ASTROBIOLOGYVolume 3 Number 2 2003copy Mary Ann Liebert Inc

Research Paper

Morphological Biosignatures and the Search for Life on Mars

SHERRY L CADY1 JACK D FARMER2 JOHN P GROTZINGER3 J WILLIAM SCHOPF4

and ANDREW STEELE5

ABSTRACT

This report provides a rationale for the advances in instrumentation and understandingneeded to assess claims of ancient and extraterrestrial life made on the basis of morphologi-cal biosignatures Morphological biosignatures consist of bona fide microbial fossils as wellas microbially influenced sedimentary structures To be recognized as evidence of life mi-crobial fossils must contain chemical and structural attributes uniquely indicative of micro-bial cells or cellular or extracellular processes When combined with various research strate-gies high-resolution instruments can reveal such attributes and elucidate how morphologicalfossils form and become altered thereby improving the ability to recognize them in the ge-ological record on Earth or other planets Also before fossilized microbially influenced sed-imentary structures can provide evidence of life criteria to distinguish their biogenic fromnon-biogenic attributes must be established This topic can be advanced by developingprocess-based models A database of images and spectroscopic data that distinguish the suiteof bona fide morphological biosignatures from their abiotic mimics will avoid detection offalse-positives for life The use of high-resolution imaging and spectroscopic instruments inconjunction with an improved knowledge base of the attributes that demonstrate life willmaximize our ability to recognize and assess the biogenicity of extraterrestrial and ancientterrestrial life Key Words BiosignaturesmdashMorphological fossilsmdashMicrofossilsmdashStromato-litesmdashExtraterrestrial lifemdashBiogenicitymdashPaleobiology Astrobiology 3 351ndash368

351

INTRODUCTION

ON EARTH TWO KINDS of morphological fossilshave been used to reveal traces of microbial

life in the ancient rock record cellularly pre-served microorganisms and laminated micro-

bially influenced sedimentary structures knownas stromatolites In order for morphological fos-sils to provide definitive evidence for life theymust be characterized by attributes that areuniquely produced by microorganisms and rec-ognizable as such Microfossils and microbially

1Department of Geology Portland State University Portland Oregon2Department of Geological Sciences Arizona State University Tempe Arizona3Massachusetts Institute of Technology Cambridge Massachusetts4University of California at Los Angeles Los Angeles California5Carnegie Institute of Geophysical Research Washington DC

influenced sedimentary structures represent twoof the three principal categories of microbialbiosignatures Chemofossils represent the thirdcategory Organic chemofossils include molecu-lar biomarkers that can be assigned to a particu-lar biosynthetic origin and the biologically fractionated isotope signature encoded duringbiosynthesis in organic compounds Inorganicchemofossils include biominerals their isotopicsignatures and anomalous concentrations or de-pletions of elements in biominerals and biologi-cally influenced sediments

Morphological fossils that retain carbonaceousremains of microbial cells (ie bona fide micro-fossils) may contain primary biomolecules or dia-genetically altered biomolecules known as bio-marker compounds (eg Summons et al 2003)Primary and diagenetically altered carbonaceousbiomolecules are usually characterized by distinc-tive carbon isotope signatures due to biologicalfractionation and preference of the light carbonisotope during metabolism and cell repair (egDes Marais et al 2003) Minerals associated withmorphological fossils may display distinctive mor-phologies isotope signatures chemical composi-tions or defect microstructures that can revealtheir biological origin (eg Banfield et al 2001)The study of how biologically produced organo-metalloids could serve as biosignatures is in its in-fancy

Morphological fossils can consist entirely ofnon-organic constituents if during fossilizationa microbial cell was completely replaced by min-erals However attempts to decipher the most an-cient fossil record on Earth have shown that min-eral-replaced morphological fossils cannot berelied upon to provide definitive evidence for lifesimply because they cannot be distinguishedfrom non-biologically produced pseudofossilsPaleobiologists searching for the earliest signs oflife on Earth and astrobiologists searching for ev-idence of extraterrestrial life face the same chal-lengemdashdistinguishing microfossils and micro-bially influenced sedimentary structures fromnon-biologically produced pseudofossils andstromatolite-like structures

Mars may prove to be the first extraterrestrialbody in our solar system to yield demonstrableevidence for the existence of life beyond Earth Inorder for a morphological fossil to be consideredpositive proof of life on Mars it must be demon-strated that the object is of biological origin andformed from Mars constituents These two crite-

ria demonstrating biogenicity and indigenicitymust be proven regardless of whether the objectis found in a martian meteorite or a targeted sam-ple returned to Earth from Mars

Demonstrating a putative fossilrsquos biogenicityrequires that the object display characteristicsuniquely attributable to microbial cells or cellu-larextracellular microbial processes The ongo-ing controversy regarding the claim of microfos-sils in martian meteorite ALH84001 exemplifiesthe difficulties encountered in proving biogenic-ity on the basis of morphological characters that are not unique to microfossils Indeed theALH84001 controversy underscores the need tobe able to distinguish the biogenically producedcharacteristics of morphological microfossilsfrom those produced non-biologically It is equallyimportant to know how to distinguish betweenbiologically and non-biologically produced stro-matolites

Proving the indigenicity of a morphologicalfossil requires demonstrating that the rock spec-imen within which it was found has not been contaminated on Mars or Earth by terrestrial mi-croorganisms (eg Steele et al 2003) Contami-nation can occur very quickly in meteorites upontheir arrival Contamination can also occur at anytime during sample collection analysis and stor-age of specimens on Mars as well as during trans-port to Earth and in terrestrial laboratories

Research strategies that improve our under-standing of the processes by which all types ofmorphological fossils are formed altered and ul-timately destroyed will facilitate a reliable as-sessment of their biogenicity and indigenicity Ofutmost importance are research strategies thatquantify the fundamental processes governingthe formation of microbial fossils and microbiallyinfluenced sedimentary structures and their abi-otic mimics within the range of conditions foundin ancient terrestrial and extraterrestrial environ-ments on early Earth and Mars It is equally im-portant to determine the constraints imposed byphysical and chemical processes that alter and de-grade ancient microbial biosignatures during di-agenesis on Earth and to establish stringent criteria for demonstrating the biogenicity of mor-phological fossils over the entire range of obser-vational scales

A conceptual framework that provides a ratio-nale for the types of advances in instrumentationand research needed to detect and interpret mor-phological fossils requires an understanding of

CADY ET AL352

(1) the use of terrestrial-based strategies to locatepotential paleobiological repositories on Mars (2)the limitations of existing criteria for interpretingthe biogenicity of fossil evidence (3) the pro-cesses that produce alter and ultimately destroymorphological fossils (4) the need for spatiallyintegrated studies to locate and interpret biosig-natures across a wide range of spatial scales (5)the need for a central database of images andspectroscopic data from biogenic and pseudo-biogenic morphological fossils and (6) the chal-lenges faced when using morphological fossils asproxies for extraterrestrial life A select numberof examples of the types of high-resolution imag-ing and spectroscopic methods needed to detectmultiple biosignatures from morphological fos-sils are provided along with a summary of rec-ommendations for instrument development andresearch areas that need to be pursued in orderto advance the search for extraterrestrial life aswell as the search for ancient life on Earth

LOCATING POTENTIALPALEOBIOLOGICAL REPOSITORIES

Terrestrial strategies for locating potential pale-obiological repositories on Mars are based uponthe strategies used by paleobiologists to studyEarthrsquos ancient fossil record of microbial life Lo-cate rock deposits that accumulated in environ-ments where fossilization of the extensive microbial communities that inhabited them wasfavored In a terrestrial context the most informa-tive assemblages of organically preserved ancientmicrofossils are found in mid-to-late Precambriansediments deposited in marine and lacustrine en-vironments Microbial fossils occur either as three-dimensional cellularly preserved (permineralized)forms embedded in an authigenic mineral matrixof stable mineralogy or as two-dimensional com-pressed or flattened forms (acritarchs) preservedin fine-grained typically clay-rich detrital sedi-ments Microbial permineralization commonly oc-curred in ancient shallow evaporative peritidal set-tings characterized by elevated salinity (eg Knoll1985) Acritarchs were preserved in ancient deepanaerobic basins especially where early diageneticmineralization occurred because of the precipita-tion of authigenic cements composed of silica car-bonate phosphate or clays

The types of rocks and paleoenvironments thathave the highest potential to capture and pre-

serve fossil biosignatures on Mars have recentlybeen reviewed by Farmer and Des Marais (1999)Paleoenvironment types include (1) mineralizingsprings (eg sinter-depositing thermal springs involcanic terrains and tufa-depositing cold springsin alkaline lake settings) (2) evaporite basins(eg terminal lake basins impact crater and vol-canic crater paleolakes and arid shorelines whereevaporite deposits inclusive of carbonates areformed) (3) mineralizing soils (eg surface du-racrust and subsoil hard pans that deposit sil-cretes calcretes and ferricretes) (4) subsurfacesedimentary systems (eg aquifers of volumi-nous extent that can sustain mineralization overa broad range of temperatures) and (5) per-mafrost and ground-ice (eg frozen soils thatpreserve microbial biosignatures in ice) In eachtype of deposit except perhaps ice-dominatedecosystems that experience repeated freeze-thaw-ing events morphological fossils could form overa wide range of spatial scales

The potential for long-term fossil preservationin each type of paleoenvironment on Mars de-pends upon the stability of the primary mineralassemblage and the amount of weathering anddiagenesis the primary deposit has endured overtime Primary aqueous mineral precipitates aremetastable their formation is kinetically notthermodynamically favored Given enough timesuch primary mineral assemblages will transformto thermodynamically favored mineral assem-blages The degree to which primary mineral as-semblages are affected by weathering and diage-netic alteration depends upon the length of timethe deposit is exposed to weathering at the sur-face (ie chemical and mechanical weatheringprocesses) and to fluids that could promote dia-genetically induced mineral phase transforma-tions Since primary aqueous mineral precipitatesoften entomb or permeate microbial cells and mi-crobially influenced sedimentary structures sub-sequent diagenesis could significantly alter pri-mary biosignatures On Mars the amount ofwater and the length of time water might havereacted with primary and secondary minerals ina sedimentary deposit are critical factors thatmust be considered when estimating the degreeto which diagenesis may have altered the deposit

On Earth the potential for long-term preser-vation also depends upon the amount of tectonicrecycling of the rocks in which the fossils are pre-served While the amount of tectonic recyclingthat mineral deposits would have experienced on

MORPHOLOGICAL BIOSIGNATURES 353

Mars is assumed to be negligible the degree towhich any tectonic-scale forces may have alteredMars rocks is not known

In general long-term preservation of microbialbiosignatures has occurred when fossils and sed-imented or precipitated biofabrics and structureswere retained in dense impermeable host rockscomposed of stable minerals that resisted chem-ical weathering dissolution or extensive reorga-nization during diagenetic recrystallization Fa-vored lithologies for long-term preservationinclude cherts and phosphorites rocks that con-tain silica and phosphate respectively Suchlithologies have long crustal residence times andalong with carbonates and shales are the mostcommon host rocks for the Precambrian micro-fossil record on Earth

Determining the location of potential paleobi-ological repositories on Mars requires an under-standing of the martian surface in terms of ele-mental abundances and mineralogy Progress inunderstanding martian surface mineralogy hasrecently been provided by data obtained by theThermal Emission Spectrometer (TES) experi-ment carried aboard the Mars Global SurveyorOrbiter (Christensen et al 2000) TES whichmaps at a spatial resolution of 3 kmpixel in themid-infrared portion of the spectrum (6ndash14 mm)detected large deposits of coarse-grained (specu-lar) hematite (iron oxide) at several locations onMars This variety of hematite on Earth formsonly in the presence of large amounts of waterand typically at elevated (hydrothermal) temper-atures (Christensen et al 2000) The hematite de-posits appear to be co-located with geomorphicfeatures previously described as having beenformed by the action of liquid water at the sur-face of Mars TES has also revealed a basic compositional difference between the southerncratered highlands of Mars (modal compositiondominantly basalt) and the northern lowlandplains [modal composition more silica-enriched(ie basaltic-andesite)] (Bandfield et al 2000)The significance of this composition difference isunclear since the geological processes responsi-ble for them are unknown

The search on Mars for common aqueouslyprecipitated mineral assemblages and rock types(eg carbonates cherts phosphorites and evap-orites) that tend to harbor fossilized microbial as-semblages on Earth is still underway Remotesensing analog studies in Death Valley (Jolliff etal 2001 Moersch et al 2001) indicate that detec-

tion of these minerals by TES may not be possi-ble at lower abundances because of the low spa-tial resolution of the instrument Thus futuremissions will need to consider higher-resolutioninstruments that can map over a broader rangeof wavelengths and at spatial resolutions 100mpixel Improved spatial resolution (100 mpixel) is being obtained with the Thermal Emis-sion Imaging System (THEMIS) on the MarsOdyssey spacecraft presently orbiting Mars Fullcoverage of the planet will be possible at this spa-tial resolution during the nominal mission How-ever the spectral resolution of this mid-infraredmapping spectrometer will be significantly lowerthan TES and it is unclear what impact this willhave on mineral detection limits at Mars In 2005the Mars Reconnaissance Orbiter (MRO) willcarry a near-infrared spectrometer (Compact Re-connaissance Imaging Spectrometer for Mars orCRISM) which will map the surface at resolu-tions as low as 10 mpixel While this hyper-spectral instrument will not achieve completecoverage of the martian surface it will providehigh-resolution mapping at targeted sites

