Facts and Problems: Amsterdam, Elsevier, p. 491 -512.

EWERS, W. E.. AND MORRtS, R. c., 1981, Studies of the Dales Gorge Member ofthe Brockman Iron Formation. Western Australia: Economic Geology, v. 76.p. 1929-1953

FERRIS, F. G., 1997, Formatlon of authigenic minerals by bacteria, in IvlcIntosh,1. M., and Groat, L. A., eds., Biological-J'vlincralogical Jnteractions: Ottawa.Ivlineralogical Association of Canada Short Course Series 25, p. 187-208.

FERRIS, F. G.. BEVERIDGE, T. 1.. AND FYFE, W. S., 1986, Iron-silica crystallitenucleation by bacteria in a geothermal sediment: Nature, v. 320, p. 609-611.

FERRIS, F. G., FYfE. W. S., AND BEVERIDGE, T 1.,1988, Meta\lic iOll binding byB,tcillus subtilis: Jmplications tor the fossilization of microorganisms:Geology, v. 16, p. 149-152.

FERRIS, F. G., TAZAKI, K., AND FYFE, W. S.. 1989. Iron oxides in acid minedrainage environments and tbeir association with bacteria: Chemical Geology,v. 74, p. 321-330.

FORTtN, D., AND BEVERIDGE, 1'. L 1997. Microbial sulfate reduction within sulficlic mine tailings: Formation of diagenetic Fe sulfides: GeomicrobiologyJournal. v. 14, p. 1-21.

FORTIN, D., FERRtS, F. G., AND BEVERIDC;E, 1'. 1., 1997, Surface-mediated mineraldevelopment by bacteria, in Banlicld, J. F., and Nealson, K. H., eds.,Geomicrobiology: Interactions Between Microbes and Minerals: Washington,D.C., Mineralogical Society of America. p. 161-180.

FOuRNtER, R. 0., 1985. The behavior of silica in 11ydrothermal solutions, inBerger, B. R., and Bethke, P. M.. eds .. Geology and Geocbemistry ofEpitllermal Systems: Reviews in Economic GeOlogy, v. 2, p. 45-61.

FOURNtER, R. 0., 1989, Geochemistry and dynamics of the Yellowstone NationalPark hydrothermal system: Annual Review of Earth and Planetary Sciences,v. 17, p. 13-53

FRANCIS, S., rVIARGULlS, L., ,\ND BARGHOORN. E. S., 1978, On the experimentalsilicification of microorganisms. II. On the time of appearance of eukaryotic

144 KURT 0. I(OlVH/t USER

organisms in tlie fossil record: Precambrian Research, v. 6 ,p . 65-100. FRAN(.OIS, L. I\i1., 1986, Extensive tleposition of banded iron formations was pos-

sible without photosynthesis: Nature, v. 320, 11. 352-354. G E R ~ I A N , C. It., C A R I P H ~ L L , A. C., AND EI>~IONI). J. M., 1991, Hydrothermal

scavcnging at tlie Mid-Atlantic Ridge: Ivfodification of trace element tlis- solved Sluxes: Earth ant1 Planetary Scicncc Lctters, v. 107, 11. 101-1 14.

GEIIORSE, IV. C., 1984, Biology of iron- and manganese-depositing bacteria: Annual Rcview of IvIicrobiology, \'. 38, 11. 515-550.

GI-IIOI~SE, W. C., A N D EHRLICI-I, M. L., 1992, Microbial biolnineralization of iron and manganese, in Skinner, H. C. IV., and Fitzpatrick, R. W., cds., Biominesalization. Processes of Iron and Manganese: Cremlingen, Catena Supplement 2 1, p. 75-99.

GOLE, M. J.. AND KLDIN, C.. 1981, Bandcd iron-formations tlirougli ~iiucli of Precambrian time: Journal of Gcology, v. 89, p. 169- 183.

G O O D ~ I N , A. M., 1956, Facies relations in the Gunflint iron Foniiation: Economic Geology, v. 51, p. 565-595.

Hi-\NER~, H. V., 1973, Rerente marine Eisenerzc auf Santorin, Griechenland. 11. Bakterogenesc von Eisenliydroxidseclitiienten: Geologische Runtlscliau, v. 62, p. 803-813.

HARMSEN, H. J. M., PRIRUR, D., AND JEAN.I.HON, C., 1997, Distribution of microorganisms in deep-sea Iiytlrotlieniial vcnt cliitiineys investigated by whole-cell hybridization and enrichment culture of ther~liophilic bubpopula- tions: Applied and Environmental Microbiology, v. 63, p. 2876-2883.

HAYES, J. hl. , K,\I~LI-\N, I. R., AND WEDEKING, K.IY., 1983, Precambrian organic geochemistry. preservation of tlie record, i~r Robertson, J. M., ed., Fluid- Mineral Equilibria in Hydrothermal Systems. Volume 1: El Paso, Society of Economic Geologists, p. 9-29.

~IENLEY, R. If1., 1984, Chemical structure of geothermal systems, in Robertson, J. PI., ed., Fluid-Mineral Equilibria in Hydrothermal Systems. Volume I: El Paso, Society of Economic Geologists, p. 9-29.

H I N ~ I A N , N. IV., AND LINI)STROI\I, R. F., 1996, Seasonal changes in silica deposition in liot spring systems: Chemical Geology, v. 132, p. 237-246.

