Microfossils of the Early Archean Apex Chert: New Evidence of the Antiquity of Life Author(S): J

Home , Allamoore Formation

Microfossils of the Early Archean Apex Chert: New Evidence of the Antiquity of Life Author(s): J. William Schopf Reviewed work(s): Source: Science, New Series, Vol. 260, No. 5108 (Apr. 30, 1993), pp. 640-646 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/2881249 . Accessed: 07/04/2012 21:09

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org REFERENCESAND NOTES 9. Y. Hiwatari,in Molecular Dynamics Simulations, F. (World Scientific, Singapore, 1988), vol. 1, p. Yonezawa, Ed., vol. 113 of Springer Series in 247. 1. H. Takayama, News Letter No. 1 (Research Group Solid State Sciences (Springer-Verlag, Heidel- 23. F. Yonezawa, S. Sakamoto, M. Hori, J. Non-Ctyst. on a new project, "Computational Physics as a berg, 1992), p. 32. Solids 137, 135 (1991). New Frontier in Condensed Matter Research" 10. S. Fujiwara and F. Yonezawa, in preparation. 24. F. Yonezawa and S. Sakamoto, Optoelectronics- under the support of the Grant-in-Aidfor Scientific 11. S. Chandrasekhar, Contemp. Phys. 29, 527 Devices Technol. 7, 117 (1992). Research on Priority Areas from the Ministry of (1988). 25. R. Car and M. Parrinello, Phys. Rev. Lett. 55, 2471 Education, Science, and Culture of Japan, Tokyo, 12. K. M. Aoki and F. Yonezawa, Phys. Rev. A 46, (1985). 1991). 6541 (1992). 26. ,_ ibid. 60, 204 (1988); G. Galli, R. M. Martin, 2. N. Metropolis, A. Rosenbluth, M. Rosenbluth, A. 13. K. Omata, K. M. Aoki, F. Yonezawa, in prepara- R. Car, M. Parrinello, ibid. 62, 555 (1989); ibid. 63, Teller, E. Teller, J. Chem. Phys. 21 1087 (1953). tion. 988 (1989); 1. Stich, R. Car, M. Parrinello, ibid., p. 3. F. Yonezawa, Ed., Molecular Dynamics Simula- 14. M. Cheng, J. T. Hsi, R. Pindak, Phys. Rev. Lett. 61, 2243; G. Seifert, G. Pastore, R. Car, J. Phys. tions, vol. 103 of the Springer Series in Solid 550 (1988). Condens. Matter 4, L179 (1992); P. Ballone, W. State Sciences (Springer-Verlag, Heidelberg, 15. B. I. Halperin and D. R. Nelson, ibid. 41, 121 Andreoni, R. Car, M. Parrinello, Phys. Rev. Lett. 1992). (1978). 60, 271 (1988). 4. S. Nos6, Ed., Prog. Theor. Phys. Suppl. 103, 1 16. R. J. Birgeneau and J. D. Lister, J. Phys. Paris 39, 27. T. Oguchi and T. Sasaki, Prog. Theor. Phys. (1991). L399 (1978). Suppl. 103, 93 (1991). 5. F. Yonezawa, in Solid State Physics, H. Ehren- 17. R. Pindak, D. E. Moncton, S. C. Davey, J. S. 28. Y. Morikawa, K. Kobayashi, K. Terakura, S. reich and D. Turnbull,Eds. (Academic Press, New Goodby, Phys. Rev. Lett. 46, 1135 (1981). Blugerl, Phys. Rev. B44, 3459 (1991). York, 1990), vol. 45, p. 179. 18. R. Geer et al., Nature 355, 152 (1992). 29. M. Tsukada, K. Kobayashi, N. Isshiki, H. 6. _ , S. Nos6, S. Sakamoto, Neue Folge 156, 19. W. E. Spear and P. G. LeComber, Solid State Kageshima, Surf. Sci. Rep. 13, 265 (1991). 77 (1988). Commun. 17, 1193 (1975). 30. W. Kohn and L. J. Sham, Phys. Rev. 140, 1133 7. F. C. Frank, Proc. R. Soc. London Ser. A 215, 43 20. D. L. Staebler and C. R. Wronski, Appl. Phys. Lett. (1965). (1952). 31, 292 (1977). 31. G. B. Bachelet, D. R. Hamann, M. Schluter, Phys. 8. J.-L. Barrat, J.-N. Roux, J.-P. Hansen, Chem. 21. K. Morigaki, Jpn. J. Appl. Phys. 27, 163 (1988). Rev. B26, 4199 (1982). Phys. 149,198 (1990); J.-L. Barratand M. L. Klein, 22. W. B. Jackson and J. Kakalios, in Amorphous 32. I thank my students S. Sakamoto, K. M. Aoki, S. Annu. Rev. Phys. Chem. 42, 23 (1991). Silicon and Related Materials, H. Fritzsche, Ed. Fujiwara, and K. Omata for collaboration.