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITYSTROMATOLITES AND RELATED

MICROBIALLY INFLUENCEDBIOFABRICS AND SEDIMENTARY

STRUCTURES

Benthic microbial communities are likely to in-fluence the microstructure of deposits that accu-mulate in their presence regardless of whetherthey occur in detrital or mineralizing environ-ments The key to proving the biogenicity of mi-crobially influenced biofabrics and structures isknowing which of their components resultedfrom the presence or influence of the organismsand whether these attributes are unique to mi-crobial processes If the attributes can form abi-otically in the absence of microorganisms thensuch attributes cannot be used to demonstratebiogenicity

To provide a framework for assessing the rel-ative roles of different biologic and abiologicprocesses in stromatolite accretion we adopt anon-genetic definition of stromatolites first pre-sented in Hofmann (1973) A stromatolite is ldquoan attached laminated lithified sedimentarygrowth structure accretionary away from a point

CADY ET AL354

or limited surface of initiationrdquo (Semikhatov et al1979) This concise definition describes the basicgeometric and textural properties of all stroma-tolites while at the same time allowing for mul-tiple or even indeterminate origins In this wayan objective evaluation can be made of the vari-ous processes that may influence stromatolite de-velopment Such an approach can lead to the formulation of specific process models that accu-rately describe stromatolite accretion dynamicsProcess models might routinely predict the rela-tive or even absolute contributions of biologicalphysical and chemical effects to stromatolite for-mation

The future value of stromatolite research lies inits potential to provide a basis for reconstructingancient environments and to understand howbenthic microbial communities interacted withtheir environments This must involve a process-oriented approach to stromatolite morphogenesisin which the correct interpretation of diageneticand recrystallization textures is as important asunderstanding microbial diversity and fossiliza-tion processes in modern microbial ecosystemsThe goal is to build an understanding of stroma-tolite development that stems from rigorousquantitative analyses of stromatolite form andlamina texture including the deconvolution of di-agenetic overprints to reveal primary fabrics di-agnostic of specific microbial and sedimentaryprocesses Critical questions regarding the factorsresponsible for the development of biofabricsbiogenic stromatolite morphologies and diage-netic textures can then be addressed For exam-ple over what length and time scales do biolog-ical physical and chemical processes operateDo any of these processesmdashwhich might be crit-ical at microscopic scalesmdashremain sensitive atlarger scales If not at what scale does the tran-sition in process response take place Questionssuch as these must be answered before we canclaim a real understanding of what fundamentalproperties such as stromatolite shape signify

The most conspicuous feature of stromatolitesis their lamination Individual laminae are thebuilding blocks of stromatolites and thereforemake up a time series of progressive albeit in-cremental accretion The morphology of anystromatolite is a function of how lamina shapeparticularly its relief evolves over time Topo-graphic anomalies that are reinforced over timegive rise to greater relief for successive laminaeand those that are stabilized give rise to greater

inheritance of shape for successive laminaeThose topographic anomalies that are dampedover time result in diminished relief

In addition to lamination the other distin-guishing feature of stromatolites is their shape Atypical stromatolite is made up of numerous suc-cessive laminae that stack one on top of the otherto form domal coniform columnar or branchingcolumnar structures Although laminae generallydescribe convex-upward structures they can alsoform concave-upward or discrete conical struc-tures In general it is thought that there is a broadbut gradational variation in the form of stromato-lites encompassing several major morphologicalmotifs (Semikhatov et al 1979) It has long beenobserved that stromatolite morphology varies as afunction of facies thus there is broad agreementthat physical environment plays a role in the gen-eration of shape (see papers in Walter 1976)

At the level of process however there is nosuch guiding consensus which severely limitsour ability to understand either paleoenviron-mental or stratigraphic variations in stromatoliteform As Hofmann (1987) stated ldquo we stillhave no stromatolite theory no model that showswhich attributes changed in what way throughtimerdquo Without a viable theory we are always atrisk of misinterpreting the genetic significance ofgrowth form

One can easily list numerous factors that mightinfluence stromatolite development includinglight intensity salinity nutrient supply currentvelocity sediment grain size distribution matcommunity diversity and degree of mineral sat-uration to name a few In detail stromatolitegrowth is dependent on many processes that are complexly interrelated Not only are theprocesses mechanistically complex they also evolveover long time scales that are difficult to repro-duce experimentally or monitor in the fieldGiven these difficulties our initial goal should beto construct simple process models of complexsystems Even this will be a difficult task in needof studies that can serve to calibrate importantmodel parameters (eg Jorgensen and DesMarais 1990) In principle the growth of stro-matolites can be described as a simple system thatdepends on only three fundamental processesgrowth and degradation of a microbial mat orbiofilm deposition of sediment and precipitationof minerals Interactions among these end-mem-ber processes should account for the bulk of stro-matolites in the record

MORPHOLOGICAL BIOSIGNATURES 355

Stromatolite growth depends on the iterativeprocess of upward growth by mats or seafloorcrusts alternating with times of sediment deposi-tion In addition these processes must be bal-anced in such a way that sediment does not over-whelm mats or microbial biofilms On furtherinspection an additional but critically importantattribute of the iterative process is revealed Thegrowth of mats tends to produce an irregular rel-atively rough surface whereas the settling of sed-iment tends to create a relatively smooth surfaceby filling in the microtopography of the under-lying mat The surface roughness of mats willvary depending on the composition and distrib-ution of the microbial community Thereforeover many iterations the surface roughness of agrowing stromatolite may be enhanced or sup-pressed depending on the competitive processesof surface roughening by mats and surfacesmoothing through sedimentation Unfortunatelyeven though it has been recognized in a qualita-tive sense for decades that microbial mats havevariable surface roughness this attribute hasnever been quantified It is recommended thatsurface roughness be measured in future studiesof modern microbial mats particularly wheresedimentation also occurs

Interface dynamics and stromatolites

The texture (ie preferred orientation) androughness of any depositional surface or interfaceis subject to certain force balances and the pres-ence of noise or randomness A widely appliedgrowth model is represented by the KPZ equa-tion (Kardar et al 1986) whose relevance to un-derstanding stromatolite growth was recentlyevaluated (Grotzinger and Rothman 1996) Thevalidity of the KPZ equation in accounting for the growth of stromatolites was tested and ten-tatively confirmed by calculating the surfaceroughness of several stromatolitic laminae andcomparing the obtained scaling exponent (andfractal dimension) with that predicted by the KPZtheory (Grotzinger and Rothman 1996) Growthof this type may characterize in a quantitativemanner the geometry of layering in many Pre-cambrian stromatolites agates botryoidal min-eral clusters travertines and certain types of stro-matolites where seafloor precipitation is thoughtto have been important For example stromato-lites formed immediately prior to the precipita-tion of some of the worldrsquos largest evaporite de-

posits are characterized by fine isopachous lam-ination and internal textures consistent with insitu precipitation (Pope and Grotzinger 1999)

In contrast for other systems the presence of adiffusing field that reflects pressure electric po-tential temperature and chemical or nutrientconcentrations may lead to a more appropriategrowth model for stromatolites since growth isfastest where concentration gradients are steep-est Initial studies of models for diffusion-limitedaggregation (DLA) strongly suggest that they arealso applicable to understanding the growth of certain stromatolites (Grotzinger and Chan1999) Whereas DLA by itself predicts highlybranched dendritic structures in the presence ofincremental sedimentation events a simple DLAmodel predicts many of the domal columnarand branching columnar stromatolite morpholo-gies observed in the geological record In thismodel an episode of upward growth by ran-domly attaching particles is taken to simulategrowth of either mats or crystals followed by anepisode of sediment settling in which the sedi-ment is allowed to settle preferentially in the mi-crodepressions formed in the underlying mat orcrystal layer If thick enough andor diffusiveenough the sediment may damp all of the initialtopography created by the underlying layerHowever if antecedent topography remainsthen the next iteration of matcrystal growth willresult in preferential accretion on those topo-graphic highs The next layer of sediment nowhas a tougher task to fill depressions giving risein the next iteration of matcrystal growth to aneven higher preference for growth on topo-graphical highs This is a particularly importantfeature of DLA models in that small perturba-tions can be amplified in time to become domi-nant features of the structure itself In this man-ner no special conditions may be required togenerate columns and branching columns in stro-matolitesmdashonly time and the positive reinforce-ment of randomly produced protuberances Thistype of growth may help account for the diver-sity of branched columnar stromatolites commonin the ancient geologic record

Biogenic versus abiogenic growth

By understanding through the use of simpleprocess models that stromatolite growth mayresult from the competitive interaction of up-ward-growth and surface roughness forced by

CADY ET AL356

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

influenced sedimentary structures represent twoof the three principal categories of microbialbiosignatures Chemofossils represent the thirdcategory Organic chemofossils include molecu-lar biomarkers that can be assigned to a particu-lar biosynthetic origin and the biologically fractionated isotope signature encoded duringbiosynthesis in organic compounds Inorganicchemofossils include biominerals their isotopicsignatures and anomalous concentrations or de-pletions of elements in biominerals and biologi-cally influenced sediments

Morphological fossils that retain carbonaceousremains of microbial cells (ie bona fide micro-fossils) may contain primary biomolecules or dia-genetically altered biomolecules known as bio-marker compounds (eg Summons et al 2003)Primary and diagenetically altered carbonaceousbiomolecules are usually characterized by distinc-tive carbon isotope signatures due to biologicalfractionation and preference of the light carbonisotope during metabolism and cell repair (egDes Marais et al 2003) Minerals associated withmorphological fossils may display distinctive mor-phologies isotope signatures chemical composi-tions or defect microstructures that can revealtheir biological origin (eg Banfield et al 2001)The study of how biologically produced organo-metalloids could serve as biosignatures is in its in-fancy

Morphological fossils can consist entirely ofnon-organic constituents if during fossilizationa microbial cell was completely replaced by min-erals However attempts to decipher the most an-cient fossil record on Earth have shown that min-eral-replaced morphological fossils cannot berelied upon to provide definitive evidence for lifesimply because they cannot be distinguishedfrom non-biologically produced pseudofossilsPaleobiologists searching for the earliest signs oflife on Earth and astrobiologists searching for ev-idence of extraterrestrial life face the same chal-lengemdashdistinguishing microfossils and micro-bially influenced sedimentary structures fromnon-biologically produced pseudofossils andstromatolite-like structures

Mars may prove to be the first extraterrestrialbody in our solar system to yield demonstrableevidence for the existence of life beyond Earth Inorder for a morphological fossil to be consideredpositive proof of life on Mars it must be demon-strated that the object is of biological origin andformed from Mars constituents These two crite-

ria demonstrating biogenicity and indigenicitymust be proven regardless of whether the objectis found in a martian meteorite or a targeted sam-ple returned to Earth from Mars

Demonstrating a putative fossilrsquos biogenicityrequires that the object display characteristicsuniquely attributable to microbial cells or cellu-larextracellular microbial processes The ongo-ing controversy regarding the claim of microfos-sils in martian meteorite ALH84001 exemplifiesthe difficulties encountered in proving biogenic-ity on the basis of morphological characters that are not unique to microfossils Indeed theALH84001 controversy underscores the need tobe able to distinguish the biogenically producedcharacteristics of morphological microfossilsfrom those produced non-biologically It is equallyimportant to know how to distinguish betweenbiologically and non-biologically produced stro-matolites

Proving the indigenicity of a morphologicalfossil requires demonstrating that the rock spec-imen within which it was found has not been contaminated on Mars or Earth by terrestrial mi-croorganisms (eg Steele et al 2003) Contami-nation can occur very quickly in meteorites upontheir arrival Contamination can also occur at anytime during sample collection analysis and stor-age of specimens on Mars as well as during trans-port to Earth and in terrestrial laboratories

Research strategies that improve our under-standing of the processes by which all types ofmorphological fossils are formed altered and ul-timately destroyed will facilitate a reliable as-sessment of their biogenicity and indigenicity Ofutmost importance are research strategies thatquantify the fundamental processes governingthe formation of microbial fossils and microbiallyinfluenced sedimentary structures and their abi-otic mimics within the range of conditions foundin ancient terrestrial and extraterrestrial environ-ments on early Earth and Mars It is equally im-portant to determine the constraints imposed byphysical and chemical processes that alter and de-grade ancient microbial biosignatures during di-agenesis on Earth and to establish stringent criteria for demonstrating the biogenicity of mor-phological fossils over the entire range of obser-vational scales

A conceptual framework that provides a ratio-nale for the types of advances in instrumentationand research needed to detect and interpret mor-phological fossils requires an understanding of

CADY ET AL352

(1) the use of terrestrial-based strategies to locatepotential paleobiological repositories on Mars (2)the limitations of existing criteria for interpretingthe biogenicity of fossil evidence (3) the pro-cesses that produce alter and ultimately destroymorphological fossils (4) the need for spatiallyintegrated studies to locate and interpret biosig-natures across a wide range of spatial scales (5)the need for a central database of images andspectroscopic data from biogenic and pseudo-biogenic morphological fossils and (6) the chal-lenges faced when using morphological fossils asproxies for extraterrestrial life A select numberof examples of the types of high-resolution imag-ing and spectroscopic methods needed to detectmultiple biosignatures from morphological fos-sils are provided along with a summary of rec-ommendations for instrument development andresearch areas that need to be pursued in orderto advance the search for extraterrestrial life aswell as the search for ancient life on Earth