HOL.L .AND, H. D., 1994. Early Pvoterozoic atmosplieric change, in Bengtson, S., etl., Early Life on Eautli: New York, Columbia Univcrsity Press, 11. 237-244.

HOLM, N. G., 1987, Biogenic influences on tlie geoclielnistry of certain feruugi- lio~is sediments of hydrothermal origin: Che~iiical Geology, v. 63, p. 45-57.

H O L ~ I , N. G., 1989, The '"1°C ratios of siderite and organic matter of a motl- ern riletalliferous liydrothemial sedimcnt and their implications for banded iron forliiations: Chemical Geology, v. 77, p. 41-45.

ILER, R. K., 1980, Isolation and characterization of particlc nuclei during the polymcri~ation of silicic acid to colloitlal silica: Journal of Colloid end Interface Science, v. 75, p. 138-148.

ISLEY, A. E., 1995, Hydrothermal plumes and the delivery of iron to bandetl iron formation: Journal of Geology, v. 103, p. 169-185.

JACOBSEN, S. B., AND PI~IENTEL-KLOSE, M. R., 1988, A Nd isotopic study of tlie Ha~iierslep and Michipicoten banded iron Fortnations: The source of REE ant1 Fe in Archean oceans: Earth and Planetary Scicnce Letters, v. 87, 11. 29-44.

J A ~ I E S , H. L., 1954, Sedimentary filcics of iron-formations: Economic Geology, v. 49, p. 235-293.

J~hres, H. L., 1983, Distribution of bandetl iron-formation in space ant1 time, in Trendall, A. F,, ant1 Morris, R. C., eds., Iuon-Forniation: Facts and Probleriis: Amsterdam, Elsevier, p. 471-490.

JANNASCH, H. LV., AND IVIRSEN, C. O., 1981, Morphological survey of microbial ~nats near deep-sea thermal vents: Applied and Environmental Microbiology, v. 41, p. 528-538.

JANNASCI-I, H. LV,, AKD MOTTL, M. J., 1985, Geomicrobiology of deep-sea hytlrotliertiial vents: Science, \'. 229, p. 717-725.

JONES, B., AND RENAU.I; R. W., 1996, Influence of tliermophilic bacteria on calcite and silica precipitation in hot springs with watcr temperatures above 90°C: Evidence fro111 Kenya and New Zealand: Canadian Journal of Earth Sciences, v. 33, p. 72-83.

JONES, B., AND RDNAU?., R. LV., 1997, Formation of silica oncoids around geysers and liot springs at El Tatio, nortliern Chile: Setlimentology, v. 44, p. 287-304.

JONES, B., RDNAUI; R. W., AND ROSEN, M. R., 1997a, Biogenicity of silica prc- cipitation around geysers and hot-spring vents, North Island, New Zealand: Journal of Sedimentary Research, v. 67, p. 88- 104.

JONES, B., I~ENAUT, R. \V,, AND ROSEN, M. R., 199713, Vertical zonation of biota in ~iiicrostromatolites associatetl with liot springs, North Islantl, New Zealand:

Palaios, v. 12, 11. 220-236. JUNIPER. S. I<., A N D FO~JQUUI., Y., 1988, Filamentous iron-silica deposits from

modern and ancient Iiytlrothcrmal sites: Canatlian Mineralogist, v. 26, p. 859- 869.

JUNIPER, S. K., A N D TEHO, B. M., 1995, hjlicrobe-metal interactions ant1 mineral deposition at liytlrotliermal vents. in Karl, D. h4., etl., Tlie Microbiology of Deep-Sea Hydrothermal Vents: Boca Raton, CRC Prcss, 11. 219-253.

JUNIPEI<, S. K., I ~ ~ A R T I N E U , P., S,\RRI\ZIN, J., AND G ~ L I N A S , Y., 1995, Microbial minclxl floc associatetl \\it11 nascent liydrotliermal activity on CoAxial Segment, Juan tle Fuca Ridge: Geophysical Research Lettcrs, v. 22, p. 179- 182.

KARL, D. M., BRITTATN, A. M., I\ND TILHICOOK, B. D., 1989, flydrotliermal ant1 lnicrobial processes at Loihi Seamount, a mid-plate hot-spot volcano: D e e p Sea Research, v. 36, 11. 1655-1673.

KARL, D. M., RIICMURTRY, G. M., MALAHOFF, A,, AND GARCIA, h,I. O., 1988, Loihi Seamount, Hawaii: A mid-plate volcano v\~ith a distincti\~e hydrothermal systern: Nature, v. 335, p. 532-535.

I<ARLIN, R., LYLE, M., AND HEATIT, G. R., 1987, Autliigenic magnctitc formation in suboxic mauinc scdiments: Nature, v. 326, p. 490-493.

KLEIN, C., AND BEUICES, N. J., 1989, Gcocliemistry atid setlimentology of a facies transition from limestone to iron-formation deposition in the early Proterozoic Transvaal Supergroup, South Africa: Economic Geology, v. 84, p. 1733-1774.

KLEIN, C., AND BEIJKES, N. J., 1992, Time distribution, stratigraphy, and scdi- mentologic setting, and geochemistry of Precambrian iron-foniiations, in Schopf, J. W., and Klein, C., etls., Tlie Proterozoic Biosphere: Cambridge, Cambridge University Press, p. 139.146.

KNOLL, A. H., 1985, Exceptional prcservation of photosynthetic organisms in silicified carbonates and silicified peats: Royal Society of London Philosopliicol Transactions, v. B31 1, 11. 1 1 1-122.

