See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/6000077

The Microbiology of Deep-Sea Hydrothermal Vents

Article  in  Science · August 1985

DOI: 10.1126/science.229.4715.717 · Source: PubMed

CITATIONS

561READS

5,234

2 authors, including:

M. J. Mottl

University of Hawai'i System

485 PUBLICATIONS   8,644 CITATIONS   

SEE PROFILE

All content following this page was uploaded by M. J. Mottl on 20 May 2014.

The user has requested enhancement of the downloaded file.

of pressure and temperature alone limitthe activities of deep-sea organisms.Microorganisms and the food chains thatdepend on them have evolved in re-sponse to the rich supply of reducedcompounds available for chemosynthe-sis. A further constraint is that the ani-mals must continually colonize newvents tens and occasionally hundreds ofkilometers away. Despite differences infaunal composition and at least partialisolation of vent fields, these communi-ties have had a long evolutionary his-tory. Fossil vent worm tubes have beenidentified in Cretaceous sulfide ores (34),and the nearest relative of the archeogas-tropod limpets date from the Paleozoic,over 200 million years ago (9). The co-occurrence of a clam, a mussel, and avestimentiferan worm at widely separat-ed sites in the Pacific and Atlantic repre-sents either an unusual distribution froma single lineage or, even more remark-ably, cases of parallel evolution.

References and Notes1. J. Brooks et al., Eos 66, 106 (1985); Canadian

American Seamount Expedition, Nature (Lon-don) 313, 212 (1985); C. Paull et al., Science 226,965 (1984).

2. R. Hessler and W. Smithey, in HydrothermalProcesses at Seafloor Spreading Centers, P. A.

Rona et al., Eds. (Plenum, New York, 1983),pp. 735-770.

3. Galapagos Biology Expedition Participants,Oceanus 22, 1 (1979); K. Crane and R. Ballard,J. Geophys. Res. 85, 1443 (1980); J. F. Grasslein Hydrothermal Processes at Seafloor Spread-ing Centers, P. A. Rona et al., Eds. (Plenum,New York, 1983), pp. 665-676.

4. D. Desbruyeres, P. Crassous, J. Grassle, A.Khripounoff, D. Reyes, M. Rio, M. Van Praet,C. R. Acad. Sci. 295, 489 (1982).

5. R. Ballard, J. Francheteau, T. Juteau, C. Ran-gan, W. Normark, Earth Planet. Sci. Lett. 55, 1(1981); R. Ballard, T. H. van Andel, R. Hol-comb, J. Geophys. Res. 87, 1149 (1982); R.Ballard, R. Hekinian, J. Francheteau, EarthPlanet. Sci. Lett. 69, 176 (1984).

6. L. Laubier and D. Desbruyeres, La Recherche161, 1506 (1984).

7. R. Turner and R. Lutz, Oceanus 27, 54 (1984);Florida Escarpment Cruise Participants, ibid.,p. 32.

8. J. Grassle, personal communication.9. J. McLean, Bull. Biol. Soc. Wash., in press.10. M. Jones, Oceanus 27, 47 (1984).11. D. Desbruyeres and L. Laubier, Oceanol. Acta

3, 267 (1980); Proc. Biol. Soc. Wash. 95, 484(1982); V. Tunnicliffe, Eos 64, 1017 (1983); per-sonal communication.

12. J. F. Grassle et al., Bull. Biol. Soc. Wash., inpress.

13. J. Enright et al., Nature (London) 289, 219(1981); P. Lonsdale, Deep-Sea Res. 24, 857(1977).

14. C. M. Cavanaugh et al., Science 213, 340 (1981);Nature (London) 302, 58 (1983); H. Felbeck,Science 213, 336 (1981); Nature (London) 293,291 (1981).

15. P. Comita and R. Gagosian, Nature (London)307, 450 (1984).

16. K. L. Smith, Ecology 66, 1067 (1985).17. G. H. Rau, Nature (London) 289, 484 (1981);

and J. I. Hedges, Science 203, 648(1979); G. H. Rau, ibid. 213, 338 (1981); B. Fry,H. Gest, J. Hayes, Nature (London) 306, 51(1983).

18. D. Rhoads et al., J. Mar. Res. 40, 503 (1982).

19. A. Williams, Proc. Biol. Soc. Wash. 93, 443(1980).

20. P. Pugh, Philos. Trans. R. Soc. London 301, 165(1983); K. Woodwick and T. Sensenbaugh,Proc. Biol. Soc. Wash., in press.

21. RISE Project Group, Science 207, 1421 (1980);F. Gaill et al., C. R. Acad. Sci. 298, 553 (1984);ibid., p. 331.

22. D. Cohen and R. Haedrich, Deep-Sea Res. 30,371 (1983).

23. R. Lutz, L. Fritz, D. Rhoads, Eos 64, 1017(1983).

24. K. Macdonald et al., Earth Planet. Sci. Lett. 48,1 (1980).

25. K. Turekian, J. K. Cochran, Y. Nozaki, Nature(London) 280, 385 (1979); K. K. Turekian and J.K. Cochran, Science 214, 909 (1981); , J.Bennett, Nature (London) 303, 55 (1983).

26. C. Berg, Bull. Biol. Soc. Wash., in press.27. Observations made by K. Turekian.28. K. Turekian et al., Proc. Natl. Acad. Sci.

U.S.A. 72, 2829 (1975).29. J. F. Grassle and L. Morse-Porteous, unpub-

lished data; H. Sanders and J. Allen, Bull. Mus.Comp. Zool. Harv. Univ. 145, 237 (1973).

30. R. D. Turner, Science 180, 1377 (1973).31. , personal communication.32. H. Jannasch and C. Taylor, Annu. Rev. Micro-

biol. 38, 487 (1984).33. J. Childress and T. Mickel, Eos 64, 1018 (1983);

J. Childress and R. Mickel, Mar. Biol. Lett. 3,73 (1982); T. Mickel and J. Childress, Biol. Bull.(Woods Hole, Mass.) 162, 70 (1982).

34. R. M. Haymon, R. A. Koski, C. Sinclair, Sci-ence 223, 1407 (1984).

35. I thank the crew of DSRV Alvin and S. Brown-Leger, L. Morse-Porteous, R. Petrecca, and I.Williams for technical assistance; I thank J. P.Grassle, L. Morse-Porteous, and J. M. Petersonfor invaluable comments and assistance in thepreparation of the manuscript. The work wassupported by grants OCE7810458, OCE7810459,OCE8115251, and OCE8311201 from the Na-tional Science Foundation. This is contributionNo. 5963 of the Woods Hole OceanographicInstitution, 71 of the GaI.pagos Rift BiologyExpedition, and 50 of the Oasis Expedition.

Deep-sea hydrothermal vents werediscovered in the 1970's after an exten-sive search along the GalApagos Rift (1,2), a part of the globe-encircling systemof sea-floor spreading axes. During thepast 7 years, more hydrothermal ventfields have been located along the EastPacific Rise. They fall into two maingroups: (i) warm vent fields with maxi-mum exit temperatures of 50 to 23°C andflow rates of 0.5 to 2 cm sec'1 and (ii) hotvent fields with maximum exit tempera-

Holger W. Jannasch is a senior scientist in theBiology Department and Michael J. Mottl, is anassociate scientist in the Chemistry Department atWoods Hole Oceanographic Institution, WoodsHole, Massachusetts 02543.

i2 AUGUST 1985

tures of 2700 to 380°C and flow rates of 1to 2 m sec'1. Hot vent fields commonlyinclude warm- and intermediate-tem-perature vents (0300TC) ("white smok-ers") as well as high-temperature vents(3500 ± 20C) ("black smokers"). A high-ly efficient microbial utilization of geo-thermal energy is apparent at thesesites-rich animal populations werefound to be clustered around these ventsin the virtual absence of a photosyntheticfood source (3-5).