LOCATING POTENTIALPALEOBIOLOGICAL REPOSITORIES

Terrestrial strategies for locating potential pale-obiological repositories on Mars are based uponthe strategies used by paleobiologists to studyEarthrsquos ancient fossil record of microbial life Lo-cate rock deposits that accumulated in environ-ments where fossilization of the extensive microbial communities that inhabited them wasfavored In a terrestrial context the most informa-tive assemblages of organically preserved ancientmicrofossils are found in mid-to-late Precambriansediments deposited in marine and lacustrine en-vironments Microbial fossils occur either as three-dimensional cellularly preserved (permineralized)forms embedded in an authigenic mineral matrixof stable mineralogy or as two-dimensional com-pressed or flattened forms (acritarchs) preservedin fine-grained typically clay-rich detrital sedi-ments Microbial permineralization commonly oc-curred in ancient shallow evaporative peritidal set-tings characterized by elevated salinity (eg Knoll1985) Acritarchs were preserved in ancient deepanaerobic basins especially where early diageneticmineralization occurred because of the precipita-tion of authigenic cements composed of silica car-bonate phosphate or clays

The types of rocks and paleoenvironments thathave the highest potential to capture and pre-

serve fossil biosignatures on Mars have recentlybeen reviewed by Farmer and Des Marais (1999)Paleoenvironment types include (1) mineralizingsprings (eg sinter-depositing thermal springs involcanic terrains and tufa-depositing cold springsin alkaline lake settings) (2) evaporite basins(eg terminal lake basins impact crater and vol-canic crater paleolakes and arid shorelines whereevaporite deposits inclusive of carbonates areformed) (3) mineralizing soils (eg surface du-racrust and subsoil hard pans that deposit sil-cretes calcretes and ferricretes) (4) subsurfacesedimentary systems (eg aquifers of volumi-nous extent that can sustain mineralization overa broad range of temperatures) and (5) per-mafrost and ground-ice (eg frozen soils thatpreserve microbial biosignatures in ice) In eachtype of deposit except perhaps ice-dominatedecosystems that experience repeated freeze-thaw-ing events morphological fossils could form overa wide range of spatial scales

The potential for long-term fossil preservationin each type of paleoenvironment on Mars de-pends upon the stability of the primary mineralassemblage and the amount of weathering anddiagenesis the primary deposit has endured overtime Primary aqueous mineral precipitates aremetastable their formation is kinetically notthermodynamically favored Given enough timesuch primary mineral assemblages will transformto thermodynamically favored mineral assem-blages The degree to which primary mineral as-semblages are affected by weathering and diage-netic alteration depends upon the length of timethe deposit is exposed to weathering at the sur-face (ie chemical and mechanical weatheringprocesses) and to fluids that could promote dia-genetically induced mineral phase transforma-tions Since primary aqueous mineral precipitatesoften entomb or permeate microbial cells and mi-crobially influenced sedimentary structures sub-sequent diagenesis could significantly alter pri-mary biosignatures On Mars the amount ofwater and the length of time water might havereacted with primary and secondary minerals ina sedimentary deposit are critical factors thatmust be considered when estimating the degreeto which diagenesis may have altered the deposit

On Earth the potential for long-term preser-vation also depends upon the amount of tectonicrecycling of the rocks in which the fossils are pre-served While the amount of tectonic recyclingthat mineral deposits would have experienced on

MORPHOLOGICAL BIOSIGNATURES 353

Mars is assumed to be negligible the degree towhich any tectonic-scale forces may have alteredMars rocks is not known

In general long-term preservation of microbialbiosignatures has occurred when fossils and sed-imented or precipitated biofabrics and structureswere retained in dense impermeable host rockscomposed of stable minerals that resisted chem-ical weathering dissolution or extensive reorga-nization during diagenetic recrystallization Fa-vored lithologies for long-term preservationinclude cherts and phosphorites rocks that con-tain silica and phosphate respectively Suchlithologies have long crustal residence times andalong with carbonates and shales are the mostcommon host rocks for the Precambrian micro-fossil record on Earth

Determining the location of potential paleobi-ological repositories on Mars requires an under-standing of the martian surface in terms of ele-mental abundances and mineralogy Progress inunderstanding martian surface mineralogy hasrecently been provided by data obtained by theThermal Emission Spectrometer (TES) experi-ment carried aboard the Mars Global SurveyorOrbiter (Christensen et al 2000) TES whichmaps at a spatial resolution of 3 kmpixel in themid-infrared portion of the spectrum (6ndash14 mm)detected large deposits of coarse-grained (specu-lar) hematite (iron oxide) at several locations onMars This variety of hematite on Earth formsonly in the presence of large amounts of waterand typically at elevated (hydrothermal) temper-atures (Christensen et al 2000) The hematite de-posits appear to be co-located with geomorphicfeatures previously described as having beenformed by the action of liquid water at the sur-face of Mars TES has also revealed a basic compositional difference between the southerncratered highlands of Mars (modal compositiondominantly basalt) and the northern lowlandplains [modal composition more silica-enriched(ie basaltic-andesite)] (Bandfield et al 2000)The significance of this composition difference isunclear since the geological processes responsi-ble for them are unknown

The search on Mars for common aqueouslyprecipitated mineral assemblages and rock types(eg carbonates cherts phosphorites and evap-orites) that tend to harbor fossilized microbial as-semblages on Earth is still underway Remotesensing analog studies in Death Valley (Jolliff etal 2001 Moersch et al 2001) indicate that detec-

tion of these minerals by TES may not be possi-ble at lower abundances because of the low spa-tial resolution of the instrument Thus futuremissions will need to consider higher-resolutioninstruments that can map over a broader rangeof wavelengths and at spatial resolutions 100mpixel Improved spatial resolution (100 mpixel) is being obtained with the Thermal Emis-sion Imaging System (THEMIS) on the MarsOdyssey spacecraft presently orbiting Mars Fullcoverage of the planet will be possible at this spa-tial resolution during the nominal mission How-ever the spectral resolution of this mid-infraredmapping spectrometer will be significantly lowerthan TES and it is unclear what impact this willhave on mineral detection limits at Mars In 2005the Mars Reconnaissance Orbiter (MRO) willcarry a near-infrared spectrometer (Compact Re-connaissance Imaging Spectrometer for Mars orCRISM) which will map the surface at resolu-tions as low as 10 mpixel While this hyper-spectral instrument will not achieve completecoverage of the martian surface it will providehigh-resolution mapping at targeted sites

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITYSTROMATOLITES AND RELATED

MICROBIALLY INFLUENCEDBIOFABRICS AND SEDIMENTARY

STRUCTURES

Benthic microbial communities are likely to in-fluence the microstructure of deposits that accu-mulate in their presence regardless of whetherthey occur in detrital or mineralizing environ-ments The key to proving the biogenicity of mi-crobially influenced biofabrics and structures isknowing which of their components resultedfrom the presence or influence of the organismsand whether these attributes are unique to mi-crobial processes If the attributes can form abi-otically in the absence of microorganisms thensuch attributes cannot be used to demonstratebiogenicity

To provide a framework for assessing the rel-ative roles of different biologic and abiologicprocesses in stromatolite accretion we adopt anon-genetic definition of stromatolites first pre-sented in Hofmann (1973) A stromatolite is ldquoan attached laminated lithified sedimentarygrowth structure accretionary away from a point

CADY ET AL354

or limited surface of initiationrdquo (Semikhatov et al1979) This concise definition describes the basicgeometric and textural properties of all stroma-tolites while at the same time allowing for mul-tiple or even indeterminate origins In this wayan objective evaluation can be made of the vari-ous processes that may influence stromatolite de-velopment Such an approach can lead to the formulation of specific process models that accu-rately describe stromatolite accretion dynamicsProcess models might routinely predict the rela-tive or even absolute contributions of biologicalphysical and chemical effects to stromatolite for-mation

The future value of stromatolite research lies inits potential to provide a basis for reconstructingancient environments and to understand howbenthic microbial communities interacted withtheir environments This must involve a process-oriented approach to stromatolite morphogenesisin which the correct interpretation of diageneticand recrystallization textures is as important asunderstanding microbial diversity and fossiliza-tion processes in modern microbial ecosystemsThe goal is to build an understanding of stroma-tolite development that stems from rigorousquantitative analyses of stromatolite form andlamina texture including the deconvolution of di-agenetic overprints to reveal primary fabrics di-agnostic of specific microbial and sedimentaryprocesses Critical questions regarding the factorsresponsible for the development of biofabricsbiogenic stromatolite morphologies and diage-netic textures can then be addressed For exam-ple over what length and time scales do biolog-ical physical and chemical processes operateDo any of these processesmdashwhich might be crit-ical at microscopic scalesmdashremain sensitive atlarger scales If not at what scale does the tran-sition in process response take place Questionssuch as these must be answered before we canclaim a real understanding of what fundamentalproperties such as stromatolite shape signify

The most conspicuous feature of stromatolitesis their lamination Individual laminae are thebuilding blocks of stromatolites and thereforemake up a time series of progressive albeit in-cremental accretion The morphology of anystromatolite is a function of how lamina shapeparticularly its relief evolves over time Topo-graphic anomalies that are reinforced over timegive rise to greater relief for successive laminaeand those that are stabilized give rise to greater

inheritance of shape for successive laminaeThose topographic anomalies that are dampedover time result in diminished relief

In addition to lamination the other distin-guishing feature of stromatolites is their shape Atypical stromatolite is made up of numerous suc-cessive laminae that stack one on top of the otherto form domal coniform columnar or branchingcolumnar structures Although laminae generallydescribe convex-upward structures they can alsoform concave-upward or discrete conical struc-tures In general it is thought that there is a broadbut gradational variation in the form of stromato-lites encompassing several major morphologicalmotifs (Semikhatov et al 1979) It has long beenobserved that stromatolite morphology varies as afunction of facies thus there is broad agreementthat physical environment plays a role in the gen-eration of shape (see papers in Walter 1976)

At the level of process however there is nosuch guiding consensus which severely limitsour ability to understand either paleoenviron-mental or stratigraphic variations in stromatoliteform As Hofmann (1987) stated ldquo we stillhave no stromatolite theory no model that showswhich attributes changed in what way throughtimerdquo Without a viable theory we are always atrisk of misinterpreting the genetic significance ofgrowth form

One can easily list numerous factors that mightinfluence stromatolite development includinglight intensity salinity nutrient supply currentvelocity sediment grain size distribution matcommunity diversity and degree of mineral sat-uration to name a few In detail stromatolitegrowth is dependent on many processes that are complexly interrelated Not only are theprocesses mechanistically complex they also evolveover long time scales that are difficult to repro-duce experimentally or monitor in the fieldGiven these difficulties our initial goal should beto construct simple process models of complexsystems Even this will be a difficult task in needof studies that can serve to calibrate importantmodel parameters (eg Jorgensen and DesMarais 1990) In principle the growth of stro-matolites can be described as a simple system thatdepends on only three fundamental processesgrowth and degradation of a microbial mat orbiofilm deposition of sediment and precipitationof minerals Interactions among these end-mem-ber processes should account for the bulk of stro-matolites in the record

MORPHOLOGICAL BIOSIGNATURES 355

Stromatolite growth depends on the iterativeprocess of upward growth by mats or seafloorcrusts alternating with times of sediment deposi-tion In addition these processes must be bal-anced in such a way that sediment does not over-whelm mats or microbial biofilms On furtherinspection an additional but critically importantattribute of the iterative process is revealed Thegrowth of mats tends to produce an irregular rel-atively rough surface whereas the settling of sed-iment tends to create a relatively smooth surfaceby filling in the microtopography of the under-lying mat The surface roughness of mats willvary depending on the composition and distrib-ution of the microbial community Thereforeover many iterations the surface roughness of agrowing stromatolite may be enhanced or sup-pressed depending on the competitive processesof surface roughening by mats and surfacesmoothing through sedimentation Unfortunatelyeven though it has been recognized in a qualita-tive sense for decades that microbial mats havevariable surface roughness this attribute hasnever been quantified It is recommended thatsurface roughness be measured in future studiesof modern microbial mats particularly wheresedimentation also occurs

Interface dynamics and stromatolites

The texture (ie preferred orientation) androughness of any depositional surface or interfaceis subject to certain force balances and the pres-ence of noise or randomness A widely appliedgrowth model is represented by the KPZ equa-tion (Kardar et al 1986) whose relevance to un-derstanding stromatolite growth was recentlyevaluated (Grotzinger and Rothman 1996) Thevalidity of the KPZ equation in accounting for the growth of stromatolites was tested and ten-tatively confirmed by calculating the surfaceroughness of several stromatolitic laminae andcomparing the obtained scaling exponent (andfractal dimension) with that predicted by the KPZtheory (Grotzinger and Rothman 1996) Growthof this type may characterize in a quantitativemanner the geometry of layering in many Pre-cambrian stromatolites agates botryoidal min-eral clusters travertines and certain types of stro-matolites where seafloor precipitation is thoughtto have been important For example stromato-lites formed immediately prior to the precipita-tion of some of the worldrsquos largest evaporite de-

posits are characterized by fine isopachous lam-ination and internal textures consistent with insitu precipitation (Pope and Grotzinger 1999)