understandingof the evolutionarystatus of early evolving microorganisms,and suggest that cyanobacterialoxygen-producing pho- tosynthesizersmay have alreadybeen extant Microfossils of the Early Archean this earlyin Earthhistory. The Archean fossil record. Unlike that Apex Chert: New Evidence of the of the laterPrecambrian Proterozoic (5), the fossil recordof the Archean is minuscule; few fossils have been detected and their Antiquity of Life studyhas been plaguedby misinterpretation and questionableresults (4). In order to J. William Schopf establishthe authenticityof Archeanmicro- fossils,five principalcriteria must be satisfied Eleven taxa (including eight heretofore undescribed species) of cellularly preserved fila- (3). The putativemicrofossils must (i) occur mentous microbes, among the oldest fossils known, have been discovered in a bedded in rocks of known provenanceand (ii) es- chert unit of the Early Archean Apex Basalt of northwestern Western Australia. This tablishedArchean age; (iii) be demonstrably prokaryotic assemblage establishes that trichomic cyanobacterium-like microorganisms indigenousto and (iv) syngeneticwith the were extant and morphologically diverse at least as early as -3465 million years ago and primarydeposition of the enclosing rock; suggests that oxygen-producing photoautotrophy may have already evolved by this early and (v) be of assuredbiological origin. All stage in biotic history. but a few of the microfossil-likeobjects reported from Archean sediments have failedto meet one or moreof these require- ments (3, 4). Among recent such examples When life originatedand the rate of evo- severelyaltered by metamorphism(1). The wasthe discoveryof authenticmicrofossils in lution and diversificationof the earlybiota most promisingterrain for such studies is rocks evidently belonging to the Early continue to be fascinatingquestions. Simi- that of the PilbaraBlock of northwestern Archean WarrawoonaGroup of Australia larly, it is unclearwhen a physiologically WesternAustralia, a region underlainby a (6), a reportunconfirmed because it has not modem ecosystembased on oxygen-produc- 30-km-thicksequence of relativelywell-pre- proved possible to relocate the geologic ing photosynthesisbecame established. The servedsedimentary and volcanic rocksthat sourceof the fossiliferoussamples (3). Sim- sole source of direct evidence relevant to are -3000 to -3500 millionyears old (Fig. ilarly, becauseof their simple morphology, such questionsis the paleobiologicrecord 1). From this region, I describea diverse solitaryunicell-like spheroids reported from contained in rocks deposited during the assemblageof filamentousmicrobial fossils severalArchean units (7, 8) arebest consid- ArcheanEon of Earthhistory [>2500 mil- detectedin the EarlyArchean (-3465 mil- ered to be possibly rather than assuredly lion years ago (Ma)]. The search for lion yearsold) Apex chert, cellularprokary- biogenic (3, 4). Other than the filamentous Archean fossils, however, is fraughtwith otes morethan 1300million years older than Apex fossilsdiscussed below, the relatively difficulty:Few Archean sedimentaryrocks any comparablesuite of fossils previously well-establishedArchean microfossil record have survivedto the present,and paleobio- reportedfrom the geologicrecord. Microfos- consistsof two typesof cyanobacterium-like logic evidencein most such units has been sils were first discoveredin this deposit in filamentsfrom the -2750-million-year-old 1986 (2); in a preliminaryaccount, three TumbianaFormation of WesternAustralia The author is in the Center for the Study of Evolution taxawere identified (3). The eight addition- (4); sheath-enclosedcolonial unicells oc- and the Origin of Life, Institute of Geophysics and al species describedhere demonstratethat curring in -3465-million-year-old sedi- Planetary Physics, the Department of Earth and Space Sciences, and the Molecular Biology Institute, Univer- the EarlyArchean biota was more diverse mentary rocks of the Towers Formation, sity of California, Los Angeles, CA 90024. than previouslyknown (3, 4), providenew also of Western Australia(2); and narrow

640 SCIENCE * VOL. 260 * 30 APRIL 1993 nonseptate bacterium-likefilaments from mapped in detail (13). Studies of petro- mineralizedin subangularto roundedsili- -3450-million-year-oldunits of the Swazi- graphicthin sections demonstratethat the ceous sedimentaryclasts less than 1 mm to land Supergroupof South Africa (3, 8, 9). three-dimensionalfossils are cellularlyper- a few millimetersin diameter(Fig. 3, A and Although stromatolites (finely layered mound-shapedsedimentary structures pro- Fig. 1. Stratigraphiccolumn (13, duced by microbial communities) have Pilbara 15), distributionof reportedstro- been reportedfrom more than 20 Archean Supergroup I (Gad 313C(per mil) matolites and microfossils (3), ap- geologic units (10), including the Tumbi- 30 WHIMCREEK 3.0 -30 -20 -10 0 GGROUP proximateages (15, 40), and car- ana, Towers, and Swazilanddeposits ( 1), 0 =Organic carbon 1 bon isotopic data (14) for geolog- their putative biological origin has been 25- * =Carbonatecarbon J ic units of the Pilbara Supergroup questioned(12). GORGE of northwesternWestern Austra- Geologicsetting. The microfossilassem- CREEK g H/C=0.09(N=1) lia. GROUP blage that I describeis foundin a sedimen- 20- tarychert unit of the Apex Basalt,a 1.5- to E 00 2.0-km-thick formationconsisting of tho- e) 15- WYMANFM 1 -3.3 leiitic pillow lava, high-magnesiumbasalt, c and komatiite interbedded with minor Yd CL EURO BASALT . PANORAMAFM -3.4C chert members that immediatelyoverlies - CYT APEX BASALT GD0 0 the TowersFormation in the lowerthird of 10 C- OWER M _ (3(D 0

Table 1. Morphological characteristics of microbial taxa from the Early globose; H, hemispheroidal; MAR, markedly; MOD, moderately; N, num- Archean Apex chert. All measurements are in micrometers. Arch., Ar- ber measured; NOT, not at all; P, pillow-shaped; Q, quadrate; R, rounded; chaeotrichon; E Eoleptonema; P., Primaevifilum;A., Archaeoscillatoriop- SC, short-cylinder; SL, slightly; SP, spheroidal; VMAR, very markedly; sis. B, blunt-rounded; C, conical; D, disc; FL, flat; FR, flat-rounded; G, VSL, very slightly; and Avg., average.