Microorganisms, mainly bacteria, areefficient geochemical agents. As pro-karyotic organisms, they lack a mem-brane-bound nucleus and thereby thecomplex genetic apparatus of the higher,

eukaryotic organisms. At the same time,bacteria retain a much wider metabolicdiversity than is found in plants andanimals. Because of the resulting bio-chemical versatility of natural microbialpopulations and the smallness, generalresistance, and dispersibility of bacterialcells, these organisms are able to exist inmore extreme environments than thehigher organisms. Therefore, the occur-rence of certain microorganisms at deep-sea vents was predictable; however,their ability to make it possible for higherforms of life to thrive with an unusualefficiency on inorganic sources of energyin the absence of light was entirely unex-pected.

Chemosynthesis

The most significant microbial processtaking place at the deep-sea vents is"bacterial chemosynthesis." The termwas coined by Pfeffer in 1897 (6) inobvious contrast to the then well-knownphotosynthesis. Both processes involvethe biosynthesis of organic carbon com-pounds from C02, with the sour'ce ofenergy being either chemical oxidationsor light, respectively. More specifically,

717

Geomicrobiology of Deep-SeaHydrothermal Vents

Holger W. Jannasch and Michael J. Mottl

Table 1. Electron sources and types of chemolithotrophic bacteria potentially occurring athydrothermal vents.

Electron donor Electron Organisms

S2-, SO, S2032 02 Sulfur-oxidizing bacteriaS2-, SO, S2032 N03- Denitrifying and sulfur-oxidizing bacteria

H2 02 Hydrogen-oxidizing bacteriaH2 NO3- Denitrifying hydrogen bacteriaH2 S, S042- Sulfur- and sulfate-reducing bacteriaH2 CO2 Methanogenic and acetogenic bacteria

NH4+, N02 02 Nitrifying bacteriaFe2+, (Mn2+) 02 Iron- and manganese-oxidizing bacteriaCH4, CO 02 Methylotrophic and carbon monoxide-

oxidizing bacteria

chemoautotrophy refers to the assimila-tion of CO2 and is coupled in some

bacteria to chemolithotrophy, the abilityto use certain reduced inorganic com-

pounds as energy sources.In the present-day terminology, the

relation between photosynthetic andchemosynthetic metabolism is illustratedin the following schematic equations,where the reduced carbon is representedas a carbohydrate, [CH2O]:

2C02 + H2S + 2H2 h.-2 [CH20] + H254 (1)

(nonoxygenic photoautolithotrophy,purple and green bacteria)

CO2 + H20 h' 1 [CH2O] + 02 (2)

(oxygenic photoautolithotrophy,green plants)

CO2 + H2S + 02 + H20-[CH2O] + H2SO4 (3)

(aerobic chemoautolithotrophy, bacteria)

2CO2 + 6H2[CH2O] + CH4 + 3H20 (4)

(anaerobic chemoautolithotrophy, bacteria)

From an evolutionary point of view,reactions 1 and 2 above are bridged bythe blue-green or cyanobacteria. In aero-

bic chemosynthesis, the possible elec-tron donors used by a large variety ofbacteria are listed in Table 1. Some ofthem are the same as those used inanaerobic chemosynthesis where freeoxygen is replaced by NO3-, elementalsulfur, s042-, or CO2 as electron accep-tors. The inorganic sources of energy are

used for the production of ATP (adeno-sine 5'-triphosphate), akin to the use oflight in phototrophy.

Differences in the average growthrates of chemolithotrophic bacteria un-

der comparable conditions are deter-mined by the amount of energy requiredfor "reverse electron transfer," a meta-

718

bolic mechanism required for generatingthe necessary negative redox potential.Some organisms have the ability to use

organic compounds simultaneously aselectron sources (mixotrophy). Since inautotrophy carbon (CO2) must be re-

duced from a higher oxidative state thanorganic carbon, more energy is requiredthan in heterotrophy. Therefore, obligatechemoautotrophic bacteria generallygrow more slowly than heterotrophs or

require larger amounts of substrate interms of energy supply.

During recent years a number of newtypes of anaerobic chemoautotrophicbacteria have been isolated and de-scribed. Among them are methanogens,acetogens, and sulfate-reducing bacteria(7). In addition, it has been shown thatcertain extremely thermophilic methano-gens are able to respire elemental sulfur(8, 9). All these metabolic types are

potential catalysts of geochemical trans-formations at deep-sea vents.

All the inorganic energy sources listedin Table 1 have been found in hydrother-mal fluids or in waters surrounding thevents except thiosulfate, the occurrenceof which has not been specifically stud-ied. Before discussing those types ofbacteria that have been isolated andthose microbial processes that have beenshown to occur, we will outline the hy-drothermal origin and the documentedoccurrence of the critical inorganic spe-cies.

Sources of H2S and Structure of the

Mixing Region

The chemistry of the vent waters indi-cates that both warm and hot vent fieldsare fed at depth by a high-temperatureend-member solution at about 350°C andthat the mixing of this solution withlargely unreacted and unheated oceanbottom water in the shallow regions of

the crust is responsible for the widerange of exit temperatures (2, 10). Thus,chemical species that are nonreactiveduring mixing define mixing lines as afunction of temperature for the warmvent waters. These lines pass throughambient seawater and extrapolate to acomposition at 350°C similar to that actu-ally measured in the hot vent waters.Most of the chemical species thought

to participate in microbiological reac-tions do not exhibit such linear mixingbehavior. These species may thereforeoriginate either at depth in the high-temperature end-member solution thathas been produced by reaction of heatedseawater with crustal rocks (11), or theymay originate in the shallow subsea-floorregion, either directly from bottom sea-water or as a result of various inorganicand organic reactions that occur on mix-ing (Fig. 1).The concentrations of relevant species

from the best-studied hot vent fields(those on the East Pacific Rise near21°N) and warm vent fields (those on theGalapagos Rift near 86°W) are shown inTable 2. Also shown are the results oftwo model calculations, the first for aconservative mixture of the hot ventwaters with ocean bottom water and thesecond for the same mixture after somesimplified inorganic reactions have oc-curred.The prominence of H2S is obvious

from Table 2. There are two possiblesources for H2S in the hot vent waters: itmay be leached from crustal basalts, or itmay be produced by reduction of S042-from seawater coupled with oxidation ofFe2+ from basalt to Fe3+. Both mecha-nisms are important in laboratory experi-ments at 300°C and above, but theyoccur only sluggishly or not at all atlower temperatures (12). It is likely thatboth mechanisms are important in thenatural system as well. The concentra-tion of sulfur in typical mid-ocean ridgebasalt (-25 mmol/kg as S2-) is similar tothat in seawater (-28 mmol/kg asS042-), and seawater circulating throughthe hydrothermal system of a mid-oceanridge apparently reacts with an amountoffresh rock about equal to its own mass(10, 13, 14). Although the hot vent wa-ters are essentially free of S042-, circu-lating seawater can be expected to losesome or all of its load of s042- asanhydrite (CaSO4), which precipitateson heating to temperatures as low as130°C (15). Thus, little seawater s042-may be delivered to the deeper, hotterparts of the system where it could bereduced to S2-. Sulfur isotopic analysesof H2S from the hot vent waters and of