In contrast for other systems the presence of adiffusing field that reflects pressure electric po-tential temperature and chemical or nutrientconcentrations may lead to a more appropriategrowth model for stromatolites since growth isfastest where concentration gradients are steep-est Initial studies of models for diffusion-limitedaggregation (DLA) strongly suggest that they arealso applicable to understanding the growth of certain stromatolites (Grotzinger and Chan1999) Whereas DLA by itself predicts highlybranched dendritic structures in the presence ofincremental sedimentation events a simple DLAmodel predicts many of the domal columnarand branching columnar stromatolite morpholo-gies observed in the geological record In thismodel an episode of upward growth by ran-domly attaching particles is taken to simulategrowth of either mats or crystals followed by anepisode of sediment settling in which the sedi-ment is allowed to settle preferentially in the mi-crodepressions formed in the underlying mat orcrystal layer If thick enough andor diffusiveenough the sediment may damp all of the initialtopography created by the underlying layerHowever if antecedent topography remainsthen the next iteration of matcrystal growth willresult in preferential accretion on those topo-graphic highs The next layer of sediment nowhas a tougher task to fill depressions giving risein the next iteration of matcrystal growth to aneven higher preference for growth on topo-graphical highs This is a particularly importantfeature of DLA models in that small perturba-tions can be amplified in time to become domi-nant features of the structure itself In this man-ner no special conditions may be required togenerate columns and branching columns in stro-matolitesmdashonly time and the positive reinforce-ment of randomly produced protuberances Thistype of growth may help account for the diver-sity of branched columnar stromatolites commonin the ancient geologic record

Biogenic versus abiogenic growth

By understanding through the use of simpleprocess models that stromatolite growth mayresult from the competitive interaction of up-ward-growth and surface roughness forced by

CADY ET AL356

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

(1) the use of terrestrial-based strategies to locatepotential paleobiological repositories on Mars (2)the limitations of existing criteria for interpretingthe biogenicity of fossil evidence (3) the pro-cesses that produce alter and ultimately destroymorphological fossils (4) the need for spatiallyintegrated studies to locate and interpret biosig-natures across a wide range of spatial scales (5)the need for a central database of images andspectroscopic data from biogenic and pseudo-biogenic morphological fossils and (6) the chal-lenges faced when using morphological fossils asproxies for extraterrestrial life A select numberof examples of the types of high-resolution imag-ing and spectroscopic methods needed to detectmultiple biosignatures from morphological fos-sils are provided along with a summary of rec-ommendations for instrument development andresearch areas that need to be pursued in orderto advance the search for extraterrestrial life aswell as the search for ancient life on Earth

LOCATING POTENTIALPALEOBIOLOGICAL REPOSITORIES

Terrestrial strategies for locating potential pale-obiological repositories on Mars are based uponthe strategies used by paleobiologists to studyEarthrsquos ancient fossil record of microbial life Lo-cate rock deposits that accumulated in environ-ments where fossilization of the extensive microbial communities that inhabited them wasfavored In a terrestrial context the most informa-tive assemblages of organically preserved ancientmicrofossils are found in mid-to-late Precambriansediments deposited in marine and lacustrine en-vironments Microbial fossils occur either as three-dimensional cellularly preserved (permineralized)forms embedded in an authigenic mineral matrixof stable mineralogy or as two-dimensional com-pressed or flattened forms (acritarchs) preservedin fine-grained typically clay-rich detrital sedi-ments Microbial permineralization commonly oc-curred in ancient shallow evaporative peritidal set-tings characterized by elevated salinity (eg Knoll1985) Acritarchs were preserved in ancient deepanaerobic basins especially where early diageneticmineralization occurred because of the precipita-tion of authigenic cements composed of silica car-bonate phosphate or clays

The types of rocks and paleoenvironments thathave the highest potential to capture and pre-

serve fossil biosignatures on Mars have recentlybeen reviewed by Farmer and Des Marais (1999)Paleoenvironment types include (1) mineralizingsprings (eg sinter-depositing thermal springs involcanic terrains and tufa-depositing cold springsin alkaline lake settings) (2) evaporite basins(eg terminal lake basins impact crater and vol-canic crater paleolakes and arid shorelines whereevaporite deposits inclusive of carbonates areformed) (3) mineralizing soils (eg surface du-racrust and subsoil hard pans that deposit sil-cretes calcretes and ferricretes) (4) subsurfacesedimentary systems (eg aquifers of volumi-nous extent that can sustain mineralization overa broad range of temperatures) and (5) per-mafrost and ground-ice (eg frozen soils thatpreserve microbial biosignatures in ice) In eachtype of deposit except perhaps ice-dominatedecosystems that experience repeated freeze-thaw-ing events morphological fossils could form overa wide range of spatial scales

The potential for long-term fossil preservationin each type of paleoenvironment on Mars de-pends upon the stability of the primary mineralassemblage and the amount of weathering anddiagenesis the primary deposit has endured overtime Primary aqueous mineral precipitates aremetastable their formation is kinetically notthermodynamically favored Given enough timesuch primary mineral assemblages will transformto thermodynamically favored mineral assem-blages The degree to which primary mineral as-semblages are affected by weathering and diage-netic alteration depends upon the length of timethe deposit is exposed to weathering at the sur-face (ie chemical and mechanical weatheringprocesses) and to fluids that could promote dia-genetically induced mineral phase transforma-tions Since primary aqueous mineral precipitatesoften entomb or permeate microbial cells and mi-crobially influenced sedimentary structures sub-sequent diagenesis could significantly alter pri-mary biosignatures On Mars the amount ofwater and the length of time water might havereacted with primary and secondary minerals ina sedimentary deposit are critical factors thatmust be considered when estimating the degreeto which diagenesis may have altered the deposit

On Earth the potential for long-term preser-vation also depends upon the amount of tectonicrecycling of the rocks in which the fossils are pre-served While the amount of tectonic recyclingthat mineral deposits would have experienced on

MORPHOLOGICAL BIOSIGNATURES 353

Mars is assumed to be negligible the degree towhich any tectonic-scale forces may have alteredMars rocks is not known

In general long-term preservation of microbialbiosignatures has occurred when fossils and sed-imented or precipitated biofabrics and structureswere retained in dense impermeable host rockscomposed of stable minerals that resisted chem-ical weathering dissolution or extensive reorga-nization during diagenetic recrystallization Fa-vored lithologies for long-term preservationinclude cherts and phosphorites rocks that con-tain silica and phosphate respectively Suchlithologies have long crustal residence times andalong with carbonates and shales are the mostcommon host rocks for the Precambrian micro-fossil record on Earth

Determining the location of potential paleobi-ological repositories on Mars requires an under-standing of the martian surface in terms of ele-mental abundances and mineralogy Progress inunderstanding martian surface mineralogy hasrecently been provided by data obtained by theThermal Emission Spectrometer (TES) experi-ment carried aboard the Mars Global SurveyorOrbiter (Christensen et al 2000) TES whichmaps at a spatial resolution of 3 kmpixel in themid-infrared portion of the spectrum (6ndash14 mm)detected large deposits of coarse-grained (specu-lar) hematite (iron oxide) at several locations onMars This variety of hematite on Earth formsonly in the presence of large amounts of waterand typically at elevated (hydrothermal) temper-atures (Christensen et al 2000) The hematite de-posits appear to be co-located with geomorphicfeatures previously described as having beenformed by the action of liquid water at the sur-face of Mars TES has also revealed a basic compositional difference between the southerncratered highlands of Mars (modal compositiondominantly basalt) and the northern lowlandplains [modal composition more silica-enriched(ie basaltic-andesite)] (Bandfield et al 2000)The significance of this composition difference isunclear since the geological processes responsi-ble for them are unknown

The search on Mars for common aqueouslyprecipitated mineral assemblages and rock types(eg carbonates cherts phosphorites and evap-orites) that tend to harbor fossilized microbial as-semblages on Earth is still underway Remotesensing analog studies in Death Valley (Jolliff etal 2001 Moersch et al 2001) indicate that detec-

tion of these minerals by TES may not be possi-ble at lower abundances because of the low spa-tial resolution of the instrument Thus futuremissions will need to consider higher-resolutioninstruments that can map over a broader rangeof wavelengths and at spatial resolutions 100mpixel Improved spatial resolution (100 mpixel) is being obtained with the Thermal Emis-sion Imaging System (THEMIS) on the MarsOdyssey spacecraft presently orbiting Mars Fullcoverage of the planet will be possible at this spa-tial resolution during the nominal mission How-ever the spectral resolution of this mid-infraredmapping spectrometer will be significantly lowerthan TES and it is unclear what impact this willhave on mineral detection limits at Mars In 2005the Mars Reconnaissance Orbiter (MRO) willcarry a near-infrared spectrometer (Compact Re-connaissance Imaging Spectrometer for Mars orCRISM) which will map the surface at resolu-tions as low as 10 mpixel While this hyper-spectral instrument will not achieve completecoverage of the martian surface it will providehigh-resolution mapping at targeted sites

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITYSTROMATOLITES AND RELATED

MICROBIALLY INFLUENCEDBIOFABRICS AND SEDIMENTARY

STRUCTURES

Benthic microbial communities are likely to in-fluence the microstructure of deposits that accu-mulate in their presence regardless of whetherthey occur in detrital or mineralizing environ-ments The key to proving the biogenicity of mi-crobially influenced biofabrics and structures isknowing which of their components resultedfrom the presence or influence of the organismsand whether these attributes are unique to mi-crobial processes If the attributes can form abi-otically in the absence of microorganisms thensuch attributes cannot be used to demonstratebiogenicity

To provide a framework for assessing the rel-ative roles of different biologic and abiologicprocesses in stromatolite accretion we adopt anon-genetic definition of stromatolites first pre-sented in Hofmann (1973) A stromatolite is ldquoan attached laminated lithified sedimentarygrowth structure accretionary away from a point

CADY ET AL354

or limited surface of initiationrdquo (Semikhatov et al1979) This concise definition describes the basicgeometric and textural properties of all stroma-tolites while at the same time allowing for mul-tiple or even indeterminate origins In this wayan objective evaluation can be made of the vari-ous processes that may influence stromatolite de-velopment Such an approach can lead to the formulation of specific process models that accu-rately describe stromatolite accretion dynamicsProcess models might routinely predict the rela-tive or even absolute contributions of biologicalphysical and chemical effects to stromatolite for-mation

The future value of stromatolite research lies inits potential to provide a basis for reconstructingancient environments and to understand howbenthic microbial communities interacted withtheir environments This must involve a process-oriented approach to stromatolite morphogenesisin which the correct interpretation of diageneticand recrystallization textures is as important asunderstanding microbial diversity and fossiliza-tion processes in modern microbial ecosystemsThe goal is to build an understanding of stroma-tolite development that stems from rigorousquantitative analyses of stromatolite form andlamina texture including the deconvolution of di-agenetic overprints to reveal primary fabrics di-agnostic of specific microbial and sedimentaryprocesses Critical questions regarding the factorsresponsible for the development of biofabricsbiogenic stromatolite morphologies and diage-netic textures can then be addressed For exam-ple over what length and time scales do biolog-ical physical and chemical processes operateDo any of these processesmdashwhich might be crit-ical at microscopic scalesmdashremain sensitive atlarger scales If not at what scale does the tran-sition in process response take place Questionssuch as these must be answered before we canclaim a real understanding of what fundamentalproperties such as stromatolite shape signify

The most conspicuous feature of stromatolitesis their lamination Individual laminae are thebuilding blocks of stromatolites and thereforemake up a time series of progressive albeit in-cremental accretion The morphology of anystromatolite is a function of how lamina shapeparticularly its relief evolves over time Topo-graphic anomalies that are reinforced over timegive rise to greater relief for successive laminaeand those that are stabilized give rise to greater

inheritance of shape for successive laminaeThose topographic anomalies that are dampedover time result in diminished relief

In addition to lamination the other distin-guishing feature of stromatolites is their shape Atypical stromatolite is made up of numerous suc-cessive laminae that stack one on top of the otherto form domal coniform columnar or branchingcolumnar structures Although laminae generallydescribe convex-upward structures they can alsoform concave-upward or discrete conical struc-tures In general it is thought that there is a broadbut gradational variation in the form of stromato-lites encompassing several major morphologicalmotifs (Semikhatov et al 1979) It has long beenobserved that stromatolite morphology varies as afunction of facies thus there is broad agreementthat physical environment plays a role in the gen-eration of shape (see papers in Walter 1976)

At the level of process however there is nosuch guiding consensus which severely limitsour ability to understand either paleoenviron-mental or stratigraphic variations in stromatoliteform As Hofmann (1987) stated ldquo we stillhave no stromatolite theory no model that showswhich attributes changed in what way throughtimerdquo Without a viable theory we are always atrisk of misinterpreting the genetic significance ofgrowth form

One can easily list numerous factors that mightinfluence stromatolite development includinglight intensity salinity nutrient supply currentvelocity sediment grain size distribution matcommunity diversity and degree of mineral sat-uration to name a few In detail stromatolitegrowth is dependent on many processes that are complexly interrelated Not only are theprocesses mechanistically complex they also evolveover long time scales that are difficult to repro-duce experimentally or monitor in the fieldGiven these difficulties our initial goal should beto construct simple process models of complexsystems Even this will be a difficult task in needof studies that can serve to calibrate importantmodel parameters (eg Jorgensen and DesMarais 1990) In principle the growth of stro-matolites can be described as a simple system thatdepends on only three fundamental processesgrowth and degradation of a microbial mat orbiofilm deposition of sediment and precipitationof minerals Interactions among these end-mem-ber processes should account for the bulk of stro-matolites in the record