KOHI.ER, B., SINGER, A., >\NU STOFERS, P., 1994, Biogenic nontronite fro111 marine white smoker chimneys: Clays and Clay Minerals, v. 42, p. 689-701.

KONHAUSER, K. O., 1997, Bacterial iron biomineraliaation in nature: FEMS Microbiology Revie\vs, v. 20, p. 315-326.

KONIIAUSER, K. O., 1998, Diversity of bacterial iron mineralizotioli: Eartli- Science Revicws, v. 43, p. 9 1 - 12 1.

KONEI~\USER, K. O., AND FERRIS, F. G., 1996, Diversity of iron and silica precipitation by microbial mats in hydrothermal waters, Iceland: Implications for Precambrian iron formations: Geology, v. 24, p. 323-326.

LI\BERGE, G. L., 1973, Possible biological origin of Precambrian iron-foniiations: Economic Geology, v. 68, p. 1098-1 109.

LEES, H., KIVOK, S. C., ANI) SUZUKI, I., 1969, Tlie themiotlpnamics of iron osi- dation by ferrobacilli: Canatlian Journal of Microbiology, v. 15, p. 43-46.

LEO, R. F,, A N D BAI~GHOOI<N, E. S., 1976, Silicification of wood: Harvard University Botanical Museum Leaflets, \'. 25, p. 1-46.

Lov~nu, D. R., 1991, Magnetite formation during ~iiicrobial dissimilatory iron reduction, in Frankel, R. B., arld Blakemore, R. P., eds., Iron Biomincuals: New York, Plenum Prcss, p. I5 1-166.

Lo\\'E. D. R., 1980, Stro~iiatolites 3,400-Myr old from the Arcliean of Western Australia: Nature, v. 284, 11. 441-443.

LO\VE, D. R., 1994, Early environments: Constraints and opport~uiities for early evolution, in Bengtson, S., ed., Early Life on Eartli: Nc\v York, Columbia University Press, 11. 24-35.

MACRAE, I. C., AND ED\\JI\RDS, J. E, 1972, Adsorption of colloidal iron by bacteria: Applicd Microbiology, v. 24, 11. 819.823.

M C N A ~ I A R A , K. J., AND A\\ 'RA~IIK, S. M., 1992, Stromatolites: A key to untler- stantling the early evolution of life: Science in Progress, v. 76, p. 345-364.

MCCONCHIE, D., 1987, The geology and geochemistry of thc Joffre and Whaleback shale members of the Brockman Iron Fortnation, Western Australia, in Appel, P. IV. U., ant1 LaBergc, G. L., eds., Precambrian Iron- Formations: Athens, Thcophrastus, p. 541-597.

MILLERO, F. J., SOTOLONGO, S., AND IZAGUIRRE, M., 1987, Tlie oxidation Itinet- ics of Fe(l1) in seawater: Geocliimica et Cosmocliimica Acta, v. 5 1, p. 703- 801.

MOJZSIS, S. J.,ARRHENIUS, G., R'ICKEEGAN, K. D., H A I ~ I S O N , T. M., N U T ~ I ~ I N , A. P., AND FRIEND, C. R. I,., 1996, Evidence for life on Earth bcfore 3,800 million years ago: Nature, v. 384, p. 55-59.

MORRIS, R. C., 1980, A text~~ral and mincralogical study of tlie relationship of


organisms in the fossil record: Precambrian Research, v. 6, p. 65-100.FRAN,:CJIS, L. M., 1986, Extensive deposition of banded iron formations was pos

sible without photosynthesis: Nature, v. 320, p. 352-354.GERMAN, C R., CAMPBELL, A. C, AND EmIOND, J. M., 1991, Hydrothermal

scavenging at the Mid-Atlantic Ridge: Modification of trace element dissolved fluxes: Earth and Planetary Science Letters, v. 107, p. 101-114.

GHIORSE, W. C., 1984, Biology of iron- and manganese-depositing bacteria:Annual Review of Microbiology, v. 38, p. 515-550.

GHiORSE, W. C, AND EHRLICH, H. L., 1992, Microbial biomineralization of ironand manganese, in Skinner, H. C W., and Fitzpatrick, R. W., cds.,Biomineralization. Processes of Iron and Manganese: Cremlingen, CatenaSupplement 21, p. 75-99.

GOLE, M. J., AND KLEIN, C, 1981, Banded iron-formations through much ofPrecambrian time: Journal of Geology, v. 89, p. 169-183.

GOODWIN, A. M., 1956, Facies relations in the Gunflint iron Formation:Economic Geology, v. 51, p. 565-595.

HANERT, H. V, 1973, Rezente marine Eisenerze auf Santorin, Griechenland. I!.Bakterogenese von Eisenhydroxidsedimenten: Geologische Rundschau, v. 62,p.803-813.

HARMSEN, H. J M., PRIEUR, D., AND JEANTHON, C, 1997, Distribution ofmicroorganisms in deep-sea hydrothermal vent chimneys investigated bywhole-cell hybridization and enrichment culture of thermophilic subpopulations: Applied and Environmental Microbiology, v. 63, p. 2876-2883.

HAYES, J. M., KAPLAN,!. R., AND WEDEKING, KW., 1983, Precambrian organicgeochemistry, preservation of the record, ill Robertson, J M., ed., FluidMineral Equilibria in Hydrothenna1 Systems. Volume 1: EI Paso, Society ofEconomic Geologists, p. 9-29.