SCIENCE, VOL. 229

sulfide minerals from the precipitatedvent chimneys indicate that H2S is de-nved mainly from the basalts, but thatthe seawater source also is important(16).The conservatively calculated H2S

concentration in Table 2 for a 12.6°Cmixture of hot vent water with seawateris in the same range as those in the warmvent waters at the same temperature.H2S undoubtedly is not conservativeduring subsurface mixing, however, asFe2+, 02, and N03- are all heavily de-pleted in the warm vent waters, presum-ably as a result of reaction with H2S.Examination of the relation betweenvent temperature and the concentrationsof species that react on mixing in theshallow subsea-floor provides insightinto the structure of the shallow crustalmixing region and the chemical process-es that occur there. This mixing region,with its large area of basalt surfaces,which serve as substrate, and its dualsource of electron donors from the hotwater end-member and electron accep-tors from seawater, is a major site ofmicrobial production.The generalized relation is shown in

Fig. 2. For a given vent field, 02 andNO3- decrease linearly from their valuesin ocean bottom water to zero at charac-teristic temperatures <20'C that varyfrom one vent field to another (Table 3).H2S decreases linearly with decreasingtemperature as 02 and N03- increase,generally going to zero at the bottom-water temperature of 2°C. An inflectiontypically occurs in the H2S-temperaturerelation where 02 goes to zero, with theslope of the H2S temperature curve be-coming steeper at higher temperatures.Other species whose concentrations de-crease from their seawater values andextrapolate to zero at temperatures -<20to 30°C in the warm vent waters arechromium, uranium, nickel, copper, cad-mium, and selenium (10).Thus distinct zones exist in the shal-

low subsea-floor mixing region that arecharacterized by particular redox condi-tions; in some cases the boundaries be-tween these zones are abrupt and iso-thermal (Fig. 2). Edmond et al. (10) haveinferred that a shallow subsurface reser-voir at 100 to 32°C is located beneath thewarm vent fields and is being tapped bythe vents. The temperature range of thisreservoir is defined by the lowest tem-peratures at which specific chemical pro-cesses occur. The inferred processes arelisted in Table 3. The minimum tempera-tures at which sulfide deposition occursin the subsea-floor reservoir are those atwhich nickel, copper, cadmium, and se-

23 AUGUST 1985

lenium go to zero; at lower tempera-tures, these species are apparently un-reactive and thus define mixing lineswith ambient seawater.The NO3- concentration extrapolates

to zero at a temperature just slightly

lower than the sulfide-related elementsand slightly above the highest tempera-ture sampled; thus the reservoir is free ofN03- and is anoxic because of the reac-tion of these species from seawater withH2S and other reduced species from the

Summary. During the cycling of seawater through the earth's crust along the mid-ocean ridge system, geothermal energy is transferred into chemical energy in theform of reduced inorganic compounds. These compounds are derived from thereaction of seawater with crustal rocks at high temperatures and are emitted fromwarm (<25°C) and hot (-350°C) submarine vents at depths of 2000 to 3000 meters.Chemolithotrophic bacteria use these reduced chemical species as sources of energyfor the reduction of carbon dioxide (assimilation) to organic carbon. These bacteriaform the base of the food chain, which permits copious populations of certainspecifically adapted invertebrates to grow in the immediate vicinity of the vents. Suchhighly prolific, although narrowly localized, deep-sea communities are thus main-tained primarily by terrestrial rather than by solar energy. Reduced sulfur compoundsappear to represent the major electron donors for aerobic microbial metabolism, butmethane-, hydrogen-, iron-, and manganese-oxidizing bacteria have also been found.Methanogenic, sulfur-respiring, and extremely thermophilic isolates carry out anaero-bic chemosynthesis. Bacteria grow most abundantly in the shallow crust whereupwelling hot, reducing hydrothermal fluid mixes with downwelling cold, oxygenatedseawater. The predominant production of biomass, however, is the result of symbioticassociations between chemolithotrophic bacteria and certain invertebrates, whichhave also been found as fossils in Cretaceous sulfide ores of ophiolite deposits.

Fig. 1. Schematic diagram showing inorganic chemical processes occurring at warm- and hot-water vent sites. Deeply circulating seawater is heated to 3500 to 400°C and reacts with crustalbasalts, leaching various species into solution. The hot water rises, reaching the sea floordirectly in some places and mixing first with cold, downwelling seawater in others. On mixing,iron-copper-zinc sulfide minerals and anhydrite precipitate. Modified from Jannasch and Taylor(54).

719

hot-water end-member. The 02 concen-tration goes to zero at a temperature 10 to11°C lower than N03 , depending on thevent field (Table 3).These observations are best explained

in terms of two distinct zones that are

shallower than the reservoir itself, inwhich the residence time of the mixedwaters is short relative to the rate ofreduction of N03- or NO2- and 02,

respectively. The warmer zone probablyconsists of the channels that connect thereservoir to the sea floor, in which 02 iSreduced completely but N03- is largelynonreactive. Both N03- and H2S coex-ist in this zone (Fig. 2), which was fre-quently sampled directly. The coolerzone probably consists of the throats ofthe vents themselves, in which the resi-dence time is so short that all species mixconservatively; H2S, 02, and NO03 co-exist in this zone.

Samples from several vents within a

single vent field define single mixinglines for reactive species. This impliesthat the temperatures that bound thevarious zones are uniform across thearea of an individual vent field. Variationin these characteristic temperatures fromone field to another (Table 3) may reflectto some extent the variations in composi-tion of the hot-water end-member feed-ing the various fields. Probably, howev-er, this variation is mainly a function ofthe shallow crustal channel geometryand the distribution of permeability andrecharge rates of seawater to the subsea-

0 4 8 12

Temperature (OC)16

Fig. 2. Relation between temperature and theconcentrations of 02, NO3-, and H2S definedby samples from individual vents in a singlewarm vent field. The specific values shownvary from field to field, but the topology istypical and has been generalized from theseven fields on the Galapagos Rift listed inTable 2. Labels refer to inferred zones withinthe subsea-ftoor mixing region.

floor reservoir. The uniformity of thecharacteristic temperatures for differentvents within a single vent field reinforcesthe notion of a subsurface reservoir cre-

ated by permeability variations in theshallow subsea-floor.

Because the inferred reservoir isanoxic, like the water in the surficialupflow channels, aerobic chemosynthet-ic microorganisms probably thrive main-ly at the margins of these zones, wheredownwelling oxygenated seawater mixeswith the major bodies of already mixedand reacted solutions. Electron-donorspecies from the reservoir would beavailable at these sites.

Sources of Other Chemical Species

Used in Chemosynthesis

In addition to H2S, the subsea-floorreservoir contains H2 (17), although inmuch lower concentrations than wouldbe expected from the high values in thehot-water end-member (Table 2). Whenseawater was reacted with basalts inlaboratory experiments (18), the result-ant concentration of H2 was lower thanthat in the natural 350°C solutions. Itapparently was controlled by the redoxstate, which was near the magnetite-hematite boundary at 3500 to 375°C. TheH2-02 redox couple approached equilib-rium faster than any other redox couple.Isotopic data on H2 from the hot ventwaters also suggest a close approach toequilibrium for H2-H20 (12). Inorganicreaction of H2 with 02 from seawater torelatively low temperatures during mix-ing could easily account for the relativelylow H2 concentrations in the warm ventwaters, which may then have been af-fected by bacterial reactions.