MORPHOLOGICAL BIOSIGNATURES 355

Stromatolite growth depends on the iterativeprocess of upward growth by mats or seafloorcrusts alternating with times of sediment deposi-tion In addition these processes must be bal-anced in such a way that sediment does not over-whelm mats or microbial biofilms On furtherinspection an additional but critically importantattribute of the iterative process is revealed Thegrowth of mats tends to produce an irregular rel-atively rough surface whereas the settling of sed-iment tends to create a relatively smooth surfaceby filling in the microtopography of the under-lying mat The surface roughness of mats willvary depending on the composition and distrib-ution of the microbial community Thereforeover many iterations the surface roughness of agrowing stromatolite may be enhanced or sup-pressed depending on the competitive processesof surface roughening by mats and surfacesmoothing through sedimentation Unfortunatelyeven though it has been recognized in a qualita-tive sense for decades that microbial mats havevariable surface roughness this attribute hasnever been quantified It is recommended thatsurface roughness be measured in future studiesof modern microbial mats particularly wheresedimentation also occurs

Interface dynamics and stromatolites

The texture (ie preferred orientation) androughness of any depositional surface or interfaceis subject to certain force balances and the pres-ence of noise or randomness A widely appliedgrowth model is represented by the KPZ equa-tion (Kardar et al 1986) whose relevance to un-derstanding stromatolite growth was recentlyevaluated (Grotzinger and Rothman 1996) Thevalidity of the KPZ equation in accounting for the growth of stromatolites was tested and ten-tatively confirmed by calculating the surfaceroughness of several stromatolitic laminae andcomparing the obtained scaling exponent (andfractal dimension) with that predicted by the KPZtheory (Grotzinger and Rothman 1996) Growthof this type may characterize in a quantitativemanner the geometry of layering in many Pre-cambrian stromatolites agates botryoidal min-eral clusters travertines and certain types of stro-matolites where seafloor precipitation is thoughtto have been important For example stromato-lites formed immediately prior to the precipita-tion of some of the worldrsquos largest evaporite de-

posits are characterized by fine isopachous lam-ination and internal textures consistent with insitu precipitation (Pope and Grotzinger 1999)

In contrast for other systems the presence of adiffusing field that reflects pressure electric po-tential temperature and chemical or nutrientconcentrations may lead to a more appropriategrowth model for stromatolites since growth isfastest where concentration gradients are steep-est Initial studies of models for diffusion-limitedaggregation (DLA) strongly suggest that they arealso applicable to understanding the growth of certain stromatolites (Grotzinger and Chan1999) Whereas DLA by itself predicts highlybranched dendritic structures in the presence ofincremental sedimentation events a simple DLAmodel predicts many of the domal columnarand branching columnar stromatolite morpholo-gies observed in the geological record In thismodel an episode of upward growth by ran-domly attaching particles is taken to simulategrowth of either mats or crystals followed by anepisode of sediment settling in which the sedi-ment is allowed to settle preferentially in the mi-crodepressions formed in the underlying mat orcrystal layer If thick enough andor diffusiveenough the sediment may damp all of the initialtopography created by the underlying layerHowever if antecedent topography remainsthen the next iteration of matcrystal growth willresult in preferential accretion on those topo-graphic highs The next layer of sediment nowhas a tougher task to fill depressions giving risein the next iteration of matcrystal growth to aneven higher preference for growth on topo-graphical highs This is a particularly importantfeature of DLA models in that small perturba-tions can be amplified in time to become domi-nant features of the structure itself In this man-ner no special conditions may be required togenerate columns and branching columns in stro-matolitesmdashonly time and the positive reinforce-ment of randomly produced protuberances Thistype of growth may help account for the diver-sity of branched columnar stromatolites commonin the ancient geologic record

Biogenic versus abiogenic growth

By understanding through the use of simpleprocess models that stromatolite growth mayresult from the competitive interaction of up-ward-growth and surface roughness forced by

CADY ET AL356

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

Mars is assumed to be negligible the degree towhich any tectonic-scale forces may have alteredMars rocks is not known

In general long-term preservation of microbialbiosignatures has occurred when fossils and sed-imented or precipitated biofabrics and structureswere retained in dense impermeable host rockscomposed of stable minerals that resisted chem-ical weathering dissolution or extensive reorga-nization during diagenetic recrystallization Fa-vored lithologies for long-term preservationinclude cherts and phosphorites rocks that con-tain silica and phosphate respectively Suchlithologies have long crustal residence times andalong with carbonates and shales are the mostcommon host rocks for the Precambrian micro-fossil record on Earth

Determining the location of potential paleobi-ological repositories on Mars requires an under-standing of the martian surface in terms of ele-mental abundances and mineralogy Progress inunderstanding martian surface mineralogy hasrecently been provided by data obtained by theThermal Emission Spectrometer (TES) experi-ment carried aboard the Mars Global SurveyorOrbiter (Christensen et al 2000) TES whichmaps at a spatial resolution of 3 kmpixel in themid-infrared portion of the spectrum (6ndash14 mm)detected large deposits of coarse-grained (specu-lar) hematite (iron oxide) at several locations onMars This variety of hematite on Earth formsonly in the presence of large amounts of waterand typically at elevated (hydrothermal) temper-atures (Christensen et al 2000) The hematite de-posits appear to be co-located with geomorphicfeatures previously described as having beenformed by the action of liquid water at the sur-face of Mars TES has also revealed a basic compositional difference between the southerncratered highlands of Mars (modal compositiondominantly basalt) and the northern lowlandplains [modal composition more silica-enriched(ie basaltic-andesite)] (Bandfield et al 2000)The significance of this composition difference isunclear since the geological processes responsi-ble for them are unknown

The search on Mars for common aqueouslyprecipitated mineral assemblages and rock types(eg carbonates cherts phosphorites and evap-orites) that tend to harbor fossilized microbial as-semblages on Earth is still underway Remotesensing analog studies in Death Valley (Jolliff etal 2001 Moersch et al 2001) indicate that detec-

tion of these minerals by TES may not be possi-ble at lower abundances because of the low spa-tial resolution of the instrument Thus futuremissions will need to consider higher-resolutioninstruments that can map over a broader rangeof wavelengths and at spatial resolutions 100mpixel Improved spatial resolution (100 mpixel) is being obtained with the Thermal Emis-sion Imaging System (THEMIS) on the MarsOdyssey spacecraft presently orbiting Mars Fullcoverage of the planet will be possible at this spa-tial resolution during the nominal mission How-ever the spectral resolution of this mid-infraredmapping spectrometer will be significantly lowerthan TES and it is unclear what impact this willhave on mineral detection limits at Mars In 2005the Mars Reconnaissance Orbiter (MRO) willcarry a near-infrared spectrometer (Compact Re-connaissance Imaging Spectrometer for Mars orCRISM) which will map the surface at resolu-tions as low as 10 mpixel While this hyper-spectral instrument will not achieve completecoverage of the martian surface it will providehigh-resolution mapping at targeted sites

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITYSTROMATOLITES AND RELATED

MICROBIALLY INFLUENCEDBIOFABRICS AND SEDIMENTARY

STRUCTURES

Benthic microbial communities are likely to in-fluence the microstructure of deposits that accu-mulate in their presence regardless of whetherthey occur in detrital or mineralizing environ-ments The key to proving the biogenicity of mi-crobially influenced biofabrics and structures isknowing which of their components resultedfrom the presence or influence of the organismsand whether these attributes are unique to mi-crobial processes If the attributes can form abi-otically in the absence of microorganisms thensuch attributes cannot be used to demonstratebiogenicity

To provide a framework for assessing the rel-ative roles of different biologic and abiologicprocesses in stromatolite accretion we adopt anon-genetic definition of stromatolites first pre-sented in Hofmann (1973) A stromatolite is ldquoan attached laminated lithified sedimentarygrowth structure accretionary away from a point

CADY ET AL354

or limited surface of initiationrdquo (Semikhatov et al1979) This concise definition describes the basicgeometric and textural properties of all stroma-tolites while at the same time allowing for mul-tiple or even indeterminate origins In this wayan objective evaluation can be made of the vari-ous processes that may influence stromatolite de-velopment Such an approach can lead to the formulation of specific process models that accu-rately describe stromatolite accretion dynamicsProcess models might routinely predict the rela-tive or even absolute contributions of biologicalphysical and chemical effects to stromatolite for-mation

The future value of stromatolite research lies inits potential to provide a basis for reconstructingancient environments and to understand howbenthic microbial communities interacted withtheir environments This must involve a process-oriented approach to stromatolite morphogenesisin which the correct interpretation of diageneticand recrystallization textures is as important asunderstanding microbial diversity and fossiliza-tion processes in modern microbial ecosystemsThe goal is to build an understanding of stroma-tolite development that stems from rigorousquantitative analyses of stromatolite form andlamina texture including the deconvolution of di-agenetic overprints to reveal primary fabrics di-agnostic of specific microbial and sedimentaryprocesses Critical questions regarding the factorsresponsible for the development of biofabricsbiogenic stromatolite morphologies and diage-netic textures can then be addressed For exam-ple over what length and time scales do biolog-ical physical and chemical processes operateDo any of these processesmdashwhich might be crit-ical at microscopic scalesmdashremain sensitive atlarger scales If not at what scale does the tran-sition in process response take place Questionssuch as these must be answered before we canclaim a real understanding of what fundamentalproperties such as stromatolite shape signify

The most conspicuous feature of stromatolitesis their lamination Individual laminae are thebuilding blocks of stromatolites and thereforemake up a time series of progressive albeit in-cremental accretion The morphology of anystromatolite is a function of how lamina shapeparticularly its relief evolves over time Topo-graphic anomalies that are reinforced over timegive rise to greater relief for successive laminaeand those that are stabilized give rise to greater

inheritance of shape for successive laminaeThose topographic anomalies that are dampedover time result in diminished relief

In addition to lamination the other distin-guishing feature of stromatolites is their shape Atypical stromatolite is made up of numerous suc-cessive laminae that stack one on top of the otherto form domal coniform columnar or branchingcolumnar structures Although laminae generallydescribe convex-upward structures they can alsoform concave-upward or discrete conical struc-tures In general it is thought that there is a broadbut gradational variation in the form of stromato-lites encompassing several major morphologicalmotifs (Semikhatov et al 1979) It has long beenobserved that stromatolite morphology varies as afunction of facies thus there is broad agreementthat physical environment plays a role in the gen-eration of shape (see papers in Walter 1976)

At the level of process however there is nosuch guiding consensus which severely limitsour ability to understand either paleoenviron-mental or stratigraphic variations in stromatoliteform As Hofmann (1987) stated ldquo we stillhave no stromatolite theory no model that showswhich attributes changed in what way throughtimerdquo Without a viable theory we are always atrisk of misinterpreting the genetic significance ofgrowth form

One can easily list numerous factors that mightinfluence stromatolite development includinglight intensity salinity nutrient supply currentvelocity sediment grain size distribution matcommunity diversity and degree of mineral sat-uration to name a few In detail stromatolitegrowth is dependent on many processes that are complexly interrelated Not only are theprocesses mechanistically complex they also evolveover long time scales that are difficult to repro-duce experimentally or monitor in the fieldGiven these difficulties our initial goal should beto construct simple process models of complexsystems Even this will be a difficult task in needof studies that can serve to calibrate importantmodel parameters (eg Jorgensen and DesMarais 1990) In principle the growth of stro-matolites can be described as a simple system thatdepends on only three fundamental processesgrowth and degradation of a microbial mat orbiofilm deposition of sediment and precipitationof minerals Interactions among these end-mem-ber processes should account for the bulk of stro-matolites in the record

MORPHOLOGICAL BIOSIGNATURES 355

Stromatolite growth depends on the iterativeprocess of upward growth by mats or seafloorcrusts alternating with times of sediment deposi-tion In addition these processes must be bal-anced in such a way that sediment does not over-whelm mats or microbial biofilms On furtherinspection an additional but critically importantattribute of the iterative process is revealed Thegrowth of mats tends to produce an irregular rel-atively rough surface whereas the settling of sed-iment tends to create a relatively smooth surfaceby filling in the microtopography of the under-lying mat The surface roughness of mats willvary depending on the composition and distrib-ution of the microbial community Thereforeover many iterations the surface roughness of agrowing stromatolite may be enhanced or sup-pressed depending on the competitive processesof surface roughening by mats and surfacesmoothing through sedimentation Unfortunatelyeven though it has been recognized in a qualita-tive sense for decades that microbial mats havevariable surface roughness this attribute hasnever been quantified It is recommended thatsurface roughness be measured in future studiesof modern microbial mats particularly wheresedimentation also occurs