HENLEY, R. W., 1984, Chemical structure of geothermal systems, in Robertson,J. M., ed., Fluid-Mineral Equilibria in Hydrothermal Systems. Volume I: EIPaso, Society of Economic Geologists, p. 9-29.

HINMAN, N. W., AND LINDSTROM, R. F, 1996, Seasonal changes in silica deposition in hot spring systems: Chemical Geology, v. 132, p. 237-246.

HOLLAND, H. D., 1994, Early Proterozoic atmospheric change, in Bengtson, S.,ed., Early Life on Earth: New York, Columbia University Press, p. 237-244.

Houd, N. G., 1987, Biogenic influences on the geochemislry of certain ferruginous sediments of hydrothermal origin: Chemical Geology, v. 63, p. 45-57.

HOLM, N. G., 1989, The 13C/12C ratios of siderite and organic matter of a modern metalliferous hydrothermal sediment and their implications for bandediron formations: Chemical Geology, v. 77, p. 41-45.

ILER, R. K, 1980, Isolation and characterization of particle nuclei during thepolymerization of silicic acid to colloidal silica: Journal of Colloid andInterface Science, v. 75, p. 138-148.

ISLEY, A. E., 1995, Hydrothermal plumes and the delivery of iron to banded ironformation: Journal of Geology, v. 103, p. 169-185.

JACOBSEN, S. B., AND PIMENTEL-KLOSE, M. R., 1988, A Nd isotopic study of theHamersley and Michipicoten banded iron formations: The source of REE andFe in Archean oceans: Earth and Planetary Science Letters, v. 87, p. 29-44.

JAMES, H. L., 1954, Sedimentary facies of iron-formations: Economic Geology,v. 49, p. 235-293.

JAMES, H. L., 1983, Distribution of banded iron-formation in space and time, inTrendall, A. F, and Morris, R. C., eds., Iron-Formation: Facts and Problems:Amsterdam, Elsevier, p. 471-490.

JANNASCH, H. W., AND WIRSEN, CO., 1981, Morphological survey of microbialmats near deep-sea thermal vents: Applied and Environmental Microbiology,v. 41, p. 528-538.

JANNASCH, H. W., AND MOTTL, M. J., 1985, Geomicrobiology of deep-seahydrothermal vents: Science, v. 229, p. 717-725.

JONES, B., AND RENAUT, R. \V, 1996, Influence of thermophilic bacteria on calcite and silica precipitation in hot springs with water temperatures above90°C: Evidence from Kenya and New Zealand: Canadian Journal of EarthSciences, v. 33, p. 72-83.

JONES, B., AND RENAUT, R. W., 1997, Formation of silica oncoids around geysersand hot springs at EI Tatio, northern Chile: Sedimentology, v. 44, p. 287-304.

JONES, B., RENAUT, R. W., AND ROSEN, M. R., 1997a, Biogenicity of silica precipitation around geysers and hot-spring vents, North Island, New Zealand:Journal of Sedimentary Research, v. 67, p. 88-104.

JONES, B., RENAUT, R. W., AND ROSEN, M. R., 1997b, Vertical zonation of biotain microstromatolites associated with hot springs, North Island, New Zealand:

Palaios, v. 12, p. 220-236.JUNIPER, S. K, AND FOUQUET, Y, 1988, Filamentous iron-silica deposits from

modern and ancient hydrothermal sites: Canadian Mineralogist, v. 26, p. 859869.

JUNIPER, S. K, AND TEBO, B. lvI., 1995, Microbe-metal interactions and mineraldeposition at hydrothermal vents, in Karl, D. lvI., ed., The Microbiology ofDeep-Sea Hydrothermal Vents: Boca Raton, CRC Press, p. 219-253.

JUNIPER, S. K, lvIARTINEU, P., SARRAZIN, J., AND GELINAS, Y, 1995, Microbialmineral floc associated with nascent hydrothermal activity on CoAxialSegment, Juan de Fuca Ridge: Geophysical Research Letters, v. 22, p. 179182.

KARL, D. M., BRITTAIN, A. M., AND TILBROOK, B. D., 1989, Hydrothermal andmicrobial processes at Loihi Seamount, a mid-plate hot-spot volcano: DeepSea Research, v. 36, p. 1655-1673.

KARL, D. M., McMURTRY, G. M., MALAHOFF, A., AND GARCIA, M. 0., 1988,Loihi Seamount, Hawaii: A mid-plate volcano with a distinctive hydrothermalsystem: Nature, v. 335, p. 532-535.

KARLIN, R., LYLE, lvI., AND HEATH, G. R., 1987, Authigenic magnetite formationin suboxic marine sediments: Nature, v. 326, p. 490-493.

KLEIN, C, AND BEUKES, N. J, 1989, Geochemistry and sedimentology of afacies transition from limestone to iron-formation deposition in the earlyProterozoic Transvaal Supergroup, South Africa: Economic Geology, v. 84,p.1733-1774.

KLEIN, C, AND BEUKES, N. J., 1992, Time distribution, stratigraphy, and sedimentologic setting, and geochemistry of Precambrian iron-formations, inSchopf, J. W., and Klein, C, eds., The Proterozoic Biosphere: Cambridge,Cambridge University Press, p. 139-146.

KNOLL, A. I-l., 1985, Exceptional preservation of photosynthetic organisms insilicified carbonates and silicified peats: Royal Society of LondonPhilosophical Transactions, v. B311, p. 111-122.