In contrast to H2, CH4 and CO are

present at much higher concentrations inthe warm vent waters than would beexpected from the concentrations in thehot vent water (Table 2). CH4 in the hotvents is almost certainly abiogenic, onthe basis of its similar concentration infresh basalts and its relatively heavyisotopic composition (19), although in-terpretation of the isotopic data has beenquestioned (17). No isotopic data are

Table 2. Comparison of the compositions of actual warm vent water at several vent fields with those calculated as mixtures of hot (350°C) ventwater with seawater (concentrations in micromoles per kilogram of solution). Vent fields: Ocean Bottom Seismograph (OBS), Southwest (SW),and Hanging Gardens (HG), all on the East Pacific Rise near 21°N; Clambake (CB), Garden ofEden (GE), Dandelions (DL), and Oyster Bed (OB)sampled in 1977, plus Mussel Bed (MB), East of Eden (EE), and Rose Garden (RG) sampled in 1979 on the Galipagos Rift near 86°W.

End-members Mixtures at 12.60C350°C waters

Compo- conservative position: reacted* 12.6°C = 800 M Si)

OBS SW HG OBS SW HG OBS SW HG CB GE DL OB MB EE RG Refer-ence

Vent watersIH2S 7300 7450 8370 219 224 251 105 127 102 20 120 260 50 500 180 (10, 11, 17)H2 1700 380 51 11 0 0 0 0.12 0.03 (17, 19)CH4 45 53 1.4 1.6 3.1 8.8 2.4 (17, 19)CO 0.31 0.67 0.0009 0.02 0.06 0.39 0.06 (17)NH3 <10 <10 <10 <0.3 <0.3 <0.3 4 4 4 2.8 0.0 2.8 0.0 (11, 17)NO2- <0.1 <0.003 2 2 2 1.3 8.2 (11, 17)N20 <0.02 0.06 <0.0006 0.0002 0.13 0.00 0.03 (17)Fe2+ 1664 750 2429 50 23 73 0 0 0 30 0.5 -3 <1 (10, 11)Mn2+ 960 699 878 29 21 26 42 11 16 15 (10, 11)£CO2 5720 2400 2510 2580 (13)

Ambient seawater02 107 104 0 0 0 0 0 0 0 0 0 0 (10, 17)N03 39 38 0 0 0 0 1.0 0 0 0 11.5 (10, 17)XCO2 2300 (11)

Temperaturet (°C)2.2 9.3 12.1 6.6 9.1 10.0 5.6 16.4 (10, 17)

*The reaction sequence used is: 2H2 + 02- 2H20; Fe2+ + H;S -* FeS + 2H+; 49H S + 76NO3- + 8H20-. 32N2 + 8NH4+ + 4NO2- + 49S42- + 18H+ + 32H20;H2S + 202 - So042- + 2H+. tTemperatures are for ambient seawater and for the highest temperature sample used in this data compilation.

720 SCIENCE, VOL. 229

Table 3. Temperatures at which the concentrations of various species in seawater decrease to zero in warm vent fields on the Galapagos Rift near86°W (10, 17). The temperature at which N03 goes to zero is considered to be the best estimate for the isotherm that bounds the shallow subsea-floor reservoir underlying the various vent fields. See Table 2 for the names of the fields.

Temper- Vent fieldSpecies Inferred locality and process ature

(OC) RG GE MB CB DL OB EE

Cr Reservoir: CrlV042-__ Cr2111O3 -32U Reservoir: UVIO2(CO3)34-_+ UIVO2 -24Se Reservoir: substitution in sulfide -20

mineralsNi, Cu, Cd Reservoir: precipitation as or in sulfide 12.5-13.5 14.4-24.2 9.9-12.2

mineralsN03- Reservoir: reduction by H2S 18.0 12.6 12.4 10.7 9.9 8.602 Reservoir and upflow channels: 6.9 9.5 9.2 9.3 6.5 6.3 3.7

reduction by H2SMaximum temperature sampled 16.4 12.7 10.0 9.3 6.6 9.9 5.6

available for CH4 or CO from the warmvents, but the anomalously high concen-trations of these two species could wellindicate a primarily biological origin,probably in the anoxic subsea-floor res-ervoir. As with N03-, CH4 behaves lin-early with temperature over the entireinterval sampled (17), indicating that,unlike H2S and 02, it is conserved in theinferred channels to the sea-floor. In atleast one warm vent field (Rose Garden),CO apparently is produced in the upflowchannels, as indicated by its inflectionpoint and slope when plotted againsttemperature.The reservoir also contains Fe2' and

Mn21 in substantial concentrations, de-rived by leaching from basalt at hightemperature. Mn2+ plots linearly againsttemperature over the entire interval sam-pled for the warm vents (10), and theselines extrapolate to concentrations simi-lar to those in the 350°C end-member(Table 2). Thus, Mn2+ is largely nonreac-tive in the shallow subsea-floor. Fe2+, bycontrast, is nonlinear over the sampledinterval in the same sense as H2S (Fig.2); thus it is being removed from solutionin the upflow channels as well as in thereservoir, probably by a combination ofsulfide and oxide deposition. It is uncer-tain to what extent Fe2+ is utilized inmicrobiological reactions, as it readilyparticipates in inorganic reactions underthese conditions.

Other electron donors present in thesubsea-floor reservoir do not originatemainly from the hot-water end-member.NH4+ and NO2- were at or below detec-tion limits in the 350°C solutions butwere readily measurable in the warmvent waters (Table 2). They almost cer-tainly derive from reduction of seawaterN03- introduced into the reservoir, byreaction mainly with H2S. Also presentat very low concentrations is N20 (17).These species together account for lessthan 20 percent of the introduced NO37;most of the rest is presumably reduced to

23 AUGUST 1985

N2. NH4' and NO2- behave linearlyversus temperature over the entire inter-val sampled for some warm vent fields(for example, Clambake); for others,however (NO2- in Oyster Beds), theydisplay inflection points indicating theirconsumption in the upflow channels.Thiosulfate has not been sought, butelemental sulfur has been detected inwarm vent effluent as well as in thechimneys of black smokers and whitesmokers. The slopes of plots of H2Sversus temperature for those warm ventsamples that are free of 02 suggest thatsulfur species with intermediate oxida-tion states are being formed on mixing aswell as so42-, although so42- is usuallydominant. Seawater also contributesS042- directly to the subsea-floor reser-voir.Among the electron acceptors, CO2 is

paramount. This species is highly en-riched in the hot vent water by theleaching of CO2 from basalt (19, 20). Itsconcentration in the warm vent waters isabout what it should be if the behavior ofCO2 on mixing is conservative (Table 2).

Microbial Populations of

Emitted Vent Waters

Without considering their specific cat-alytic function, one can assess abun-dance of natural bacterial populations bydetermining cell concentrations or bymeasuring growth rates using unspecifictracers. The milky-bluish waters (Fig.3A) flowing from some of the warmvents (60 to 23°C, 1 to 2 cm sec- 1)contain between 105 and 109 cells permilliliter (2, 4, 5). Independent of thetemperatures measured, the large rangeof numbers is due to the dilution of ventwater at the point of sampling. Visiblebacterial aggregates add to this heteroge-neity and may represent dislodged piecesof microbial mats (4, 5). When contami-nation by ambient water was strictly

prevented, we were unable to find signif-icant numbers of microscopically visiblebacteria in hot (3380 to 350°C) vent wa-ter. In contrast, 4.7 x 105 cells werecounted in vent water at 304°C (21) whenthe temperature was determined frommagnesium concentrations (22). Thisfinding indicated an unspecified amountof seawater intrusion prior to or duringsampling.