Interface dynamics and stromatolites

The texture (ie preferred orientation) androughness of any depositional surface or interfaceis subject to certain force balances and the pres-ence of noise or randomness A widely appliedgrowth model is represented by the KPZ equa-tion (Kardar et al 1986) whose relevance to un-derstanding stromatolite growth was recentlyevaluated (Grotzinger and Rothman 1996) Thevalidity of the KPZ equation in accounting for the growth of stromatolites was tested and ten-tatively confirmed by calculating the surfaceroughness of several stromatolitic laminae andcomparing the obtained scaling exponent (andfractal dimension) with that predicted by the KPZtheory (Grotzinger and Rothman 1996) Growthof this type may characterize in a quantitativemanner the geometry of layering in many Pre-cambrian stromatolites agates botryoidal min-eral clusters travertines and certain types of stro-matolites where seafloor precipitation is thoughtto have been important For example stromato-lites formed immediately prior to the precipita-tion of some of the worldrsquos largest evaporite de-

posits are characterized by fine isopachous lam-ination and internal textures consistent with insitu precipitation (Pope and Grotzinger 1999)

In contrast for other systems the presence of adiffusing field that reflects pressure electric po-tential temperature and chemical or nutrientconcentrations may lead to a more appropriategrowth model for stromatolites since growth isfastest where concentration gradients are steep-est Initial studies of models for diffusion-limitedaggregation (DLA) strongly suggest that they arealso applicable to understanding the growth of certain stromatolites (Grotzinger and Chan1999) Whereas DLA by itself predicts highlybranched dendritic structures in the presence ofincremental sedimentation events a simple DLAmodel predicts many of the domal columnarand branching columnar stromatolite morpholo-gies observed in the geological record In thismodel an episode of upward growth by ran-domly attaching particles is taken to simulategrowth of either mats or crystals followed by anepisode of sediment settling in which the sedi-ment is allowed to settle preferentially in the mi-crodepressions formed in the underlying mat orcrystal layer If thick enough andor diffusiveenough the sediment may damp all of the initialtopography created by the underlying layerHowever if antecedent topography remainsthen the next iteration of matcrystal growth willresult in preferential accretion on those topo-graphic highs The next layer of sediment nowhas a tougher task to fill depressions giving risein the next iteration of matcrystal growth to aneven higher preference for growth on topo-graphical highs This is a particularly importantfeature of DLA models in that small perturba-tions can be amplified in time to become domi-nant features of the structure itself In this man-ner no special conditions may be required togenerate columns and branching columns in stro-matolitesmdashonly time and the positive reinforce-ment of randomly produced protuberances Thistype of growth may help account for the diver-sity of branched columnar stromatolites commonin the ancient geologic record

Biogenic versus abiogenic growth

By understanding through the use of simpleprocess models that stromatolite growth mayresult from the competitive interaction of up-ward-growth and surface roughness forced by

CADY ET AL356

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

or limited surface of initiationrdquo (Semikhatov et al1979) This concise definition describes the basicgeometric and textural properties of all stroma-tolites while at the same time allowing for mul-tiple or even indeterminate origins In this wayan objective evaluation can be made of the vari-ous processes that may influence stromatolite de-velopment Such an approach can lead to the formulation of specific process models that accu-rately describe stromatolite accretion dynamicsProcess models might routinely predict the rela-tive or even absolute contributions of biologicalphysical and chemical effects to stromatolite for-mation

The future value of stromatolite research lies inits potential to provide a basis for reconstructingancient environments and to understand howbenthic microbial communities interacted withtheir environments This must involve a process-oriented approach to stromatolite morphogenesisin which the correct interpretation of diageneticand recrystallization textures is as important asunderstanding microbial diversity and fossiliza-tion processes in modern microbial ecosystemsThe goal is to build an understanding of stroma-tolite development that stems from rigorousquantitative analyses of stromatolite form andlamina texture including the deconvolution of di-agenetic overprints to reveal primary fabrics di-agnostic of specific microbial and sedimentaryprocesses Critical questions regarding the factorsresponsible for the development of biofabricsbiogenic stromatolite morphologies and diage-netic textures can then be addressed For exam-ple over what length and time scales do biolog-ical physical and chemical processes operateDo any of these processesmdashwhich might be crit-ical at microscopic scalesmdashremain sensitive atlarger scales If not at what scale does the tran-sition in process response take place Questionssuch as these must be answered before we canclaim a real understanding of what fundamentalproperties such as stromatolite shape signify

The most conspicuous feature of stromatolitesis their lamination Individual laminae are thebuilding blocks of stromatolites and thereforemake up a time series of progressive albeit in-cremental accretion The morphology of anystromatolite is a function of how lamina shapeparticularly its relief evolves over time Topo-graphic anomalies that are reinforced over timegive rise to greater relief for successive laminaeand those that are stabilized give rise to greater

inheritance of shape for successive laminaeThose topographic anomalies that are dampedover time result in diminished relief

In addition to lamination the other distin-guishing feature of stromatolites is their shape Atypical stromatolite is made up of numerous suc-cessive laminae that stack one on top of the otherto form domal coniform columnar or branchingcolumnar structures Although laminae generallydescribe convex-upward structures they can alsoform concave-upward or discrete conical struc-tures In general it is thought that there is a broadbut gradational variation in the form of stromato-lites encompassing several major morphologicalmotifs (Semikhatov et al 1979) It has long beenobserved that stromatolite morphology varies as afunction of facies thus there is broad agreementthat physical environment plays a role in the gen-eration of shape (see papers in Walter 1976)

At the level of process however there is nosuch guiding consensus which severely limitsour ability to understand either paleoenviron-mental or stratigraphic variations in stromatoliteform As Hofmann (1987) stated ldquo we stillhave no stromatolite theory no model that showswhich attributes changed in what way throughtimerdquo Without a viable theory we are always atrisk of misinterpreting the genetic significance ofgrowth form

One can easily list numerous factors that mightinfluence stromatolite development includinglight intensity salinity nutrient supply currentvelocity sediment grain size distribution matcommunity diversity and degree of mineral sat-uration to name a few In detail stromatolitegrowth is dependent on many processes that are complexly interrelated Not only are theprocesses mechanistically complex they also evolveover long time scales that are difficult to repro-duce experimentally or monitor in the fieldGiven these difficulties our initial goal should beto construct simple process models of complexsystems Even this will be a difficult task in needof studies that can serve to calibrate importantmodel parameters (eg Jorgensen and DesMarais 1990) In principle the growth of stro-matolites can be described as a simple system thatdepends on only three fundamental processesgrowth and degradation of a microbial mat orbiofilm deposition of sediment and precipitationof minerals Interactions among these end-mem-ber processes should account for the bulk of stro-matolites in the record

MORPHOLOGICAL BIOSIGNATURES 355

Stromatolite growth depends on the iterativeprocess of upward growth by mats or seafloorcrusts alternating with times of sediment deposi-tion In addition these processes must be bal-anced in such a way that sediment does not over-whelm mats or microbial biofilms On furtherinspection an additional but critically importantattribute of the iterative process is revealed Thegrowth of mats tends to produce an irregular rel-atively rough surface whereas the settling of sed-iment tends to create a relatively smooth surfaceby filling in the microtopography of the under-lying mat The surface roughness of mats willvary depending on the composition and distrib-ution of the microbial community Thereforeover many iterations the surface roughness of agrowing stromatolite may be enhanced or sup-pressed depending on the competitive processesof surface roughening by mats and surfacesmoothing through sedimentation Unfortunatelyeven though it has been recognized in a qualita-tive sense for decades that microbial mats havevariable surface roughness this attribute hasnever been quantified It is recommended thatsurface roughness be measured in future studiesof modern microbial mats particularly wheresedimentation also occurs

Interface dynamics and stromatolites

The texture (ie preferred orientation) androughness of any depositional surface or interfaceis subject to certain force balances and the pres-ence of noise or randomness A widely appliedgrowth model is represented by the KPZ equa-tion (Kardar et al 1986) whose relevance to un-derstanding stromatolite growth was recentlyevaluated (Grotzinger and Rothman 1996) Thevalidity of the KPZ equation in accounting for the growth of stromatolites was tested and ten-tatively confirmed by calculating the surfaceroughness of several stromatolitic laminae andcomparing the obtained scaling exponent (andfractal dimension) with that predicted by the KPZtheory (Grotzinger and Rothman 1996) Growthof this type may characterize in a quantitativemanner the geometry of layering in many Pre-cambrian stromatolites agates botryoidal min-eral clusters travertines and certain types of stro-matolites where seafloor precipitation is thoughtto have been important For example stromato-lites formed immediately prior to the precipita-tion of some of the worldrsquos largest evaporite de-

posits are characterized by fine isopachous lam-ination and internal textures consistent with insitu precipitation (Pope and Grotzinger 1999)

In contrast for other systems the presence of adiffusing field that reflects pressure electric po-tential temperature and chemical or nutrientconcentrations may lead to a more appropriategrowth model for stromatolites since growth isfastest where concentration gradients are steep-est Initial studies of models for diffusion-limitedaggregation (DLA) strongly suggest that they arealso applicable to understanding the growth of certain stromatolites (Grotzinger and Chan1999) Whereas DLA by itself predicts highlybranched dendritic structures in the presence ofincremental sedimentation events a simple DLAmodel predicts many of the domal columnarand branching columnar stromatolite morpholo-gies observed in the geological record In thismodel an episode of upward growth by ran-domly attaching particles is taken to simulategrowth of either mats or crystals followed by anepisode of sediment settling in which the sedi-ment is allowed to settle preferentially in the mi-crodepressions formed in the underlying mat orcrystal layer If thick enough andor diffusiveenough the sediment may damp all of the initialtopography created by the underlying layerHowever if antecedent topography remainsthen the next iteration of matcrystal growth willresult in preferential accretion on those topo-graphic highs The next layer of sediment nowhas a tougher task to fill depressions giving risein the next iteration of matcrystal growth to aneven higher preference for growth on topo-graphical highs This is a particularly importantfeature of DLA models in that small perturba-tions can be amplified in time to become domi-nant features of the structure itself In this man-ner no special conditions may be required togenerate columns and branching columns in stro-matolitesmdashonly time and the positive reinforce-ment of randomly produced protuberances Thistype of growth may help account for the diver-sity of branched columnar stromatolites commonin the ancient geologic record

Biogenic versus abiogenic growth

By understanding through the use of simpleprocess models that stromatolite growth mayresult from the competitive interaction of up-ward-growth and surface roughness forced by

CADY ET AL356

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

Stromatolite growth depends on the iterativeprocess of upward growth by mats or seafloorcrusts alternating with times of sediment deposi-tion In addition these processes must be bal-anced in such a way that sediment does not over-whelm mats or microbial biofilms On furtherinspection an additional but critically importantattribute of the iterative process is revealed Thegrowth of mats tends to produce an irregular rel-atively rough surface whereas the settling of sed-iment tends to create a relatively smooth surfaceby filling in the microtopography of the under-lying mat The surface roughness of mats willvary depending on the composition and distrib-ution of the microbial community Thereforeover many iterations the surface roughness of agrowing stromatolite may be enhanced or sup-pressed depending on the competitive processesof surface roughening by mats and surfacesmoothing through sedimentation Unfortunatelyeven though it has been recognized in a qualita-tive sense for decades that microbial mats havevariable surface roughness this attribute hasnever been quantified It is recommended thatsurface roughness be measured in future studiesof modern microbial mats particularly wheresedimentation also occurs

Interface dynamics and stromatolites

The texture (ie preferred orientation) androughness of any depositional surface or interfaceis subject to certain force balances and the pres-ence of noise or randomness A widely appliedgrowth model is represented by the KPZ equa-tion (Kardar et al 1986) whose relevance to un-derstanding stromatolite growth was recentlyevaluated (Grotzinger and Rothman 1996) Thevalidity of the KPZ equation in accounting for the growth of stromatolites was tested and ten-tatively confirmed by calculating the surfaceroughness of several stromatolitic laminae andcomparing the obtained scaling exponent (andfractal dimension) with that predicted by the KPZtheory (Grotzinger and Rothman 1996) Growthof this type may characterize in a quantitativemanner the geometry of layering in many Pre-cambrian stromatolites agates botryoidal min-eral clusters travertines and certain types of stro-matolites where seafloor precipitation is thoughtto have been important For example stromato-lites formed immediately prior to the precipita-tion of some of the worldrsquos largest evaporite de-

posits are characterized by fine isopachous lam-ination and internal textures consistent with insitu precipitation (Pope and Grotzinger 1999)

In contrast for other systems the presence of adiffusing field that reflects pressure electric po-tential temperature and chemical or nutrientconcentrations may lead to a more appropriategrowth model for stromatolites since growth isfastest where concentration gradients are steep-est Initial studies of models for diffusion-limitedaggregation (DLA) strongly suggest that they arealso applicable to understanding the growth of certain stromatolites (Grotzinger and Chan1999) Whereas DLA by itself predicts highlybranched dendritic structures in the presence ofincremental sedimentation events a simple DLAmodel predicts many of the domal columnarand branching columnar stromatolite morpholo-gies observed in the geological record In thismodel an episode of upward growth by ran-domly attaching particles is taken to simulategrowth of either mats or crystals followed by anepisode of sediment settling in which the sedi-ment is allowed to settle preferentially in the mi-crodepressions formed in the underlying mat orcrystal layer If thick enough andor diffusiveenough the sediment may damp all of the initialtopography created by the underlying layerHowever if antecedent topography remainsthen the next iteration of matcrystal growth willresult in preferential accretion on those topo-graphic highs The next layer of sediment nowhas a tougher task to fill depressions giving risein the next iteration of matcrystal growth to aneven higher preference for growth on topo-graphical highs This is a particularly importantfeature of DLA models in that small perturba-tions can be amplified in time to become domi-nant features of the structure itself In this man-ner no special conditions may be required togenerate columns and branching columns in stro-matolitesmdashonly time and the positive reinforce-ment of randomly produced protuberances Thistype of growth may help account for the diver-sity of branched columnar stromatolites commonin the ancient geologic record