Medial cells cells Trichomes

Taxon Width Length Medial Terminal N Attenuated cd Max- Cosrce N shdape shmiape N toward mum Range Avg. Range Avg. sape shape apices at septa length

Arch. septatum, n. sp. 31 0.5 to 0.6 0.5 0.5 to 0.8 0.6 Q R 2 NOT NOT 34 E. apex, n. sp. 24 0.7 to 1.2 1.0 0.8 to 1.4 1.1 SP FR/H 3 NOT MOD 31 P. minutum, n. sp. 72 1.2 to 2.1 1.6 0.8 to 2.0 1.5 Q FR 8 NOT VSL/NOT 28 P. delicatulum Schopf, 1992 673 1.8 to 3.2 2.5 0.7 to 2.2 1.5 D/SC FL/FR 51 NOT SL/NOT 46 P. amoenum Schopf, 1992 554 2.0 to 5.0 3.8 1.5 to 4.5 2.8 Q/D H 47 MOD MOD/NOT 89 A. disciformis, n. gen., n. sp. 123 3.0 to 5.5 4.2 0.8 to 2.2 1.5 D H/G 12 SU/MOD MOD/MAR 39 P. conicoterminatum Schopf, 188 4.0 to 6.0 5.0 1.4 to 3.2 2.2 D/SC B/C 29 SUMOD SUMOD 72 1992 P. laticellulosum, n. sp. 122 6.0 to 8.5 7.0 2.5 to 5.0 3.5 SC/Q P 13 SL/NOT SL/NOT 82 A. grandis, n. gen., n. sp. 26 8.0 to 11.5 9.0 1.0 to 3.5 2.0 D FR/H 2 NOT NOT 45 P. attenuatum, n. sp. 49 4.0 to 12.0 7.5 1.0 to 4.0 3.0 D FR 5 VMAR SL/MOD 35 A. maxima, n. gen., n. sp. 15 15.0 to 19.5 16.5 3.0 to 6.0 4.5 D FR/H 2 NOT NOT 69

SCIENCE *m VOL. 260 * 30 APRIL 1993 641 B). These small fossiliferous clasts make up Eleven taxa (Table 1) of filamentous, nized and embeddedin mucilage, the fila- -5 percent of the rock and are distinguish- darkbrown to blackcarbonaceous microfos- ments exhibit neitherthe subparallelorien- able from other clastic components by their sils, includingeight new species(see appen- tation nor the laminarorganization typical fine-grained relatively homogeneous tex- dix), have been identifiedin the deposit. of most stromatoliticmicrobiotas (16). Mi- ture and their grayish brown to dark brown Solitary unicell-like spheroidsof possible crofossilshave not been detectedin stroma- color. The remainder of the bedded chert is but uncertainbiological origin also occur tolite-like laminatedclasts that also occur composed of similarly rounded unfossilifer- (Fig. 5, K and L). The assuredfossils occur in the unit (Fig. 3C). ous clasts and lithic fragments, less than 1 as irregularlydistributed and randomlyori- In comparisonwith permineralizedmi- mm to more than 20 cm in size, and of ented solitaryfilaments (Fig. 3B) surround- crobiotasfrom the later Precambrian(16), siliceous matrix. Laterally, over a distance ed by more or less homogeneousbrown to the Apex assemblageis highly carbonized of less than 100 m from the fossiliferous dark brown kerogen. The kerogenis floc- and poorlypreserved. Of the thousandsof locality, the fossiliferous bed merges into a culent and composed of very fine (<0.3 fragmentsof cellular filamentsdetected in continuous brecciated gray chert unit, 20 to ,um) particles, which might originally have the deposit (by examinationof a total area 30 m thick, that is concordant and inter- been mucilaginous. Single taxa (especially, of -450 cm2of 150-gum-thickpetrographic fingering with associated volcanic rocks. Primaevifilumminutum, n. sp.; P. laticellulo- thin sections), less than 1 percent are suf- The variety and textures of the detrital sum, n. sp.; and P. attenuatum, n. sp.) or ficientlypreserved to warrantdetailed study clasts, the occurrence of cross-bedding in particular pairs or groups of taxa (for exam- and formal description. Such alteration parts of the outcrop, and the stratigraphic ple, P. delicatulumand Archaeoscillatoriopsis makes taxonomic delineation difficult. relations to adjacent units demonstrate that disciformis,n. gen., n. sp.; or P. delicatulum, However, like modem filamentous mi- the fossiliferous chert is a primary sedimen- P. amoenum, and P. conicoterminatum)tend crobes (17, 18), membersof discrete size tary deposit and not of secondary diagenetic to predominate in individual clasts. Al- classes of relatively well-preservedApex or intrusive origin. The fossils are from though possibly representing a benthic mi- filamentsexhibit taxonomicallyuseful lim- samples collected on two occasions: in June crobial community that was loosely orga- ited rangesof consistentlyco-occurring ter- 1982 (Fig. 3, A to N; Fig. 4, A to D and F to H; and Fig. 5, A, B, E to G, and K and L) and in August, 1986 (Fig. 30; Fig. 4, E, I, and J; and Fig. 5, C, D, and H to J). Paleobiology. The Apex filaments meet all criteria required of bona fide Archean microfossils. (i) Their occurrence in rocks of known provenance has been substantiat- ed by replicate sampling of the fossiliferous locality. (ii) As summarized above, the stratigraphic relations and Early Archean (-3465 Ma) age of the fossiliferous cherts are well documented. (iii) The kerogenous [and iron-stained (Fig. 3L and Fig. 5, D to F)l fossils are encased within and unques- tionably indigenous to the Apex chert, as demonstrated by their occurrence in petrographic thin sections (Figs. 3 to 5). (iv) They are localized in organic-rich clasts (Fig. 3B) shown by petrographic relations [for example, cross-cutting veinlets that transect both the clasts and their encom- passing matrix; figure 1.5.4A in (3)] to be primary components of the sedimentary chert unit, assuredly syngenetic with its deposition. (v) As discussed below, their evident cellular organization, and their morphological complexity and similarity to younger prokaryotes, both fossil and modem, firmly establish their biogenicity. The presence of these microfossils in a petrographically distinctive population of clasts and their absence from all other clasts and the surrounding matrix (Fig. 3, A and B) indicate that the filaments predate dep- Fig. 3. Microfossiliferous(A and B) and laminatedstromatolite-like clasts (C), and carbonaceous osition of the chert unit and were initially and iron-stained(L) microfossils(with interpretive drawings) shown in thin sections of the Early preserved in older rocks, some part of which ArcheanApex chert of WesternAustralia. Except as otherwiseindicated, magnification of all parts denoted scale in to and and show of the sinuous three- was eroded, transported, and redeposited as by (N). (D K) (N 0) photomontages dimensionalmicrofossils. (A) Microfossiliferousclast; area denoted by dashed lines shown in (B). the a detrital component of bedded chert. (B) Arrowspoint to minutefilamentous microfossils, randomly oriented in the clast. (C) Portionof a Whether the microfossils are greatly older clast showing stromatolite-likelaminae. (D and E) Archaeotrichionseptatum, n. sp. (D, holotype). than or essentially penecontemporaneous (F) Eoleptonemaapex, n. sp. (holotype).(G and H) Primaevifilumminutum, n. sp. (G, holotype).(I, with deposition of the Apex chert is un- J, and K) Primaevifilumdelicatulum Schopf, 1992 (I, holotype) (3). (L, M, N, and 0) Archaeoscil- known. latoriopsisdisciformis, n. gen., n. sp. (M, holotype).