Since aerobic chemosynthesis resultsin higher productivity than anaerobicchemosynthesis, the availability of theelectron donor and oxygen under favor-able growth conditions will be decisive.From this point of view, bacterial pro-ductivity should be highest in the vicinityof warm vents where the slow emissionof sources of reduced chemical energyinto oxygenated seawater forms slowlymoving plumes. In contrast, the forcefulemission of hydrothermal fluid from thehot vents results in a quick dispersal andfast dilution of energy sources in thewater column, eventually leading tochemical oxidations. The observation ofmaximum populations of animals in theimmediate vicinity of warm vent plumesand heavy bacterial mats near warmleakages at the base of hot vent chim-neys supports these assumptions.Biomass measurements can also be

based on determinations of adenosinetriphosphate (ATP) or total adenylates(22). Data of Karl et al. (5) demonstratethat the microbial biomass of warm ventplumes, determined as ATP, was two tothree times that of the photosynthetic-heterotrophic microbial populations ofsurface waters at the same site (Galapa-gos Rift). The ratio of guanosine 5'-triphosphate to ATP, also measured inthis study (5), has been interpreted as anindicator of growth rates. It correlatedwell with the data derived from biomassdeterminations (5).The most recent developments in the

measurement of growth rates of naturalmicrobial populations are based on the

721

use of tritiated nucleotides (adenine orthymidine) for incorporation into RNAand DNA (23). It is assumed that theassimilation of these marker substratesdoes not affect growth by stimulatingATP production. In a recent study withsamples collected from a hot smokerorifice, higher adenine incorporationrates were found at 90°C than at 210 and50°C (24).

In addition to their occurrence inwarm vent water plumes (Fig. 3A), largemicrobial populations are also found (i)as mats covering almost indiscriminatelyall surfaces exposed to warm ventplumes (Fig. 3, B and C) and (ii) insymbiotic tissues within certain vent in-vertebrates (see below). Quantitativedata on microbial activities at these twosites have not yet been obtained.

Sulfur-Oxidizing Bacteria andRates of Chemosynthesis

The predominant chemosyntheticallyusable chemical energy at the vents ap-pears in the form of sulfur compounds.This predominance is reflected in theease and success with which sulfur-oxi-dizing bacteria can be isolated (25). Ingeneral, the types of sulfur bacteriafound at the deep-sea vents do not differgreatly from those isolated from otherH2S-rich environments. There is one ex-

ception to this rule: the common occur-rence of the genus Thiobacillus appearsto be replaced by a prevalence of thegenus Thiomicrospira (25).

Pure-culture isolations resulted in awide range of metabolic types of sulfurbacteria including acidophilic obligatechemoautotrophs, mixotrophs (whichsimultaneously assimilate inorganic andorganic carbon), and facultative chemo-autotrophs (25). Since the presence oforganic carbon can be expected to bewidespread within the vent communi-ties, the facultative chemoautotrophsmay well represent the predominant typeof sulfur bacteria. The demonstrated ex-cretion of organic carbon by obligatechemoautotrophs indicates the possibleoccurrence of these bacteria even in thesubsurface vent systems (25). The pref-erence for a neutral pH range favorsthe facultative (polythionate-producing)chemoautotrophs in the well-bufferedseawater environment (26). This bio-chemical versatility of sulfur bacteria,together with the relatively high concen-trations of reduced sulfur compounds,appears to be the key to their predom-inance at the vents and to their role asprimary chemosynthetic producers com-pared to the other types of chemolitho-autotrophic bacteria.As in the measurement of photosyn-

thesis, 14CO2 was used as a substrate todetermine rates of chemosynthesis. With

the aid of the research submersible Al-vin, arrays of six 200-ml syringes werefilled in situ from a joint inlet (27). Theyfacilitated replica and control samplingsand were used for in situ incubationexperiments (Fig. 4). At the base of the21°N black smoker, the in situ rate ofCO2 incorporation by natural microbialpopulations in warm water leakages wasapproximately 10-6 pLM ml-' day-' (27).When parallel samples were incubated inthe ship's laboratory (atmospheric pres-sure) at 3°C, the rate was virtually thesame (indicating a minimal effect of hy-drostatic pressure). This result was cor-roborated by data on the metabolic ratesof a pure culture isolate (Thiomicrospira,strain L-12) as affected by pressure (24).

In a second shipboard incubation at23°C, the in situ temperature of thewarm-water leakages, the rate of CO2incorporation increased one and a halforders of magnitude (27). This behaviorindicates the "mesophilic" growth char-acteristic of the total natural population.A similar response was found in purecultures. An addition of 1mM thiosulfateas an accessory energy source in allthree experiments resulted in substantialrate increases. This immediate use ofreduced sulfur confirmed the predom-inance of sulfur-oxidizing bacteria in thenatural population (27).

Different types of dense bacterial matshave been observed at various vents

Fig. 3 (left). (A) Filtrate (bacterial cells) of turbid water emitted from awarm vent (Mussel Bed, Galpagos Rift vent site, 21°C) on aNuclepore filter (pore size, 0.2 ,um; scale bar, 1 ,um). (B) Scanningelectron micrograph of a microbial mat grown within a warm ventplume: a, Beggiatoa; b, Thiothrix-like filaments; c, stalks of prosthe-cate cells (Hyphomonas) (scale bar, 5 ,uam). (C) Transmission electronmicrograph of the outer surface of a hot vent chimney: a, Methyloloc-cus-type cells with an internal membrane structure; b, mostly thick-walled Thiothrix-like filaments embedded in ferromanganese deposits(scale bar, I R,m). Fig. 4 (right). At left, syringe array for the insitu incubation of six parallel vent water samples to measure rates of14CO2 assimilation; at right, Riftia pachyptila, most of the tubes withgills extended; tube length, 0.5 to 1.0 m. Location: warm vent at thebase of a black smoker at 21°N, 109°W; depth, 2610 m; Alvin dive1223. [Photo by D. M. Karl]

SCIENCE, VOL. 229722

(28). The genera Thiothrix and Beggiatoaappear to be predominant according tomorphological criteria. During prepara-tions for the isolation of these organisms,the capacities of marine Beggiatoa forthe fixation of N2 and for facultativechemolithoautotrophy have been dem-onstrated (29). Whitish microbial matsand streamers were commonly observedat the base of hot vents. They representsites of substantial chemosynthetic pro-duction and active grazing by a variety ofinvertebrates.Thick mats of Beggiatoa-like fila-

ments, partly floating above the bottom,were observed in situ at exploratorydives at the Guaymas Basin vent site(2000 m deep) in the Gulf of California(30). Collected and fixed specimensshowed a filament width of up to 100 ,um.At this site, hot vents are overlayed byabout 200 m of sediment. A substantialinput of photosynthetically produced or-ganic matter from the water column tothe sediments further distinguishes thissite from all others studied so far. Highconcentrations of NH3 (-4 mM) havealso been reported (31), suggesting che-mosynthesis by nitrification. A majorgeochemical-biological study of this siteis planned for mid-1985.