Biogenic versus abiogenic growth

By understanding through the use of simpleprocess models that stromatolite growth mayresult from the competitive interaction of up-ward-growth and surface roughness forced by

CADY ET AL356

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

microbial mats and dampening of surface reliefby sediment settling it becomes easy to see howthe growth of abiotic marine crusts might sub-stitute for mats and create the same end resultThe good news is that we may now have a the-ory that can account for the growth of a re-markable range of stromatolites The bad newsis that this theory predicts that we can no longeraccept only morphological descriptions of stro-matolites as evidence of their biogenicity Thisdoes not mean that stromatolites may not havegrown in the presence of biogenic influences Itmeans however that in many cases morphol-ogy may be a non-unique parameter Biogenic-ity cannot easily be demonstrated on the basisof relationships observed at the outcrop scale itis essential to examine lamination textures pet-rographically and demonstrate the presence offabrics or textures (primary or secondary)uniquely attributable to the presence of micro-bial mats or biofilms (Cady and Farmer 1996Knoll and Semikhatov 1998) For many ancientstromatolites such an approach may not be pos-sible because of an indecipherable level of dia-genetic recrystallization

Biosignatures produced by microbial biofilms

Although we have emphasized the limitationsfaced when attempting to distinguish the bio-genicity of microbially influenced sedimentarystructures there are a number of structures formedby benthic microbial communities that occur asbiofilms rather than flat-laminated microbial matcommunities For example the high-temperaturesiliceous sinter known as geyserite displays mi-crostructural attributes that result from the pres-ence of benthic microbial biofilms on accretionarysurfaces (Cady and Farmer 1996) Microbial bio-films also occur as cryptoendoliths that form in-side minerals such as the biofilm communitiesthat occur at the sediment surface in AntarcticDry Valleys (Friedmann et al 1988 Wynn-Williams et al 1999) If life exists at the surfaceof Mars it would likely occur as endoliths Thesubmicroscopic attributes of these types of de-posits are clearly influenced by microbialbiofilms yet they may not be detectable by thecriteria established for recognizing the biogenic-ity of stromatolites The fate of biofilms in naturalenvironments the fossilization potential of thesestructures and how these factors relate to estab-lished criteria for demonstrating biogenicity are

priority research topics that have yet to be sys-tematically addressed

LIMITATIONS OF CRITERIA FORDEMONSTRATING BIOGENICITY BONA

FIDE MICROFOSSILS

Recall that any claim of morphological fossilsas representing life that once existed beyondEarth must be accompanied by proof that the ob-jects were produced biologically or biosyntheti-cally and that they are not terrestrial contami-nants Of the two methods traditionally used todetect ancient microfossils (eg Schopf 1999)maceration is the easier and faster of the two tech-niques Macerations are carried out by dissolvingrocks in mineral acids (hydrochloric acid for lime-stones hydrofluoric acid for cherts and silt-stones) Because of their carbonaceous composi-tion organic-walled microfossils pass throughthe technique unscathed Abundant fossils areconcentrated in the resulting sludge-like acid-re-sistant residue that can be slurried onto a micro-scope slide for study Unless specimens are prepared in such a way as to avoid air- and wa-ter-borne contaminants however this techniqueis subject to the problems posed by contaminantsbeing identified as false-positives for life

Petrographic thin section analysis the othertechnique traditionally used to detect microfos-sils in ancient rocks provides a means to evalu-ate the indigenicity of purported fossil objectsThe possibility of laboratory contamination canbe ruled out if the fossils are clearly embeddedwithin the mineral matrix as evidenced in thinsections of the rock within which they are foundConsider for example microfossils in cherts To-gether with fine-grained clastic sediments suchas siltstones cherts are one of the most fossilifer-ous rock types known in the early geologicrecord Ancient fossil-bearing cherts are made upof cryptocrystalline interlocking grains of quartzThe grains precipitated initially as opaline silicataking thousands of years to transform into a full-fledged quartz chert during which time the mi-croorganisms became petrified (technically ldquoper-mineralizedrdquo) and embedded within a solidchunk of rock The quartz grains that formed in-side the cells and surrounded them on all sidesdeveloped so slowly that they grew through thecell walls instead of crushing them As a result

MORPHOLOGICAL BIOSIGNATURES 357

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

the petrified fossils are preserved as unflattenedthree-dimensional bodies Except for their quartz-filled interiors and the brownish color of theiraged organic matter (kerogen) they bear a strik-ing resemblance to living microorganisms In pet-rographic thin sections only those objects that areentirely entombed in rock can be considered fos-sil so it is easy to exclude contaminants that set-tle onto the surface of a section or are embeddedin the resin used to cement the sliver of rock ontothe glass thin-section mount

Optical microscopy of thin sections provides ameans to establish that fossil-like objects datefrom the time a rock formed rather than havingbeen sealed later in cracks and crevices In thisway it is possible to establish that the objects aresyngenetic with a primary mineral phase ratherthan one of secondary or later genesis Considerfor example microfossiliferous stromatolitic chertsWhen such cherts first form many contain cavi-ties where gases (often oxygen carbon dioxidehydrogen or methane) given off by the microbialcommunity accumulate in small pockets Laterthese cavities can be sealed and filled by a sec-ond generation of quartz laid down from seepinggroundwater sometimes tens or hundreds of mil-lions of years after the primary chert precipitatedMicroscopic organisms trapped in these cavitiesand petrified by the second-generation quartzwould be true fossils that are younger than therock unit itself Fortunately the various genera-tions of quartz in a chert can be distinguishedfrom each other Rather than having interlockinggrains the quartz variety that rapidly fills cavi-ties is a type known as chalcedony which followsthe smooth contours of the infilled pocket to formdistinctive botryoidal masses Secondary quartzin cracks or veinlets is also easy to identify sinceit is angular and its grains are much larger thanthose first formed

Because special equipment is needed to pre-pare thin sections and their study is exceedinglytime-consuming some workers have focusedtheir hunt for ancient fossils on acid-resistant rockresidues In relatively young (Proterozoic) Pre-cambrian rocks where the fossil record is wellenough known that misidentification of contam-inants and fossil-like artifacts can be avoided thistechnique is useful simple and fast But to avoidmistakes when examining older (Archean) Pre-cambrian deposits where the fossil record is notnearly so well known or of course when exam-ining extraterrestrial samples use of the more rig-

orous thin-section technique is essential Sincecriteria by which the biogenicity of bona fide mi-crofossils identified by light microscopy have yetto be established and verified at the scale of de-tection (imaging and spectroscopic) obtainedwith electron microscopy we recommend furtherresearch of this topic

Although transmission (TEM) and scanning(SEM) electron microscopy have been used tocharacterize microfossils previously detected inmacerations or thin sections these techniqueshave not yet proven reliable for demonstratingthe biogenicity of microfossils not identified byother means In fact such techniques have re-vealed a new set of challenges posed by the dis-covery of submicroscopic objects characterizedby microbial morphologies that lack detectableorganic components (eg McKay et al 1996Westall 1999) In addition to established criteriafor assessing microfossil biogenicity (Schopf andWalter 1983 Buick 1990 Schopf 1999) a num-ber of new criteria have been proposed for as-sessing the biogenicity of submicroscopic-sizedmicrofossil-like objects (Westall 1999) Regard-less of the size of purported morphological fos-sils the evidence sought for life should be posi-tivemdashevidence that affirms the biological originof the features detected (Schopf 1999) Evidencethat is neutral (consistent either with biology ornon-biological processes) is by its nature inade-quate to establish the existence of past life andinterpretations based on negative reasoning or in-ference by default are likely to prove erroneousSuch stringency is warranted in the search for lifebeyond Earth as well as in the search for the old-est evidence of life on Earth

FORMATION AND ALTERATION OFMICROFOSSILS

Geomicrobiological and mineralogical studiesof modern ecosystems recognized as analogs forearly Earth or Mars environments along withlaboratory-based experiments designed to simu-late microbial fossilization and stromatolite for-mation continue to improve our understandingof the processes involved in producing morpho-logical fossils As discussed by Cady (1998) sincemicroorganisms can occupy nearly every avail-able habitat where water and available carbonand energy sources exist the number of potentialpaleobiological sites in which evidence of life

CADY ET AL358

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

may be preserved has expanded Indeed anystructural discontinuity in rocks through whichmineralizing fluids have passed freely should besearched for morphological fossils including mi-crofossils and microbially influenced fabrics andstructures Many modern extreme ecosystemsand their ancient analog deposits have not beensystematically assessed as potential paleobio-logical repositories (eg subsurface rocks per-mafrost regions paleosols and hydrothermaland epithermal deposits) Because of their highpotential to demonstrate life it is important to un-derstand how microfossils are formed preservedand altered The formation and preservation ofmicrofossils depend upon the intrinsic character-istics of the microorganisms the chemical andphysical characteristics of their environment andthe amount of post-depositional alteration and di-agenesis they experience (eg Cady 2001)

Intrinsic characteristics of microorganisms

The intrinsic characteristics of microorganismssignificantly influence whether they become fos-silized in either chemically mineralizing environ-ments or in detrital sedimentary environmentsStudies to date indicate differences in the cell walltype (eg Westall 1997) the presence of recalci-trant sheaths (eg Cady and Farmer 1996) orother exopolymers (eg Allen et al 2000) andthe reactivity of biomolecules within cells (egFortin et al 1997) can influence the susceptibilityof microorganisms to mineralization The pro-pensity of some microorganisms to sequesteranomalous concentrations of trace metals such asiron a process that can enhance microbial preser-vation has also been shown to depend upon thetype density and distribution of biomoleculesthat make up microbial cell walls and various ex-tracellular components (eg Ferris et al 1989Beveridge et al 1997) Future studies to deter-mine when and how the composition of cellular andextracellular components changes and whetherthose changes alter the susceptibility of a micro-organism to fossilization will provide constraintsneeded to assess the range of conditions overwhich trace metal concentrations can be used asbiogenic indicators Furthermore since the reac-tivity of cellular surfaces depends upon the mi-croenvironment around the cell rather than thebulk composition of any external fluid further ex-periments should concentrate on analyzing fluidgeochemistry at the cellular scale

Extrinsic geochemical characteristics

As reviewed by Farmer and Des Marais (1999)microbial preservation on Earth occurs in envi-ronments where microbial cells are entombed ina fine-grained authigenically precipitated min-eral matrix or where they are rapidly buried byfine-grained sediments These processes protectcellular remains from oxidative degradation

In detrital sedimentary systems preservationis enhanced by rapid burial In addition fine-grained detrital-rich environments often sustainanoxic conditions that favor early diagenetic min-eralization This can also enhance preservation byreducing permeability and helping to create a closed chemical system that arrests cellulardegradation

In chemically mineralizing sedimentary systemsenvironments that favor rapid mineral depositioncan entomb organisms while they are still aliveThis process arrests degradation and enhances thepreservation of important aspects of morphologygrowth habit and distribution of organisms withinmicrobial mats and biofilms Rapid reduction inporosity and permeability is important for cellularpreservation in chemically mineralizing systems aswell sustained migration of oxidizing fluids canultimately remove all traces of cellular remains andpromote early diagenetic mineral transformationsIt is recommended that the extrinsic geochemicalcharacteristics that affect preservation in modernecosystems that could have analogs on Mars be sys-tematically studied

Mineral and biomolecule diagenesis

Whether microfossils can be detected and reli-ably identified in the ancient geological record ul-timately depends upon their diagenetic historyDiagenesis occurs when an increase in the reac-tivity of a phase (either due to an increase in tem-perature or pressure or via waterndashrock inter-action) results in its structural or chemicaltransformation Both primary biomolecules andmineral phases can be diagenetically alteredConsideration of the preservation potential of alltypes of biomolecules suggests that the variousclasses of lipids especially glycolipids and lipo-polysaccharides are most resistant to degrada-tion in all types of depositional environments (DeLeeuw and Largeau 1993) The conversion of pri-mary lipids to their diagenetic counterparts in-volves information loss through structural alter-ation However the original class of lipid can

MORPHOLOGICAL BIOSIGNATURES 359

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

often be identified even after such diageneticchanges occur (Summons 1993)

Microfossils are generally preserved in primarykinetically favored mineral assemblages that ulti-mately recrystallize to thermodynamically stablemineral assemblages An example of mineral dia-genesis that alters chert deposits with time is thetransformation of primary opaline silica phases tomicrocrystalline quartz varieties through interme-diary cryptocrystalline opal phases (Hesse 1989)Recrystallization of primary aqueous mineral ma-trices often obliterates the fine-scale morphologicalfeatures of microfossils which are finer than theminimum grain size of the predominant diageneticphase (eg microquartz by definition has an aver-age grain size of 5ndash20 mm) Although recrystalliza-tion of primary mineral assemblages is known tooccur as a function of time it is recommended thatthe details of how such processes alter biosigna-tures be systematically studied