642 SCIENCE * VOL. 260 * 30 APRIL 1993 .;GZWS9...... RESEARCHARTICLE minal cell shapes, medial cell shapes and the -2080-million-year-oldGunflint Iron resemble trichomic (nonensheathed or dimensions,and degrees of trichomicatten- Formation(21), both of Canada. But the thinly ensheathed) oscillatoriaceancyano- uation (for example, compareFig. 3, L to fossil record from the greater than 1300 bacteria. Cell widths of the Apex taxa N; Fig. 4, F to H; and Fig. 5, A to C). million years intervening between these range from 0.5 p.m (Fig. 3, D and E) to These characteristics,and the similarityof depositsand the Apex chert is essentially 19.5 p.m (Fig. 5F) and average -5.0 p.m the delimitedsize classes (Table 1) to the undeciphered(5, 22). Although there is (Table 1). Modem filamentous bacteria size rangesof modem microbialtaxa (17, thus a profoundgap in the record, the tend to be quite narrow, predominantly 18), make it unlikely that any of the de- morphologicalsimilarity of the Apex fossils < 1.5 p.min diameter,whereas most oscil- scribedspecies representvariants of other to septate filamentousprokaryotes, both latoriaceantrichomes are notably broader membersof the assemblage.Indeed, as is Proterozoic(16) and modem (17, 18), in- (Fig. 6). On the basis of morphometric documentedfor younger Precambrianmi- dicates that they are almost certainlypro- analysesof more than 500 taxa of modem crobiotas(19), the incompletepreservation karyotesand part of an evolutionarycon- filamentousmicrobes, I have suggestedthat of the Apex fossilssuggests that the original tinuum that extends from the Early fossil septate filaments <1.5 p.m wide be assemblageprobably included more taxa Archeanto the present.This interpretation regardedas "probablebacteria," those 1.5 than the 11 speciesidentified. seemssupported by the occurrencein Apex p.m to 3.5 pumwide as (undifferentiated) The evolutionaryrelations of these mi- filamentsof bifurcatedcells and cell pairs "prokaryotes,"and those >3.5 p.mbroad as crofossils to younger fossils and modem [Fig.5, H to J; figure1.5.6, F and G, in (3)] "probablecyanobacteria" (23). Applying microorganismsare of interest.As currently that evidently reflectthe originalpresence these criteriato the Apex fossils,I interpret documented,the fossilrecord is moreor less of partialseptations and, thus, of cell divi- two taxa (Archaeotrichionseptatum, n. sp., continuousand relativelywell known from sion like that occurringin extant prokary- and Eoleptonenaapex, n. sp.) as probable about 2100 Ma to the present, beginning otic filaments(3). bacteria;two taxa (Primaevifilumminutum, with the diversemicrobiotas of the -2100- In comparisonwith modemprokaryotes, n. sp., and P. delicatulum)as eitherbacteria million-year-oldBelcher Group (20) and most of the Apex microbes particularly or cyanobacteria;and the remainingseven species, nearly two-thirdsof the taxa (and -63 percent of measuredspecimens) as probablecyanobacteria. Because the size ranges of filamentous bacteriaand cyanobacteriaoverlap (Fig. 6), the suggested affinities are not absolute. Nevertheless, the pattern of size distribu- tion exhibited by the Apex assemblageis more like that of modem oscillatoriaceans than of noncyanobacterialprokaryotes (Fig. 6). Furthermore,several of the Apex taxa, particularlythose with broad trichomes (Primaevifilumlatice11ulosum, n. sp.; Archae- oscillatoriopsisgrandis, n. gen., n. sp.; and A. maxima,n. gen., n. sp.), differin cell size from almost all bacteriabut are essen- tiallyindistinguishable from specific oscillatoriaceans,both Proterozoic(Oscillatoriopsis 20 spp.) and modem (Oscillatornaspp.). If the Apex filamentshad been discoveredin later Precambriansediments, in which fossil os- cillatoriaceansare well knownand relatively widespread(23), or if they had been detectedin a modem microbialcommunity and morphologywere the only criterionby which to infer biologicalrelationships, the majoritywould be interpretedas oscillatoriacean cyanobacteria.However, because the affinitiesof these fossilsin the Procary- otae cannot be demonstratedunequivocal- ly, I formallydescribe them as "prokaryotes IncertaeSedis"; and becausethe phylogenet- ic relations between them and the much (1300 to 2800 million years)younger, predominantly cyanobacterialfossil taxa to which they bear specific resemblanceare therefore undetermined, they have not referredto describedPro- Fig. 4. Carbonaceous microfossils(with interpretive drawings) shown in thin sections of the Early been previously ArcheanApex chert of WesternAustralia. Magnification of (D, E, I, and J) denoted by scale in (E); terozoicspecies (see appendix). magnificationof all other parts shown by scale in (A). (A, B, C, and D) and (F, G, H, and 1)show Evolutionaryimplications. The rangeof photomontages of the sinuous three-dimensionalmicrofossils. (A, B, C, D, and E) Primaevifilum morphologies exhibited by the Apex fila- amoenumSchopf, 1992 (A, holotype)(3). (F, G, H, I, and J) P. conicoterminatumSchopf, 1992 (H, ments indicates that if the majority are holotype) (3); arrowsin (I) point to conical terminalcells. oscillatoriaceans, this primitive family of