Microbial CH4 Oxidation

Next to reduced sulfur, CH4 may be asubstantial source of energy for chemo-synthesis at those deep sea vents whereit has been reported to be present inconsiderable quantities. Although quan-titatively less abundant than H2 in thehigh-temperature vents, CH4 is moreabundant in the warm vents (Table 2).Evidence for its microbial oxidation is,at this time, stronger than that for H2oxidation.Methanotrophic bacteria are included

in the disparate group of the methylo-trophic microorganisms, which compriseall those metabolic types that metabolizeCl compounds (32). CH4 may serve asthe source of both energy and carbon(2CH4 + 202 -* 2[CH2OJ + 2H20), butCO2 may be incorporated as well. Allmethanotrophs are strictly aerobic, of-ten microaerophilic (33), Gram-negativerods, cocci, or vibrios and are character-ized by typical intracellular membranestructures. Methane-utilizing bacteriamay also co-oxidize the CO that mayoccur in vent water (17), without gainingenergy in the form of cell carbon throughenzymes that normally catalyze otherprocesses (34).

Microbial CH4 oxidation at the ventswas first suggested when the typical23 AUGUST 1985

morphological characteristics were ob-served in transmission electron micro-graphs from bacterial mats (Fig. 3C) (28).Up to 20 percent of the cells surveyed insections of mats collected from variousparts of the vents showed the pairedvesicular membranes that distinguishmethanotrophic cells from similar struc-tures found in ammonium oxidizers.Both CH4- and methylamine-oxidizingbacteria were successfully isolated frommicrobial mats, filtered vent water, clamgill tissue, and Riftia trophosome (seebelow), and the pure cultures obtainedwere preliminarily grouped as type Imethanotrophs (33).

Hydrogen as a Microbial

Source of Energy

Many different types of microorga-nisms oxidize H2, but only a few are ableto use the energy gained for the fixationof CO2 and can be described as chemo-lithoautotrophs (Table 1). Within thisgroup the term "H2 bacteria" is usedonly for aerobic organisms. Formerlygrouped in the genus Hydrogenomonas,the aerobic H2-oxidizing bacteria arespread over many known genera (35). Allof them are facultative autotrophs. Assuch, they possess ecological advantagessimilar to those for the facultativelyautotrophic sulfur-oxidizing bacteria.They combine the properties of hetero-trophic growth with the use of the Calvincycle enzymes. The net equation forautotrophic growth is 6H2 + 202 + CO2

[CH20] + 5H20.Little is known about the ecology of

aerobic hydrogen bacteria except thattheir occurrence in nature is as wide-spread as that of biological H2-producingprocesses. As in the case of sulfide oxi-dizers, the chemosynthetic use of geo-thermally produced H2 at the vents rep-resents a primary production of organiccarbon. No specific study of aerobichydrogen bacteria at the vents has yetbeen undertaken. An organism with astrong growth stimulation by H2 wasisolated incidentally from a Riftia tro-phosome sample (36).Anaerobic hydrogen-oxidizing bacte-

ria are known as methanogens and aceto-gens because of their products (Table 1).They are commonly found at anoxicniches where CO2 and H2 are present asthe result of fermentation. In hydrother-mal fluid both compounds are producedgeothermally. The production of CH4,H2, and CO was observed experimental-ly at about 100°C in certain media inocu-lated with samples of black smoker wa-ter (37).

An extremely thermophilic methano-gen of the genus Methanococcus wasisolated from the base of the 21°N blacksmoker (Fig. 4) (38). This organismshowed an optimal growth rate of 0.036hour' (a doubling time of 28 minutes) at86°C. These results demonstrate the ex-istence of a potential biological CH4 pro-duction at the vents. The absence ofisotopic evidence in support of thisobservation is not necessarily conclusivebecause of microbial patchiness.Although denitrifying H2 oxidizers

may exist in vent systems wherever theNO3-containing bottom seawater mixeswith rising hydrothermal fluid, theSo42-_ and sulfur-reducing equivalentsare geochemically more significant. Bothmetabolic types of bacteria do exist buthave not yet been isolated from ventwaters. The respiration of elemental sul-fur has recently been demonstrated to bea common property of extremely ther-mophilic methanogens and other archae-bacteria (8, 9). Above temperatures of-80°C, this microbial sulfur respirationoccurs in addition to an abiological re-duction.

Microbial Iron and Manganese Oxidation

Deposits of iron and manganese ox-ides cover most surfaces exposed inter-mittently to plumes of hydrothermal andbottom seawater or to mixes of the two.The color of these encrustations rangesfrom almost black to light brown. Scan-ning electron microscopy reveals densemicrobial mats. A large variety of micro-bial forms are deeply embedded in themetal oxide deposits (Fig. 3, B and C).Not enough data exist to permit esti-

mates of the rate of mat formation. How-ever, when various types of materials(glass, plexiglass, steel, membrane fil-ters, and clam shells) enclosed in a pro-tective rack were placed into the openingof an active warm (21°C) vent for -10months, all surfaces were evenly black-ened (28, 30).

Nondispersive x-ray spectroscopyshowed a decrease of the Fe2+/Mn2+ratio in these layers with increasing dis-tance from vent openings (28), an obser-vation attributable to the different sol-ubility products of the two metals. X-raydiffraction determinations of the depos-its resulted in a correlation with the miner-al todorokite, (Mn,Fe,Mg,Ca,K,Na2) -

(Mn5O12) * 3H2O, which, in its fine-grained and poorly crystalline state, ischaracteristic of marine ferromanganesedeposits.The role of bacteria in the oxidative

deposition of iron is difficult to prove in723

neutral or alkaline waters where Fe2+undergoes rapid spontaneous oxidationin contact with dissolved oxygen. Het-erotrophically grown bacteria have beenshown to accumulate Fe3+ deposits, butno physiological significance of thisprocess has ever been demonstrated inthe marine environment.Although iron lithotrophy has been

demonstrated for acid freshwaters andsoils, true manganese lithotrophy has notbeen proven (39). The oxidation of Mn2+in seawater (pH -8.1) is more likely thanthe biological oxidation of Fe2+. Twobacterial isolates from the GalapagosRift vent region oxidized Mn2+ either ingrowing cultures or in cell extracts (39).The oxidation was heat-labile and inhib-ited by azide (NaN3), potassium cyanide(KCN), and antimycin A. The "oxi-dase" was inducible by reduced manga-nese and was not constitutive as in iso-lates obtained from manganese nodules.Since ATP synthesis was coupled withMn2+ oxidation it appears that Mn2+-oxidizing bacteria do contribute to thechemosynthetic production at deep-seahydrothermal vents.

The Role of Elevated Temperatures

The transfer of thermal to chemicalenergy takes place at temperaturesabove 350TC (Fig. 1). Thermophilic C02-,so42--, and So-reducing bacteria thatuse H2 as the source of electrons (Table1) are the best candidates for possiblemicrobial activities in hot zones wherebottom seawater mixes below the sur-face with rising hydrothermal fluid. Mi-crobial growth has been measured so farat temperatures up to H10°C in culturesof extremely thermophilic bacteria iso-lated from shallow and deep marine hotvents (40).The free 02 in this mix of hydrother-

mal fluid and bottom seawater may bequickly consumed biologically as well aschemically, and both aerobic and anaer-obic microorganisms may exist in sub-surface vent systems. Most aerobic bac-terial isolates obtained from the turbidwaters emitted by some of the GalapagosRift warm vents were "mesophilic," thatis, exhibited growth optima at tempera-tures of 250 to 35°C (24). "Extremelythermophilic" isolates obtained from thevarious types of shallow and deep hotvents are all anaerobic with growthranges from 650 to H10°C and growthoptima from 860 to 105°C (40). Most ofthese isolates belong to the "archaebac-teria," which are distinguished from the"eubacteria" and from all eukaryotic

724

organisms by their specific ribosomalRNA nucleotide sequences (41).A heterotrophic bacterium that grows

on a complex organic medium (peptoneand yeast extract) in a temperature rangefrom 550 to 98°C with an optimum at-88°C has recently been isolated from ashallow marine hot spring as well as fromdeep-sea vents (40). It has the facultativerespiration of elemental sulfur and someother characteristics in common with themethanogenic archaebacteria (19). Themethanogenic vent isolate discussedabove (38) differs from all other archae-bacteria in having a unique macrocyclicglycerol diether instead of a tetraether asthe polar membrane lipid (42), which issuspected of affecting the membrane flu-idity at high temperatures.