KOHLER, B., SINGER, A., AND STOFFERS, P., 1994, Biogenic nontronite frommarine white smoker chimneys: Clays and Clay Minerals, v. 42, p. 689-701.

KONHAUSER, K 0., 1997, Bacterial iron biomineralisation in nature: FEMSMicrobiology Reviews, v. 20, p. 315-326.

KONI-IAUSER, K. 0., 1998, Diversity of bacterial iron mineralization: EarthScience Reviews, v. 43, p. 91-121.

KONHAUSER, K 0., AND FERRIS, F G., 1996, Diversity of iron and silica precipitation by microbial mats in hydrothermal waters, Iceland: Implications forPrecambrian iron formations: Geology, v. 24, p. 323-326.

LABERGE, G. L., 1973, Possible biological origin of Precambrian iron-formations: Economic Geology, v. 68, p. 1098-1109.

LEES, H., KWOK, S. C, AND SUZUKI, I., 1969, The thermodynamics of iron oxidation by ferrobacilli: Canadian Journal of Microbiology, v. 15, p. 43-46.

LEO, R. F, AND BARGHOORN, E. S., 1976, Silicification of wood: HarvardUniversity Botanical Museum Leaflets, v. 25, p. 1-46.

LOVLEY, D. R., 1991, Magnetite formation during microbial dissimilatory ironreduction, in Frankel, R. B., and Blakemore, R. P., eds., Iron Biominerals:New York, Plenum Press, p. 151-166.

LOWE, D. R., 1980, Stromatolites 3,400-Myr old from the Archean of WesternAustralia: Nature, v. 284, p. 441-443.

LOWE, D. R., 1994, Early environments: Constraints and opportunities for earlyevolution, in Bengtson, S., eel., Early Life on Earth: New York, ColumbiaUniversity Press, p. 24-35.

MACRAE, I. C, AND EDWARDS, J F, 1972, Adsorption of colloidal iron by bacteria: Appliedlvlicrobiology, v. 24, p. 819-823.

McNAMARA, K J., AND AWRAMIK, S. M., 1992, Stromatolites: A key to understanding the early evolution of life: Science in Progress, v. 76, p. 345-364.

]'vICCONCHIE, D., 1987, The geology and geochemistry of the Joffre andWhaleback shale members of the Brockman Iron Formation, \VesternAustralia, in Appel, P W. U., and LaBerge, G. L., eds., Precambrian IronFormations: Athens, Theophrastus, p. 541-597.

MILLERO, F J., SOTOLONGO, S., AND IZAGUIRRE, M., 1987, The oxidation kinetics of Fe(II) in seawater: Geochimica et Cosmochimica Acta, v. 51, p. 793801.

MOJZSIS, S. J., ARRHENIUS, G., McKEEGAN, K D., HARRISON, T. M., NUBIAN, A.P, AND FRIEND, C R. L., 1996, Evidence for life on Earth before 3,800 million years ago: Nature, v. 384, p. 55-59.

MORRIS, R. C, 1980, A textural and mineralogical study of the relationship of

BIFs: H YDROTHERIWIL BACTERIAL BIOMIIVERA LIZATIOl\' 145

iron ore to banded iron-forrn;~tion in tlie Hamersley Iron Province of \?restern Austmlia: Economic Geology, v. 75, p. 184-209.

MORRIS, R. C.. 1993. Genetic ~nodellirlg for bandetl iron-formation of thc Hamersley Group, Pilbara Cmton, Western Australia: Precanlbrian Researcli, \I. 60, p. 243-286.

~ ~ O K R I S , R. C.. A N D HORIVITZ, R. C., 1983, Tlie origin of the iron-for~iintion-rich Harilersley Group of Western Austmlia-Deposition on a platrorn~: Precambrian Research, 11. 21.11. 273-397.

NEAI.SON, K. H., AND R~YERS, C. R., 1990, Iron reduction by bacteria: A potential role in the genesis of bandetl iron formations: American Journal of Science, v. 290A, p. 35-45.

OEHLER, J. H., 1976, Experimental studies in Precambrian paleontology: Structural and chcmical changes in bluc-green algae during simulated fossilization in synthetic chert: Geological Society of America Bulletin, v. 87, p. 117-129.

OEHLER, J. H., AND SCHOPF, J. \v., 1971, Artificial microfossils: Experimental studies of pennineralization of blue-green algae in silica: Science, v. 174, p. 1229-1231.

O H ~ ~ ~ J I - O , H., KAKEGA\VI\, l'.. AND L.o\\JE. D. R., 1993, 3.4-billio~i-year-old biogenic pyrites from Barbertoti, South Africa: Sulfur isotope evidence: Science, v. 262, p. 555-557.

PERRY, E. C., JR., TAN, F. C., AND R~OREY, G. B., 1973, Geology and stable isotope geochemistry of the Biwnbik iron formation, northeni Minnesota: Economic Geology, v. 68, p. I 110-1 125.

SAWICKI, J. A,, BRO\\'N. D. A,, AND BEVERIDGE, T. J., 1995, Microbial precipitation of siderite and protoferrihydrite in a biofilm: Canadian Mineralogist, v. 33, p. 1-6.

SCHIDLO\VSI<I, M., 1976, Archean atmosphere and evolutio~i of tlie tel-restrial oxygen budget, in Windley, B. F., ed., Tlie Early History of the Earth: London, John Wiley, p. 525-535.