SCIENCE * VOL. 260 * 30 APRIL 1993 643 filamentouscyanobacteria was already high- (ii) The reactants requiredfor oxygenic additionallines of evidence, however, are ly diverse at Apex time. Although some photosynthesis,CO2 and H20, and mate- not conclusive; all but the latter, which cyanobacteriaare capable of temporarily rials possiblyrepresenting products of this necessarily incorporatesmodel-dependent carryingout anoxic (bacterial)photosyn- process, sedimentaryorganic matter and uncertainties,would be equally consistent thesis (24), oxygen-producingphotoau- oxidizediron minerals,were presentin the with the presenceof solely anoxic bacterial totrophy is a universal, presumablyearly- EarlyArchean environment(3). (iii) The photosynthesizers(3). Moreover,it is con- evolving characteristicof the group. The isotopic compositionsof EarlyArchean or- ceivable that the extemal similarityof the presenceof diverseoscillatoriaceans in the ganic and carbonatecarbon (for example, Apex microorganismsto younger oxygen- Apex assemblagewould thus seem to imply Fig. 1) areevidently indicative of photosyn- producing oscillatoriaceansmasks signifi- that this relativelyadvanced level of phys- thetic C02-fixation like that occurringat cant differencesof internalbiochemical ma- iological evolution had been attained at relativelyhigh CO2 concentrationsin ex- chinery (27); thus, their morphologymay least as earlyas -3465 Ma. tant microbialpopulations (25). (iv) Cal- provide a weak basis on which to infer Fourother lines of evidence seem con- culations based on models of the early paleophysiology.To addressthis issue, ad- sistent with the possible Early Archean globalecosystem, and ceriumand europium ditionaldata are neededregarding the ecol- existence of 02-producing oscillatori- concentrationsin Archean banded iron- ogy and communitystructure of the Apex aceans: (i) Early Archean stromatolites formations,suggest that 02-producingpho- assemblageand the evolutionaryrelations (10) werepresumably produced by photoau- tosynthesis and aerobic respirationboth that link these prokaryotesto the later, totroph-dominatedmicrobial communities. date from the EarlyArchean (26). These relatively well-documented Precambrian fossil record. Whether or not 02-producingphotoau- totrophs are representedamong the Apex fossils, the morphologicaldiversity of the assemblageis striking. In particular, the Apex filaments exhibit a greaterrange of diametersthan those reportedfrom all but 7 of the 70 other septatefilament-containing Precambrianunits known [Fig. 7 and 4P~~~~~~~~V data in (22, 23, 28)]. Because filament diameteris a principaltaxonomic charac- ter for such microorganisms(17, 18), the assemblageis notable also for its taxonom- - ~~~~~~~~~~~~~~~~~~`1-^m-7 ic diversity.The Apex assemblageis more Aa~~~~~~~~~~~ .::: . *,: diverse taxonomicallythan 92 percent of the 126 other septate filament-containing or tubularsheath-containing Precambrian assemblages known, and is more than twice as diverse as the average diversity (-5 taxa per formation)of all such assemblages (22, 28). Evidently, cellular filamentous microbesoriginated and diversi- fied early in Earth's history and have subsequently exhibited an exceedingly slow, hypobradytelic(27) rate of morphological evolution. The Apex microfossils thus providea glimpseof the diversityand evolutionarystatus of EarlyArchean life, a primitive microbial biota having evolu- tionary roots that must predate, perhaps substantially,-3465 Ma.

40 Bacteria (80 taxa) Apexfilaments .30 ~~~~~~~~(11 taxa) Oscillatoriacean

Fig. 5. Carbonaceous and iron-stained(D, E, and F) microfossils(with interpretive drawings) and possible microfossils(K and L) shown in thin sections of the EarlyArchean Apex chert of Western Australia.Magnification of (C, F, H, 1,and J) denoted by scale in (F);magnification of all other parts shown by scale in (B). (A to J) show photomontagesof the sinuous three-dimensionalmicrofossils. (A, B, and C) Primaevifilumlaticellulosum, n. sp. (A, holotype);pillow-shaped terminal cells are Cell width (jmi) indicated by arrows in (A) and (C). (D and E) Archaeoscillatoriopsisgrandis, n. gen., n. sp. (0, holotype). (F) Archaeoscillatoriopsismaxima, n. gen., n. sp. (holotype).(G) Primaevifilumattenua- Fig. 6. Cell widths of modern septate filamen- turn,n. sp. (holotype).(H, I, and J) Poorlypreserved trichomes showing bifurcatedcells and cell tous bacteria and oscillatoriacean cyanobacte- pairs (at arrows).(K and L) Solitaryunicell-like possible microfossils,in equatorial(left) and polar ria <20 ,um in diameter (23) compared with views (right). those of taxa from the Apex chert (Table 1).