Bacterial growth at temperatures up to250°C by a natural population collectedfrom a hot vent has also been reported,but the experimental proof of this studyis still being contested (21). Other studieswith natural populations collected fromthe immediate vicinity of hot vents re-sulted in the microbial production ofgases at 100°C (37) and in the incorpo-ration of adenine into RNA and DNA atrates that were higher at 90°C than at 21°and 50°C (24). It has also been speculat-ed that the particular conditions of deep-sea hydrothermal vents might lead to asynthesis of organic compounds and ulti-mately to the origin of life (43).Thorough analysis of particulate or-

ganic carbon has only been done at con-siderable distances from warm ventemissions (44). The results demonstrateda rather quick passage and completetransformation of microbially producedorganic compounds into those character-istic of certain grazers (zooplankton).Concern about bacterial growth at hotvents is not so much a question of wheth-er there is a substantial addition to pri-mary production but rather the questionof the problem of biological activity at anupper temperature limit per se.

In the early spring of 1984, densecommunities of marine invertebrateswere also discovered at a depth of 3200m at the base of the West Florida Es-carpment, a site without volcanic or geo-thermic activity (45). In this area H2S-containing ground water with a salinityabout one-third higher than that of theambient seawater seeps from jointedlimestone formations. The types of ani-mals found here are similar to thosedescribed from the vent sites of the EastPacific Rise, but the individuals as wellas the total quantities are smaller. Thepresence of H2S has not been measured,but it is inferred from the odor of the

collected samples. The temperatures ofthese nongeothermal seepages are nearambient, that is, about 0.15°C aboveambient when measured at a depth of 10cm in the sediment.From the distribution pattern of inver-

tebrates at the tectonic vent sites, itappears that the spotty occurrence ofelevated temperature is of secondary im-portance for the abundance of these pop-ulations. The overriding factors seem tobe the availability of inorganic chemicalspecies and the efficiency of their use inchemosynthesis.

Symbiotic Chemosynthesis

One major evolutionary developmentis responsible for the unusual amounts ofbiomass found at the deep-sea vents: anew type of symbiosis between chemo-synthetic bacteria and certain marine in-vertebrates. Symbiosis is not commonlya topic of geomicrobiology, but this new-ly discovered highly efficient transforma-tion of geothermal or geochemical ener-gy for the production of organic carbonposes a new situation.The predominant part of the biomass

observed at the warm deep-sea vents isgenerated by the symbiotic associationof prokaryotic cells in the clam Calpyto-gena magnifica and the pogonophorantube worm Riftia pachyptila (46) (Fig. 4).The microbial symbionts have not yetbeen isolated, but their prokaryotic na-ture, DNA base ratio, genome size, andenzymatic activities identify them asbacteria (36, 47). They are found withinthe gill cells of C. magnifica and, as aseparate "trophosome" tissue, withinthe body cavity of R. pachyptila. Thetrophosome may amount to 60 percent ofthe worm's wet weight.The animal's dependence on the mi-

crobial symbiont has developed to thepoint where all ingestive and digestivemorphological features have been lost.Through an active blood system the ani-mal provides the bacteria in the tropho-some with H2S and free 02. It appearsthat the spontaneous reaction of the twodissolved gases is prevented or slowedby the presence of an HS--binding pro-tein (48). The isolation of CH4-oxidizingbacteria from Calyptogena gill tissue andRiftia trophosome (33) indicates, but cer-tainly not conclusively, that chemosyn-thesis by CH4 assimilation (ribulose mo-nophosphate pathway) may also takeplace. Enzymes associated with both theATP-producing system and the Calvincycle have been found in Riftia andCalyptogena. Physiological work on pu-

SCIENCE, VOL. 229

rified preparations of symbionts fromRiftia and the newly described vent mus-sel Bathymodiolus thermophilus (49)showed that their chemoautotrophic ac-tivities differ greatly with respect to tem-perature and the type of electron donorused (50).

Probably because of heavy predationof dying vent communities, fossilizedanimal remains in metal-rich deposits ofancient sea-floor spreading centers andpresently mined ophiolites have onlyrarely been found (51). Evidence formicrobial activities at similar sites hasbeen based on the results of sulfur iso-tope analyses (52).The most significant geomicrobiologi-

cal point of the deep-sea vent discoveryis the dependence of entire ecosystemson geothermal (terrestrial) rather thansolar energy. Were a catastrophic dark-ening of the earth's surface to occur (53),the chance of survival of such ecosys-tems is the highest of any community inthe biosphere. The chemosynthetic exis-tence of organisms in the deep sea alsosuggests a possible occurrence of similarlife forms in other planetary settingswhere water may be present only in theabsence of light. It is surprising that, asfar as we know, science fiction writersdid not turn their attention to geochemi-cally supported complex forms of lifeuntil such forms were actually discov-ered in the deep sea.

References and NotesI. D. L. Williams, R. P. von Herzen, J. G. Sclater,

R. N. Anderson, Geophys. J. R. Astron. Soc.38, 587 (1974); R. F. Weiss, P. Lonsdale, J. E.Lupton, A. E. Bainbridge, H. Craig, Nature(London) 267, 600 (1977).

2. J. B. Corliss et al., Science 203, 1073 (1979).3. P. F. Lonsdale, Deep-Sea Res. 24, 875 (1977).4. H. W. Jannasch and C. 0. Wirsen, BioScience

29, 592 (1979).5. D. M. Karl, C. 0. Wirsen, H. W. Jannasch,

Science 207, 1345 (1980).6. W. Pfeffer, Planzenphysiology (Engelmann,

Leipzig, 1897), vol. 1.7. R. K. Thauer and J. G. Momris, in The Microbe

1984, D. P. Kelly and N. G. Carr, Eds. (Cam-

bridge Univ. Press, Cambridge, 1984), vol. 2,pp. 123-168.

8. K. 0. Stetter and G. Gaag, Nature (London)305, 309 (1983).

9. S. Belkin, C. 0. Wirsen, H. W. Jannasch, Appl.Environ. Microbiol. 49, 1057 (1985).

10. J. M. Edmond et al., Earth Planet. Sci. Lett. 46,1(1979); ibid., p. 19.

11. K. L. Von Damm, thesis, Woods Hole Oceano-graphic Institution-Massachusetts Institute ofTechnology (1983).

12. M. J. Mottl, H. D. Holland, R. F. Conf, Geo-chim. Cosmochim. Acta 43, 869 (1979); W. C.Shanks III, J. L. Bischoff, R. J. Rosenbauer,ibid. 45, 1977 (1981).

13. H. Craig, et al., Eos 61, 992 (1980).14. M. J. Mottl, Geol. Soc. Am. Bull. 94, 161 (1983).15. J. L. Bischoff and W. E. Seyfried, Jr., Am. J.

Sci. 278, 838 (1978); R. M. Haymon and M.Kastner, Earth Planet. Sci. Lett. 53, 363 (1981).