SCHISSEL, D.. AND AKO, P., 1992, The niajor early Proterozoic iron and inan- ganese tleposits and tlieil. tectonic setting: Economic Geology, \'. 87. p. 1367- 1374.

SCHOPF, J. \V.. AND PACI<ER, B. I\?., 1987, Early Arclican (3.3-billion to 3.5-billion-year-oltl) microfossils from Warrawoona Gsoup, Australia: Science, v. 237, 11. 70-72.

SCHOPF, J. W., 1993, h,licrofossils of tlie early Archean Apex cliert: New evidence of tlie antiquity of lifc: Science, v. 260, p. 610-646.

SCHULTZE-LAM. S., FERRIS, F. G., KONHAUSER, K. 0.. AND WIESE. R. G., 1995. In situ silicification of an Icelandic 1101 spring microbial mat: I~iiplications for niicrofossil for~nation: Canadian Journal of Earth Sciences, v. 32, p. 2021- 2026.

SCH\\JERTRIANN, U., AND FITZPATRICK. R. Mr., 1992, Iron minerals in surface en\+ ronmellts, in Skinner, H. C. \Y., and Fitzpatrick, R. W., eds., Biomineralizatio~i. Processes of Iron and Manganese: Cremlingen, cat en;^

Vel.lag, p. 7-30. SIEI'EI~, R.. 1992, Tlie silica cycle in the Precambrian: Geocliimica et

Cosniochi~u~ca Acta, \I. 56, 11. 3265-3272. SIEVER, R., AND WOODFORD, N., 1973, Sorption of silica by clay minerals:

Geocliimica et Cosmochirnica Acta, v. 37, p. 185 1- 1880. SIGLEO, A. C., 1978, Organic geochemistry of silicified wood, Petrified Forest

Nat~onal Park, Arizona: Geochimica et Cosmoclii~uicn Acta, v. 42, p. 1397- 1405.

S ~ h r o ~ s o ~ , B. M., 1985, Sedimentological constraints on the origins of Precambrian iron-formations: Geological Society of America Bulletin, v. 96, p. 244-252.

SI~IONSON, B. M., AND HASSLER, S. W., 1996, Was the deposition of largc Precambrian iron formations linked to niajor oiarine transgressions?: Journal of Geology, \'. 104, p. 665-676.

STEEI-EL, C. I., ANL) VAN CI~I'ELLEN. P., 1990, A Iie\\i kinetic approach to model-

ing wntcr-rock interaction: The role of nucleation, precursors, and Ost\valcl ripening: Geochimica et Cosmocliimica Acta, \,. 54, p. 7,657-2677.

STRAUB, K. L., BENL, M., SCHINK, B., AND WIDDEL, F,. 1996, Anaerobic, nitratedependent microbial oxidation of ferrous iron: Applied and Environmental Microbiology, v. 62, p. 1458-1460.

TI IO~IPSON, J. B., AND FERIIIS, F. G., 1990, Cpanobacterial precipitation of gypsum. calcite, and magnesite from natural alkaline lake water: Geology, v. 18. 11. 995-998.

TRENDALL, A. F., 1965, Progress report on the Brocltman Iron Formation in the Witte~ioom-'iam],ire area: Western Australia Geolo~ical Survey Annual Report. 1964, p. 55-65.

TRENDALL, A. F., A N D BLOCKLBY, J. G., 1970, Tlie iron formationc of the Precarubrian Hamersley Group, Western Austrnlia: \vitli special reference to tlie associated crocitlolite: \Vestern Australia Geologic;~l Su~vep Bulletin, v. 119, 336p.

TUNNICLIFFE, V., AND FONTAINE. A. R., 1987, Faunal co~nposition and organic surface encrustations at hydrothermal vents on tlie southern Juan tle Fuca Ridge: Journal of Geo~liysical Research, v. 92, 11. 11303-1 1314.

TURCOTTE, D. L., 1980, 011 tlie thermal e\rolution of the earth: Earth and Planetary Science Z.etters. v. 48, p. 53-58.

URRUTIA, h4. M., AND BE\TRIDGE, T. J., 1993, Mechanism of silicate binding to the bacterial cell \+all in Bacillus subtilis: Journal of Bacteiiology, v. 175, p. 1936-1945.

URI?UTIA, M. M., AND BEVERIDGE, T. J., 1994. Furrrlalion of fine-grained ~netal and silicate precipitates on a bacterial surface (Bacillus subtilis): Chemical Geology, v. 116, p. 261-280.

VEIZER, J.. 1983, Geologic evolution of tlie A~.cliea~i-early Proterozoic earth, in Schopf, J. \T', ed., Earth's Earliest Biosphere. Its Origin and E\~olution: Princeton, Princeton University Press, p. 240-259.

Velz~ll, J.. Conlpsl~N. \V., HOEFS, J., AND NIELSEN, H., 1982, Mantle buffering of the early oceans. Nat~~rwissenscl~aften, v. 69, p. 173-180.

\VALKM, J. C. G., 1984. Suboxic diagenesis in banded iron formations: Nature, v. 309, 11. 340-342.

WALKER, J. C. G., KLEIN, C., SCHIULO\VSKI, h4., SCHOPI:, J. W., STEVENSON, D. .I., AND WALTER, M. R., 1983. Erivironmental evolution of the Archean-early Proterozoic eartli, in Schopf, J.W., ed.. Earth's Earliest Biosphere. Its Origin and Evolution: Princeton, Princeton University Press, p. 260-290.