644 SCIENCE * VOL. 260 * 30 APRIL1993 ...... RESEARCHARTICLE

Fig. 7. Range and aver- Diagnosis: Uniseriate unbranched trichomes, age diameters of septate 236 oun apparentlynot ensheathed, having the characteris- prokaryotic filaments re- 71 formations tics specifiedin Table 1. Etymology:With reference ported from Precambrian E to occurrence in the Apex chert. Type specimen: sediments, grouped in 60=L Trichome in Fig. 3F. Remarks:E. apex is morpho- 50-million-year-long in- logically comparable to unnamed narrow cellular - tervals based on estimat- Average E filamentsreported from the 1425-million-year-old ed formation ages (22, T '-'diameter 40.c Gaoyuzhuang Formation of China (31) and the 23, 28). The -3465-mil- Apex -1050-million-year-old Allamoore Formation of Texas (32), and to the modern bacteriumBeggiatoa lion-year-old Apex chert filaments e minima[(17), p. 114]. contains broader fila- 20 ments than those known Genus Primaevifilum Schopf, 1983 (6). from 22 of the 28 younger , - 0 Type species:Primaevifilum septatum Schopf, 1983 (6). intervals (dashed line); PHANERO- P R 0 T E R 0 Z 0 I C I A R C H E A N Primaevifilum minutum, n. sp. (Fig. 3, G and Ga, billion years ago. ZOICI p R E C A M B R I A N H; Table 1). 0 1 2 3 4 Diagnosis: Uniseriate unbranched trichomes, Age (Ga) apparentlynot ensheathed, having the characteristics specifiedin Table 1. Etymology:With reference to small diameter in comparisonwith other species Appendix: Systematic Paleontology Fig. 3M, 57.5/112.4 (W19/1). PPRG V.63730[1], of Primaevifilum.Type specimen: Trichome in Fig. colection 2644. Specimen 2644-2; section 2644-2B Type locality.[Schopf collection 4 of 6/15/82 and 3G. Remarks:P. minutumis morphologicallycom- (label right) with "X"at front left, 68.1/123.6 (file 8, PrecambrianPaleobiology Research Group collec- parableto unnamednarrow septate filamentsreport- row off slide): Fig. 3C, V.63731[1], 5 1.5/102.3 (L25/ tions 1458, 2006, and 2644 (29).] Outcrops of ed from the -2080-million-year-old Gunflint Iron 3); section -2C (label right) with "X"at front left, bedded chert from the Apex Basalt (Warrawoona Formationof Canada (33) and the 1025-million- 69.7/123.6 (file 7, row off slide):Fig. 3E, V.63732[1], Group, Pilbara Supergroup)on a hill immediately year-old Valyukhta Formationof Siberia (34). 50.0/108.3 (S26/2); Fig. 5L, V.63732[2], 35.3/116.6 south of Chinaman Creek and -12 km west of Primaevifilum laticellulosum, n. sp. (Fig. 5, A (file42, rowoff slide). Specimen2644-4; section -4C Marble Bar, Western Australia (Fig. 2), at grid to C; Table 1). (labelleft) with "X"at frontleft, 58.1/112.1 (W18/2): reference number 799558 on the Marble Bar Diagnosis: Uniseriate unbranched trichomes, Fig. 4D, V.63733[1], 20.3/99.5 057/2); Fig. 5E, 1:100,000 Australian National Topographic Map apparentlynot ensheathed, having the characteris- V.63733[2], 19.9/96.2 (E57/2). 1986 Collections: No. 2855, and at 21011'4"S and 119?42'36"E in tics specifiedin Table 1. Etymology:With reference PPRGcollection 2006. Specimen2006-1; section 2006- Archean map unit "Acj" of the Marble Bar struc- to large diameter in comparisonwith other species 1A (labelright) with "X"at frontleft, 71.5/135.4 (file tural belt on the Geological Survey of Western of Primaevifilum.Type specimen: Trichome in Fig. 4, row off slide):Fig. 4E, V.63734[1], 58.0/118.3 (file Australia 1:250,000 Marble Bar Geological Map 5A. Remarks:P. laticellulosumis similar in medial 18, rowoff slide);Fig. 4J, V.63734[2], 53.5/120.4 (file Sheet SF 50-8. cell size and shape to an unnamed oscillatoriacean- 23, row off slide); Fig. 5H, V.63734[3], 47.8/121.4 andstage coordinates of type trichome reported from the -1500-million- Repository figuredspecimens. (file29, rowoff slide). Specimen2006-2; section - 2A (Acquisitionnumbers for specimensreposited in the year-old Barney Creek Formationof Australia (35) (labelright) with "X"at front left, 69.1/134.9 (file 7, permanentcollections of The NaturalHistory Muse- and to the modem cyanobacterium Oscillatoria row off slide):Fig. 30, V.63735[1], 40.0/113.4 (X37/ um, CromwellRoad, LondonSW7 5BD, microscope tenuis[(18), p. 223], but differsfrom these by having 1); Fig. 51, V.63735[2], 38.3/110.5 (U38/2); Fig. 5J, stage coordinates on Leitz Orthoplan 2 automatic pillow-shapedterminal cells (Fig. 5, A to C). V.63735[3], 43.0/105.7 (P34/1). Specimen 2006-3; photomicroscopeUCLA No. 874-002635 and, in Primaevifilum attenuatum, n. sp. (Fig. 5G; section -3A (label right) with "X" at front left, parentheses,England Finder Slide coordinates.)1982 Table 1). 71.2/135.9 (file 6, row off slide):Fig. 41, V.63736[1], Collections:Schopf collection 4 of Rock spec- Uniseriate 6/15182. 