16. J. F. Kerridge, R. M. Haymon, M. Kastner,Earth Planet. Sci. Lett. 66, 91 (1983); M. Arnoldand S. M. F. Sheppard, ibid. 56, 148 (1981); M.M. Styrt et al., ibid. 53, 382 (1981).

17. M. D. Lilly, J. A. Baross, L. I. Gordon, inHydrothermal Processes at Seafloor SpreadingCenters, P. A. Rona et al., Eds. (NATO Confer-ence Series IV, Plenum, New York, 1983), vol.12, pp. 411-449; M. D. Lilly, M. A. de Angelis,L. I. Gordon, Nature (London) 300, 48 (1982); J.A. Baross, M. D. Lilly, L. I. Gordon, ibid. 298,366 (1982).

18. W. E. Seyfried, Jr., and D. R. Janecky, Extend-ed Abstracts, Fourth Int. Symp. on Water-RockInteraction (International Association of Geo-chemists and Cosmochemists, Misasa, Japan,1983), p. 433; D. R. Janecky and W. E. Seyfried,Jr., ibid., p. 210; Geochim. Cosmochim. Acta,in press.

19. J. A. Welhan and H. Craig, in HydrothermalProcesses at Seafloor Spreading Centers, P. A.Rona et al., Eds. (NATO Conference Series IV,Plenum, New York, 1983), vol. 12, pp. 391-409.

20. M. J. Mottl and H. D. Holland, Geochim. Cos-mochim. Acta 42, 1103 (1978).

21. J. A. Baross, J. W. Deming, R. R. Becker, inCurrent Perspectives of Microbial Ecology, M.J. Klug and C. A. Reddy, Eds. (AmericanSociety of Microbiology, Washington, D.C.,1984), pp. 186-195; J. D. Trent, R. A. Chastain,A. A. Yayanos, Nature (London) 307, 737(1984); R. H. White, ibid. 310, 430 (1984).

22. R. E. McDuff and J. M. Edmond, Earth Planet.Sci. Lett. 57, 117 (1982).

23. D. M. Karl, Microbiol. Rev. 44, 739 (1980);Appl. Environ. Microbiol. 42, 802 (1981); C. D.Winn and D. M. Karl, ibid. 47, 835 (1984).

24. D. M. Karl et al., Mar. Biol. Lett. 5, 227 (1984).25. E. G. Ruby, C. 0. Wirsen, H. W. Jannasch,

Appl. Environ. Microbiol. 42, 317 (1981); E. G.Ruby and H. W. Jannasch, J. Bacteriol. 149, 161(1982).

26. J. H. Tuttle, C. 0. Wirsen, H. W. Jannasch,Mar. Biol. 73, 293 (1983).

27. H. W. Jannasch, in The Microbe 1984, D. P.Kelly and N. G. Can, Eds. (Cambridge Univ.Press, Cambridge, 1984), vol. 2, pp. 97-122.

28. H. W. Jannasch and C. 0. Wirsen, Appl. Envi-ron. Microbiol. 41, 528 (1981).

29. D. C. Nelson, J. B. Waterbury, H. W. Jannasch.Arch. Microbiol. 133, 172 (1982); D. C. Nelsonand H. W. Jannasch, ibid. 136, 262 (1983).

30. H. W. Jannasch, in Hydrothermal Processes at

Seafloor Spreading Centers, P. A. Rona et al.,Eds. (NATO Conference Series IV, Plenum,New York, 1983), vol. 12, pp. 677-709.

31. J. M. Edmond, personal communication.32. R. Whittenbury and H. Dalton, in The Prokary-

otes, M. P. Starr et al., Eds. (Springer, Berlin,1981), vol. 1, pp. 894-902.

33. R. S. Hanson, personal communication.34. G. A. Zavarzin and A. N. Nozhevnikova, Mi-

crob. Ecol. 3, 305 (1977); Y. Park and G. Hege-man, in Microbial Chemoautotrophy, W. R.Strohl and 0. H. Tuovinen, Eds. (Ohio StateUniv. Press, Columbus, 1984), pp. 211-218.

35. H. G. Schlegel, in Marine Ecology, 0. Kinne,Ed. (Wiley, New York, 1975), vol. 2, pp. 9-60.

36. H. W. Jannasch and D. C. Nelson, in CurrentPerspectives of Microbial Ecology, M. J. Klugand C. A. Reddy, Eds. (American Society ofMicrobiology, Washington, D.C., 1984), pp.170-176.

37. J. A. Baross, M. D. Lilley, L. I. Gordon, Nature(London) 298, 366 (1982).

38. W. J. Jones et al., Arch. Microbiol. 136, 254(1983).

39. H. L. Ehrlich, in Microbial Chemoautotrophy,W. R. Strohl and 0. H. Tuovinen, Eds. (OhioState Univ. Press, Columbus, 1984), pp. 47-56;Ecol. Bull. (Stockholm) 35, 357 (1983).

40. W. Zillig et al., Naturwissenschaften 69, 197(1982); K. 0. Stetter, Nature (London) 300, 258(1982); S. Belkin and H. W. Jannasch, Arch.Microbiol. 141, 181 (1985).

41. C. R. Woese, Zentralb. Bakteriol. Parasitenkd.Infektionskr. Hyg. I Abt. Orig. C3, 1 (1982).

42. P. B. Comita and R. B. Gagosian, Science 222,1329 (1983).

43. J. B. Corliss, J. A. Baross, S. E. Hoffman,Oceanol. Acta N°SP, 59 (1981).

44. P. B. Comita, R. B. Gagosian, P. M. Williams,Nature (London) 307, 450 (1984).

45. C. K. Paull et al., Science 226, 965 (1984).46. K. J. Boss and R. D. Turner, Malacologia 20,

161 (1980); M. L. Jones, Proc. Biol. Soc. Wash.93, 1295 (1980).

47. C. M. Cavanaugh et al., Science 213, 340 (1981);C. M. Cavanaugh, Nature (London) 302, 56(1983); D. C. Nelson, J. B. Waterbury, H. W.Jannasch, FEMS Microbiol. Lett. 24, 267 (1984);H. Felbeck, J. J. Childress, G. N. Somero, inThe Mollusca, P. W. Hochachka, Ed. (Academ-ic Press, London, 1984), vol. 2, pp. 331-358.

48. A. G. Arp and J. J. Childress, Science 213, 342(1981); ibid. 219, 295 (1983).

49. V. C. Kenk and B. R. Wilson, Malacologia 26,253 (1985).

50. D. C. Nelson, S. Belkin, H. W. Jannasch, inpre aration.

51. R. M. Haymon, R. A. Koski, C. Sinclair, Sct--ence 223, 1407 (1984).

52. A. J. Boyce, M. L. Coleman, M. J. Russell,Nature (London) 306, 545 (1983).

53. P. R. Ehrlich et al., Science 222, 1293 (1983).54. H. W. Jannasch and C. D. Taylor, Annu. Rev.

Microbiol. 38, 487 (1984).55. We thank C. 0. Wirsen, D. C. Nelson, J. B.

Waterbury, and S. J. Molyneaux for their col-laboration in many aspects of this work and J.M. Peterson for assistance in the preparation ofthe manuscript. Our research is supported byNational Science Foundation grants OCE82-14896, OCE83-08631 (H.W.J.), and OCE83-10175 (M.J.M.). Contribution 5764 of the WoodsHole Oceanographic Institution.

23 AUGUST 1985 725

View publication statsView publication stats