\VALSH, M. M., A N D Lo\\'e, D. R., 1985, Filamentous microfossils from the 3,500-My~old Onver\vacht Group. Barberton Mountain Land. South Africa: Nature, 11. 314, p. 530-532.

~VALTER, hI. R., 1976, Hot-spring sediments in Yello\%~stone National Park, in Walter, h4. R., ed., Stromatolites: Amsterdam, Elsevier, 11. 489-498.

W!\LTER, M. R.. BAULD, J., AND BROCK, T. D., 1972, Siliceous algal and bacterial stromatolites in hot spring and geyser effluents of Yellowstone National Park: Science, \'. 178, p. 402-405.

WALTER, M. R., BUICK, R., AND DUNLOP, J. S. I<., 11)80. Stromatolites 3,400- 3,500 Myr old from the North Pole area, Western Austmlia: Nature. v. 284, 11. 443-445.

\VAT:TER, M. R., BAULD, J., DES MARAIS, D. J., AXD SCHOPF, J. W.. 1992, A general comparison of microbial mats and microbial stro~natolites: Britlging tlie gap between the modern and thc fossil, in Scllupf, J. IV., and Klein, C., eds., Tlie Proterozoic Biosphere: Cambridge, Cambridge Uni\~ersity Press, p. 335- 338.

WESTALL, F., BONI, L., AND GUERZONI, E.. 1995, The experimental silicification of ~iiicroorganisnis: Palaeunlology, v. 38, p. 495-528.

WIDUEL, F., SCFINELL, S., HEISING, S., EHRENREICH,A.: ASSI\IUS, B.. AND SCHINK, B., 1993, Felrous iron oxitlation by anoxygenic pliototropliic bacteria: Nature, v. 362, p. 834-836.

WILLIARIS, L. A,, AND CRERAR, D. A,, 1085, Silica diagenesis, 11. General mech- anisms: Journal of Sedimentary Petrology. \,. 55, p. 312-321.

BIFs: HYDROTHERMAL BACTERIAL BlOlvllNERALIZATlON 145

iron ore to banded iron-formation in the Hamersley Iron Province of WesternAustralia: Economic Geology, v. 75, p. 184-209.

MORRIS, R. C, 1993, Genetic modelling for banded iron-formation of theHamersley Group, Pilbara Craton, Western Australia: Precambrian Research,v. 60, p. 243-286.

MORRIS, R. C, AND HORWITZ, R. C., 1983, The origin of the iron-formation-richHamersley Group of Western Australia-Deposition on a platform:Precambrian Research, v. 21, p. 273-297.

Nr,ALSON, K. H., AND rVlYERS, c. R., 1990, Iron reduction by bacteria: A potential role in the genesis of banded iron formations: American Journal ofScience, v. 290A, p. 35-45.

OEHLER, J. H., 1976, Experimental studies in Precambrian paleontology:Structural and chcmical changes in blue-green algae during simulated fossilization in synthetic chert: Geological Society of America Bulletin, v. 87,p. 117-129.

OEHLER, J. H., AND SCHOPF, J. W., 1971, Artificial microfossils: Experimentalstudies Df permineralization of blue-green algae in silica: Science, v. 174,p. 1229-1231.

OHi\lOTO, H., KAKEGAWA, T., AND LOWE, D. R, 1993, 3.4-billion-year-old biogenic pyrites from Barberton, South Africa: Sulfl1l' isotope evidence: Science,v. 262, p. 555-557.

PERRY, E. C, JR., TAN, F. C, AND MOREY, G. B., 1973, Geology and stable isotope geochemistry of the Biwabik iron formation, northern Minnesota:Economic Geology, v. 68, p. 1110-1125.

SAWICKI, 1. A., BROWN. D. A., AND BEVERIDGE, 1'. J., 1995, Microbial precipitation of siderite and protofeITihydrite in a biofilm: Canadian Mineralogist,v. 33, p. 1-6.

SCHIDLOWSKI, M., 1976, Archean atmosphere and evolution of the terrestrialoxygen budget, in Windley, B. F., ed., The Early History of the Earth: London,John Wiley, p. 525-535.

SCHtSSEL, D.. AND ARO, E, 1992, The major early Proterozoic iron and manganese deposits and their tectonic setting: Economic Geology, v. 87, p. 13671374.

SCHOPF, J. W., AND PACKER, B. M., 1987, Early Archean (3.J-billion to J.5-billion-year-old) microfossils from Warrawoona Group, Australia: Science,v. 237, p. 70-72.

SCHOPF, J. w., 1993, Microfossils of the early Archean Apex chert: New evidence of the antiquity of life: Science, v. 260. p. 640-646.

SCHULTZE-LAM, S., FERRIS, F. G., KONHAUSER, K. 0., AND WIESE, R. G., 1995,In situ silicification of an Icelandic hot spring microbial mat: Implications formicrofossil formation: Canadian Journal of Earth Sciences, v. 32, p. 20212026.

SCHWERTMANN, 0., AND FITZPATRICK, R. W., 1992, Iron minerals in surface environments, in Skinner, H. C. W., and Fitzpatrick, R. W., eds.,Biomineralization. Processes of Iron and Manganese: Cremlingen, CatenaVerlag, p. 7-30.

SIEVER, R.. 1992, The silica cycle in the Precambrian: Geochimica etCosmochlll1ica Acta, v. 56, p. 3265-3272.

SIEVER, R., AND WOODFORD, N., 1973, Sorption of silica by clay minerals:Geochimica et Cosmochimica Acta, v. 37, p. 1851-1880.