40.9/111.8 (V36/3); section -3B (labelleft) with "X" Diagnosis: unbranched trichomes, imen 4 of 6/15/82-1;petrographic thin section number at frontleft, 68.3/133.9 (file8, row off slide):Fig. 5D, apparentlynot ensheathed, having the characteris- 4 of 6/15/82-1B (slide label right) with diamond tics in Table 1. With reference V.63737[1], 27.4/115.7 (Z50/3); section -3C (label specified Etymology: scribed"X" at front left, stage coordinates64.4/137.9 right)with "X"at frontleft, 67.4/132.4 (file8, row off to markedattenuation of the trichome. Type spec- (EnglandFinder Slide file 11, row off slide):Fig. 3K, imen: Trichome in 5G. Remarks:P. slide): Fig. 5C, V.63738[1], 28.3/97.1 (F49/circle). Fig. attenua- NaturalHistory Museum No. V.63164[4], 25.2/118.4 in Thin sectionscontaining topotypes of Apex taxa have tum, croissant-shaped complete specimens, differs (file 52, row off slide); Fig. 4B, V.63164[6], 25.4/ alsobeen repositedin the permanentcollections of the from previouslyreported filamentous microfossils by 120.1 (file 52, row off slide); Fig. 4G, V.63164[9], the markedattenuation of its trichomes. WesternAustralian Museum, Perth, Australia. 13.1/105.9 (P64/4);Fig. 4H, V.63164[1], 23.9/118.9 (file 53, row off slide); Fig. 5A, V.63164[10], 24.0/ Genus Archaeoscillatoriopsis, n. gen. 118.6 (file 53, row off slide); Fig. 5B, V.63164[11], Description of new taxa Type species: Archaeoscillatoriopsisdisciformis, n. 38.8/103.7 (N38/circle);Fig. 5F, V.63164[12], 46.1/ gen., n. sp. 131.6 (file 30, row off slide). Section - 1C (label KingdomProcaryotae Murray 1968, Incertae Sedis. Diagnosis: Trichomes uniseriate, unbranched, right)with "X"at frontleft, 66.6/138.8 (file9, rowoff Genus ArchaeotrichionSchopf, 1968 (30). cylindrical or slightly to moderatelytapered toward slide):Fig. 4F, V.63727[l], 68.0/131.8 (file 9, row off Type species: Archaeotrichioncontortum Schopf, apices, not at all to moderately or markedly con- slide);Fig. 5G, V.63727[2], 32.0/103.5 (N45/1). Sec- 1968 (30). stricted at septa, apparently not ensheathed, and tion - 1D (label right) with "X"at front left, 64.9/ Archaeotrichionseptatum, n. sp. (Fig. 3, D and commonly slightly to moderatelybent or disrupted; 130.0 (file 11, rowoff slide):Figs. 3A, B, V.63165[5], E; Table 1) medial cells disc-shaped, ranging from 3.0 to 19.5 65.0/97.0 (Fli/circle); Fig. 3H, V.63165[6], 65.1/ Diagnosis: Uniseriate unbranched trichomes, wm wide and from 0.8 to 6.0 wm long; terminal 97.3 (F11/3);Fig. 31, V.63165[2], 30.2/111.9 (U47/ possibly enclosed by a thin sheath, having the cells flat-rounded,hemispheroidal, or globose. Ey 3); Fig. 3J, V.63165[7], 65.1/97.9 (G 1/1); Fig. 3N, characteristics specified in Table 1. Etymoloy: mology: With reference to Archean age and mor- V.63165[8], 31.9/112.5 (W45/circle); Fig. 5K, With reference to cellularity (for example, Fig. phological similarity to fossil (Oscillatoriopsisspp.) V.63165[9], 31.9/112.6 (W45/circle). Section -1E 3D). Type specimen: Trichome in Fig. 3D. Re- and modern (Oscillatoriaspp.) oscillatoriaceans. (labelright) with "X"at frontleft, 61.9/139.2 (file 14, marks: The type species (A. contortum), first Archaeoscillatoriopsis disciformis, n. gen., n. row off slide):Fig. 3L, V.63728[1], 04.1/108.6 (S74/ described from the -850-million-year-old Bitter sp. (Fig. 3, L to 0; Table 1). circle);Fig. 4C, V.63728[2], 35.0/123.2 (file 42, row Springs Formation of central Australia (30), was Diagnosis:As for the genus, having the charac- off slide). Section - 1F (label left) with "X"at front established to include nonseptate threadlike fila- teristics specified in Table 1. Etymology: With left, 73.3/137.7 (file 2, row off slide); Fig. 3D, ments 0.5 to 0.7 [Lmin diameter. Although iden- reference to disc-shaped medial cells. Type speci- V.63166[3], 45.0/97.3 (F32/3);Fig. 4A, V.63166[l], tical in diameter, well-preservedspecimens of A. men: Trichome in Fig. 3M. Remarks:A. disciformis 68.3/115.1 (Z8/1). Section - 1G (labelleft) with "X" septatumare cellular (Fig. 3D). is morphologically comparable to specimens of at frontleft, 72.1/137.0 (file 4, row off slide):Fig. 3F, Gunflintiagrandis reported from the - 1950-million- V.63729[l], 52.3/113.0 (W24/3); Fig. 3G, Genus EoleptonemaSchopf, 1983 (6). year-old Tyler Formation of Michigan (36) and to V.63729[2], 53.4/113.9 (X23/3). PPRG colection Type species: Eoleptonemaaustralicum Schopf, 1983 the modern cyanobacteriumOscillatoria grunowiana 1458. Specimen1458-4; section 1458-4A (labelright) (6). 1(18), p. 216]. with "X"at frontleft, 70.1/123.8 (file6, row offslide): Eoleptonemaapex, n. sp. (Fig. 3F; Table 1). Archaeoscillatoriopsis grandis, n. gen., n. sp.

SCIENCE * VOL. 260 * 30 APRIL 1993 645 (Fig. 5, D and E; Table 1). lution, M. Schidlowski, S. Golubic, M. M. Kimberly, 26. K. M. Towe, Nature 348, 54 (1990); Palaeogeogr. Diagnosis:As for the genus, having the charac- D. M. McKirdy, P. A. Trudinger, Eds. (Springer- Palaeoclimatol. Palaeoecol. 97,113 (1991). teristics specified in Table 1. Etymology: With Verlag, Berlin, 1992), pp. 509-518; E. S. 27. J. W. Schopf, in (3), pp. 583-600. Barghoorn and J. W. Schopf, Science 152, 758 referenceto largediameter in comparisonwith most 28. _ , in ibid., pp. 11 19-1166. (1966); M. D. Muir and P. R. Grant, in The Early 29. T. B. Moore and J. W. Schopf, in ibid., pp. other taxa of Archaeoscillatoriopsis.Type specimen: History of the Earth (Wiley, London, 1976), pp. Trichome in Fig. 5D. Remarks:A. grandisis mor- 603-693. 595-604; A. H. Knoll and E. S. Barghoorn, Sci- 30. J. W. Schopf, J. Paleontol. 42, 651 (1968). phologically comparableto Oscillatoriopsismedia de- ence 198, 396 (1977). 31. _ , W. 0. Zhu, Z. L. Xu, J. Hsu, Precambrian 8. M. M. Walsh, Precambrian Res. 54, 271 (1992). scribed from the -1250-million-year-old Sukhaya Res. 24, 335 (1984). 9. __ and D. R. Lowe, Nature314, 530 (1985). Tunguska Formation of Siberia (37) and reported 32. A. V. Nyberg and J. W. Schopf, ibid. 16, 129 10. M. R. Walter, in (1), pp. 187-213; H. J. Hofmann, from the -900-million-year-old Deoban Limestone (1981). R. P. Sage, E. N. Berdusco, Econ. Geol. 86,1023 of India (38), and to the modem cyanobacterium 33. P. E. Cloud, Jr., Science 148, 27 (1965), Oscillatoriachalybea [(18), p. 219]. (1991). 11. G. R. Byerly, D. R. Lowe, M. M. Walsh, 34. J. W. Schopf et al., Precambrian Res. 4, 269 (1977). n. Nature319, Archaeoscillatoriopsis maxima, gen., n. sp. 489 (1986); M. R. Walter, R. Buick, J. S. R. Dunlop, 35. J. H. Oehler, Alcheringa 1, 315 (1977). (Fig. 5F; Table 1). ibid. 284, 443 (1980); D. R. Lowe, ibid. 284, 441 36. P. Cloud and K. Morrison, Geomicrobiol. J. 2,161 Diagnosis:As for the genus, having the charac- (1980). (1980). teristics specified in Table 1. Etymology: With 12. R. Buick, J. S. R. Dunlop, D. I. Groves, Alcheringa 37. C. V. Mendelson and J. W. Schopf, J. Paleontol. referenceto very large diameterin comparisonwith 5, 161 (1981); R. Buick, Palaios 5, 441 (1991). 56, 42 (1982). all other taxa of Archaeoscillatoriopsis.Type speci- 13. A. H. Hickman, W Aust. Geol. Surv. Bull. 127, 1 38. M. Shukla, V. C. Tewari, V. K. Yadav, Palaeobot- men: Trichome in Fig. 5F. Remarks:A. maximais (1983); and S. L. Lipple, Explanatory anist35, 347 (1986). morphologicallycomparable to unnamed broad os- Notes MarbleBar 1:250,000 GeologicalMap Se- 39. J. W. Schopf and Yu, K. Sovietov, Science 193, ries (Western Australia cillatoriacean trichomes from the -650- Geological Survey, Perth, 143 (1976). reported 1978), pp. 1-24. million-year-old Chichkan Formation of Kazakh- 40. T. S. Blake and N. J. McNaughton, in Archean and 14. H. Strauss and T. B. Moore, in (3), pp. 709-798. Proterozoic Basins of the Pilbara, Westem Austra- stan (39) and to the modem cyanobacteriumOscil- 15. R. I. Thorpe, A. H. Hickman, D. W. Davis, J. K. lia: Solution and Mineralization Potential, J. R. latoriaantillarum [(18), p. 2421. Mortensen, A. F. Trendall, Precambrian Res. 56, Muhling, D. I. Groves, T. S. Blake, Eds. (Publ. 9, 169 (1992). Univ. W. Aust. Geol. Dept. and Univ. Extension, REFERENCESAND NOTES 16. M. R. Walter, J. P. Grotzinger, J. W. Schopf, in (3), Perth, 1984), pp. 1-22. pp. 253-260; J. W. Schopf, in ibid., pp. 1055- 41. I thank the other participants in the Precambrian 1. J. M. Hayes, I. R. Kaplan, K. W. Wedeking, in 1117. Paleobiology Research Group (PPRG) fieldwork of Earth's Earliest Biosphere, J. W. Schopf, Ed. 17. R. E. Buchanan and M. E. Gibbons, Bergey's 1982 (K. J. Armstrong, D. Blight, D. J. Chapman, J. (Princeton Univ. Press, New Jersey, 1983), pp. Manualof DeterminativeBacteriology (Williams & M. Hayes, A. H. Hickman, C. Klein, D. D. Radke, J. 93-1 34. Wilkins, ed. 8, Baltimore, 1974). C. G. Walker, and M. R. Walter), supported by 2. J. W. Schopf and B. M. Packer, Abstr. 5th Meet. 18. T. V. Desikachary, Cyanophyta (Indian Council National Aeronautics and Space Administration Int. Soc. Study Origin Life, 163 (1986); Science Agricultural Research, New Delhi, 1959). (NASA) grant NAGW-825 (to the PPRG), and B. M. 237, 70 (1987). 19. A. H. Knoll, S. Rossi, P. K. Strother, Precambrian Packer for the fieldwork of 1986, supported by 3. J. W. Schopf, in The Proterozoic Biosphere, J. W. Res. 38, 257 (1988). NASA grant NGR 05-007-407 (to J.W.S.). Laboratory Schopf and C. Klein, Eds. (Cambridge Univ. 20. H. J. Hofmann, J. Paleontol. 50,1040 (1976). studies were supported by National Science Foun- Press, New York, 1992), pp. 25-39. 21. E. S. Barghoorn and S. A. Tyler, Science 147, 563 dation grant BSR 86-13583 and NASA grant NAGW- 4. __ and M. R. Walter, in (1), pp. 214-239. (1965). 2147. Helpful reviews of this paper were provided 5. J. W. Schopf, in (3), pp. 179-183. 22. C. V. Mendelson and J. W. Schopf, in (3), pp. by S. M. Awramik,J. K. Bartley, S. Bengtson, T. R. 6. S. M. Awramik, J. W. Schopf, M. R. Walter, Pre- 865-951. Fairchild,A. N. Glazer, H. J. Hofmann, R. J. Horod- cambrian Res. 20, 357 (1983). 23. J. W. Schopf, in ibid., pp. 195-218. yski, A. H. Knoll,C. Marshall,C. V. Mendelson, T. B. 7. H. D. Pflug, Univ. Witwatersrand Econ. Geol. Res. 24. D. M. Ward, J. Bauld, R. W. Castenholz, B. K. Moore, B. Runnegar, E. Schultes, J. Shen-Miller,K. Unit Info. Circ. 28 (University of the Witwa- Pierson, in ibid., pp. 309-324. M. Towe, and one anonymous reviewer. tersrand, Johannesburg, South Africa, 1966), pp. 25. J. W. Schopf, in Early Life on Earth, S. Bengtson, 1-14; __ and E. Reitz, in Early Organic Evo- Ed. (Columbia Univ. Press, New York, in press). 17 December 1992; accepted 12 March 1993

646 SCIENCE * VOL. 260 * 30 APRIL 1993