SIGLEO, A. C., 1978, Organic geochemistry of silicified wood, Petrified ForestNational Park, Arizona: Geochimica et Cosmochimica Acta, Y. 42, p. 13971405.

SIMOKSON, B. M., 1985, Sedimentological constraints on the origins ofPrecambrian iron-formations: Geological Society of America Bulletin, v. 96,p.244-252.

SIMONSON, B. M., AND HASSLER, S. W., 1996, Was the deposition of largePrecambrian iron formations linked to major marine transgressions?: Journalof Geology, v. 104, p. 665-676.

STEEFEL, C I., AND VAN CAPELLEN, P., 1990, A new kinetic approach to model·

ing water-rock interaction: The role of nucleation, precursors, and Ostwaldripening: Geochimica et Cost11ochimica Acta, v. 54, p. 2657-2677.

STRAUB, K. L., BENZ, M., SCHINK, B., AND WIDDEL, F., 1996, Anaerobic, nitratedependent microbial oxidation of ferrous iron: Applied and EnvironmentalMicrobiology, v. 62, p. 1458-1460.

THOMPSON, J. B., AND FERHIS, F. G., 1990, Cyanobacterial precipitation of gypsum, calcite, and magnesite from natural alkaline lake water: Geology, v. 18,p.995-998.

TRENDALL, A. E, 1965, Progress report on the Brockman Iron Formation in theWittenoom-Yampire area: Western Australia Geological Survey AnnualReport, 1964, p. 55-65.

TRENDALL, A. F., AND BLOCKLEY, J. G., 1970, The iron formations of thePrecambrian Hamersley Group, Western Australia: with special reference tothe associated crocidolite: Western Australia Geological Survey Bulletin, v.119,336 p.

TUNNICLIFFE, V, AND FONTAINE. A. R, 1987, Faunal composition and organicsurface encrustations at Ilydrothermal vents on the southern Juan de FucaRidge: Journal of Geophysical Research, v. 92, p. 11303-11314.

TURCOTTE, D. L., 1980, On the thermal evolution of the earth: Earth andPlanetary Science Letters, v. 48, p. 53-58.

URRUTIA, M. M., AND BEVERIDGE, 1'. 1., 1993, Mechanism of silicate binding tothe bacterial cell wall in Bacillus subtilis: Journal of Bacteriology, v. 175,p. 1936-1945.

URRUTIA, M. M., AND BEVERIDGE, 1'. J., 1994, Formation of fine-grained metaland silicate precipitates on a bacterial surface (Bacillus subtilis): ChemicalGeology, v. 116, p. 261-280.

YElZER, J., 1983, Geologic evolution of the Archean-early Proterozoic eartb, inSchopf, J. w., ed., Earth's Earliest Biosphere. Its Origin and Evolution:Princeton, Princeton University Press, p. 240-259.

VEiZER, J., COMPSTON, W., HOEFS, J., AND NIELSEN, H., 1982, Mantle bufferingof tile early oceans. Naturwissenscl1aften, v. 69, p. 173-180.

WALKER, J. C. G., 1984, Suboxic diagenesis in banded iron formations: Nature,v. 309, p. 340-342.

WALKER, J. C G., KLEIN, C, SCHIDLOWSKI, M., SCHOPF, J. W., STEVENSON, D. J.,AND WALTER, M. R, 1983, Environmental evolution of the Archean-earlyProterozoic earth, in Schopf, J.W., ed., Earth's Earliest Biosphere. Its Originand Evolution: Princeton, Princeton University Press, p. 260-290.

WALSH, M. M., AND LOWE, D. R., 1985, Filamentous microfossils from the3,500-Myr-old Onverwacht Group, Barberton Mountain Land, South Africa:Nature, v. 314, p. 530-532.

WALTER, J\t R., 1976, Hot-spring sediments in Yellowstone National Park, inWalter, M. R., ed., Stromatolites: Amsterdam, Elsevier, p. 489-498.

WALTER, M. R, BAULD, J., AND BROCK, T D., 1972, Siliceous algal and bacterial stromatolites in hot spring amI geyser effluents of Yellowstone NationalPark: Science, v. 178, p. 402-405.

WALTER, M. R., BUICK, R., AND DUNLOP, J. S. R, 1980, Stromatolites 3,4003,500 Myr old from the North Pole area, Western Australia: Nature, v. 284,p.443-445.

WAUI'R, M. R., BAULD, J., DES MARAIS, D. J., AND SCHOPF, J. W., 1992, A general comparison of microbial mats and microbial stromatolites: Bridging thegap between the modern and the fossil, in Schopf, 1. W., and Klein, C, eds.,The Proterozoic Biosphere: Cambridge, Cambridge University Press, p. 335338.

WESTALL, E, BONI, L., AND GUERZONI, E., 1995, The experimental silicificationof microorganisms: Palaeontology, v. 38, p. 495-528.

WIDDEL, F., SCHNELL, S., HEISING, S., EHRENREICH, A., ASSMUS, B.. AND SCHINK,n., 1993, Ferrous iron oxidation by anoxygenic phototrophic bacteria: Nature,v. 362, p. 834-836.

WILLIAMS, L. A., AND CRERAR, D. A, 1985, Silica diagenesis, 11. Generalmechanisms: Journal of Sedimentary Petrology, Y. 55, p. 312-321.

hydrothermal bacterial biomineralization: potential …...ganese-oxidizing bacteria are also locally...

Documents