Chemical and structural discontinuities

Regardless of the mechanism by which a mi-croorganism is fossilized or altered during diagen-esis the resultant microfossil will go undetectedunless it differs either in composition or in struc-tural organization from the mineral matrix that sur-rounds it The most illustrative examples of chem-ical and structural discontinuities are thosebetween acritarchs (organically preserved cells) orpermineralized microorganisms (cellularly pre-served cells) and the mineral matrices in whichthey are preserved Environmental perturbationscan also produce compositional differences be-tween fossilized microbes and their surroundingmineral matrix The temporal changes that occurin fluids within evaporative environments or influid mixing zones can be preserved in the lami-nated crusts that develop around microbial cellsThe sequence of minerals may reveal informationabout the metabolic character of the microorgan-ism preserve a cast of a microbial cell and help re-construct details about the nature of paleoenviron-ments The survival of such features as a functionof time and increasing diagenesis is a fundamentalresearch question we recommend pursuing

SPATIALLY INTEGRATEDPALEOBIOLOGICAL STUDIES

Advances in our understanding of the afore-mentioned processes and the ways in which mi-

crobes interact with and influence their environ-ment must be accompanied by technological ad-vances that will improve our ability to detectbiosignatures of life preserved within ancientrocks formed on Earth or beyond Paleontologi-cal interpretations including those made for ex-traterrestrial materials rely upon comparisonsbetween the biology and environmental charac-teristics of modern and fossilized ecosystemsMorphological fossil remains preserve importantaspects of the behavior and distribution of mi-croorganisms and biomarkers and isotope sig-natures provide details about their metabolismand functional capacity Biominerals reflect boththe geochemistry of the microbial milieu and thebiogeochemical processes by which micro-organisms interact with their environment Thestructural and geochemical characteristics of sed-imentary repositories reflect the paleoenviron-ments in which microbial populations lived

A number of factors complicate the use of mod-ern analogs for interpreting the paleobiology andpaleoenvironment of ancient paleobiological re-positories The simple shapes and limited sizerange of microbial cells often preclude diagnos-tic paleobiological interpretations based entirelyon morphology The progressive taphonomic alteration of both physical and biochemical characteristics of microfossils limits the utility of chemical and mineralogical biosignatures toyounger relatively unaltered rocks The secularchanges observed in biogeochemical cycles (egcarbon sulfur iron) during Earthrsquos early historyindicate that many modern-day ecosystems can-not be regarded as direct homologs of ancientecosystems Even with these caveats many pale-obiological studies include parallel geomicrobio-logical studies of modern ecosystems Modernanalog studies can help constrain and quantifythe fundamental biological physical and chemi-cal processes that ultimately lead to the forma-tion of a fossil record

An example of a scale-integrated frameworkfor paleontological investigations based on acomparison of modern and ancient analogs isshown in Fig 1 The examples illustrate the sim-ilarity of attributes at different observation scalesof modern hydrothermal spring deposits in Yel-lowstone National Park (Farmer 1999) and of an-cient siliceous hydrothermal spring deposits innortheast Queensland Australia (Walter et al1998) For each spatial scale environmental andbiological comparisons reflect the lateral facies

CADY ET AL360

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

MORPHOLOGICAL BIOSIGNATURES 361

FIG 1 Scale-integrated framework for paleontological investigations based on a comparison of modern (Yel-lowstone National Park) and ancient (northeast Queensland Australia) hydrothermal spring deposit analogs Theexamples illustrate the similarity of attributes at different observation scales At a regional scale of hundreds of kilo-meters the geological context of this continental volcanic terrane provides important constraints for interpreting themajor geological environments and facies or depositional trends A At the scale of a local outcrop of a single hotspring system meters to kilometers environmental parameters such as thermal and pH gradients trends in com-munity distribution and systematic changes in sinter texture and mineralogy are apparent B and C At the mi-croenvironmental and microfacies scale of the hot spring system meters to centimeters distinctive mat communitytypes surface biofabrics and major sinter types are evident D Investigation of the mats at the microscale centime-ters to submillimeters reveals that each mat type exhibits distinctive mat compositions Vertical and lateral changesin the distribution of organisms can be related to small-scale changes in environmental characteristics as well as cor-responding microstructural changes in sinter structure characteristics that include lamina shapes and internal fabricE At the level of the microbial ultrastructure and mineral microstructure it is possible to obtain nanometer-scale in-formation about the physical conformation and chemical composition of biomolecules

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

and vertical temporal changes in the system overdifferent time scales

Integrating observations over all of the obser-vational scales discussed above is important forunderstanding the origin of emergent propertiesof the entire system Spatially integrated studiesalso provide important constraints for evaluatinghypotheses regarding the nature of the ecosys-tem For example basic mat structures and orga-nizational modes arise from the growth andmotility or taxis of the component species Thismeans that to understand the higher-order matfeatures requires some knowledge of how themotile species interacts with its environment atthe next lower level of organization A knowledgeof motility responses of various taxa in modernecosystems provides in turn a context for test-ing ideas about the origin of biosedimentary fab-rics preserved in ancient deposits Clearly an ap-proach that integrates data across spatial scalesaffords a more robust framework for interpretingpaleobiology and paleoenvironments

In taking an analog approach it is assumed thatthe laws of physics and chemistry are universallyconstant now and in the past On this basis theresults of investigations of terrestrial analog sys-tems can be extrapolated and applied to otherplanets as well as to early Earth within the con-straints of known environmental differences Thepractical risks are that (1) we still have a lot tolearn about the environments on early Earth andMars and (2) life on Mars could be entirelyunique possessing unknown properties that can-not be predicted from an Earth-centric modelNevertheless even with these uncertainties it islogical to begin with what we know about lifebased on the one example we have for study andto assume the invariability of natural chemicaland physical laws as a framework for definingappropriate strategies for exploration The alter-native is to take a nonndashEarth-centric approach tothe study of microbial biosignatures (Conrad and Nealson 2001) The essence of analog andnonndashEarth-centric approaches is the same searchfor biosignatures left by microorganisms

MORPHOLOGICAL BIOSIGNATUREDATABASE

As discussed in this report a variety of high-res-olution techniques can be used and developed toidentify microbial biosignatures A relatively large

and diverse database of images and spectroscopicdata indicative of life will continue to accumulatein the peer-reviewed literature At the same timemillions of images and spectra from dubiofossils(of unknown origin) and pseudofossils (abioticmimics) will also continue to accumulate yet theywill rarely appear in publication On the otherhand those abiotic mimics that have appeared inthe literature have often been mistaken usually be-cause of the lack of a comparative database for rec-ognizing biogenicity in such objects

Since the same techniques used to identify bonafide biosignatures can be used to distinguish themfrom their abiotic mimics it would be extremelybeneficial to have an easily accessible (ie web-based) database of both types of objects Clearlythere is a need for a common database of bona fidebiosignatures dubio-biosignatures and pseudo-biosignatures An investigator could search thedatabase for similar objects and obtain detailedinformation about them (eg protocol and tech-nique used to obtain the image and spectrum in-formation about the sample location and collec-tion protocol whether the object is common inthe sample whether the object has been found inother types of samples etc)

Such a database would improve systematicallythe ability to identify bona fide biosignatures andestablish efficiently and with a uniform level ofquality control the constraints of the varioushigh-resolution techniques The database wouldalso solve some of the problems posed by re-stricted (less accessible) sample sets We thereforerecommend that NASA investigate the utility ofestablishing and maintaining this type of data-base to avoid detection of false-positives for life

MARS THE CHALLENGES OF USINGMORPHOLOGICAL FOSSILS AS PROXIES

FOR EXTRATERRESTRIAL LIFE

Lessons learned from ALH84001

The initial interpretations of the ALH84001 me-teorite included nanometer-scale morphologicalfeatures likened to the remains of nanobacteria(McKay et al 1996) At the nanometer scale how-ever there is a convergence in the morphology ofbiosynthetic and inorganic forms since both typesof structures form in response to similar physicaland chemical processes (eg diffusion mini-mization of surface energy etc) This conver-

CADY ET AL362

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

gence of biosynthetic and inorganic forms at thesubmicroscopic scale underscores the problemsassociated with using only morphology as a bio-genic indicator Important lessons learned fromthe controversy over the origin of the putativenanobacteria discovered in ALH84001 are thelack of specificity of simple nanometer-scaleforms and our inability to assess the biogenicityof such features based on morphology alone Wetherefore recommend the development of ana-lytical methods that will allow non-destructiveintegrated morphological chemical and miner-alogical analysis at the nanometer scale

Bradley et al (1998) showed that nanometer-scale structures similar to those found inALH84001 could be produced artificially whenmetallic coatings were applied to mineral speci-mens a standard procedure used during thepreparation of specimens for high-resolutionSEM studies of the type reported by McKay et al(1996) However the conclusions of Bradley et al(1998) were challenged by studies in which thenanometer-scale structures were characterizedusing atomic force microscopy (Steele et al 1998)Indeed studies to examine the range of possibleinorganic processes and forms that could explainsuch features are warranted and mark a neces-sary condition for assessing biogenicity

In addition nanometer-scale fossilization pro-cesses are not well understood Consequently thetaphonomic framework for assessing the potentialbiogenicity of nanostructures is poorly definedThe structures reported by McKay et al (1996)could represent microbial preservation by eitherpermineralization or the wholesale mineral re-placement of microbial cells In the former case thepotential to preserve cellular components such asa cell wall would be an important attribute nec-essary for demonstrating biogenicity Our pre-sent inability to conduct targeted sectioning ofspecific nanometer-scale features imaged by sub-microscopic-scale methods and to map their lightelement composition precludes any further as-sessment of the biogenicity of the purported mi-crofossils based on such criteriamdashclear prioritiesfor instrument development and future technol-ogy advances needed to support the study of pur-ported morphological biosignatures

Planetary protection-related issues

Regardless of the technique of investigation itwill be necessary to understand how alteration

and sterilization methods affect morphologicalbiosignatures It appears likely that samples re-turned from Mars or other planetary bodies willbe released from quarantine for scientific studyonly after they have been certified harmless orsterilized and shown to be devoid of biohaz-ardous components A number of potential ster-ilization methods have been proposed includingdry or wet steam heat g- or electron beam radi-ation alkylating chemicals such as formaldehydeand ethylene oxide and oxidizing chemicals suchas hydrogen peroxide chlorine dioxide ozoneparacetic acid or combinations of one of more ofthese with plasma It appears probable that theprotocol selected for sterilization will employ dryheat (eg 135degC for 24 h the method used in 1976to sterilize Viking spacecraft) g-irradiation (60degC1 Mrad dosage) or a combination of both tech-niques A new report from the National ResearchCouncilrsquos Committee on the Origins and Evolu-tion of Life titled The Quarantine and Certificationof Martian Samples (2002) updates and extendsthese recommendations for sterilization Regard-less of the method chosen for sterilization we rec-ommend that thorough studies be carried out inadvance of sample return to determine what ef-fect if any the selected protocol may have on thefidelity of morphological biosignatures

EXAMPLES OF INSTRUMENTS CAPABLEOF IMPROVING OUR ABILITY TO

ASSESS BIOGENICITY

Time-of-flight secondary ion mass spectrometry(ToF-SIMS) and SEM

Techniques such as ToF-SIMS combined withhigh-resolution field emission gun SEM (FEG-SEM) provide a means to relate morphologicalfeatures to chemical signatures (Fig 2) By mark-ing the sample or using tell-tale features on thesurface of a sample it is possible to gain in situsurface sensitive data of both positive and nega-tive ions from 0 to 10000 AMU at very high peakresolutions (1000ndash6000) with a beam spot sizethat ranges from 50 to 200 nm depending on theinstrument The distribution of any peaks in thespectra can then be correlated with features onthe surface of the sample Although the inclusionof a secondary electron detector in the chamberof a ToF-SIMS instrument can generate SEM im-ages they generally suffer from low resolution

MORPHOLOGICAL BIOSIGNATURES 363

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

It is possible at the present time to overcome thisobstacle by imaging the same area analyzed witha ToF-SIMS instrument with high-resolution SEMafter ToF-SIMS analysis This technique has al-ready been successfully applied to fossilized bac-terial biofilms (Toporski et al 2001)

Although the combination of surface mappingand surface analysis using ToF-SIMS with FEG-SEM imaging is extremely promising this com-bination of non-destructive techniques is still inits infancy and we recommend further develop-ment be pursued Specific recommendations arean improved imaging capacity and mass resolu-tion for the ToF-SIMS an improved understand-ing of the mechanisms by which the spectral signatures are produced (a 61 AMU shift some-times occurs with higher-molecular-weight com-pounds) an increase in the database of informa-tion about various types of microbes (eg hopaneshave been detected with this instrument) andcorrelation of chemistry with morphology duringsequential stages of preservation

Analytical high-resolution TEM (HRTEM) ionmicroprobe Raman spectroscopy andsynchrotron x-ray tomography

Again the aim is to correlate chemical mole-cules of organic remains with morphological fea-tures indicative of their cells or cellular remainsand to determine the technological limitations ofinstruments that have yet to be fully exploitedOther than being evidently ldquocarbonaceousrdquo thechemical composition of Precambrian micro-scopic fossils is generally not known informationthat if preserved could provide valuable insightinto the original biochemical makeup and physi-ological capabilities of preserved microorgan-isms Additional methods that have yet to be systematically utilized to correlate micromor-phology with chemical composition include (1)analytical HRTEM equipped with high-resolu-tion spectrometers (eg an electron energy lossspectrometer) (2) ion microprobe analyses of thecarbon (and other elemental) isotopic composi-tion of individual microfossils to reveal aspectsof their physiology a technique that has beenused with notable success in an initial pilot study(House et al 2000) (3) Raman spectroscopy ofmicroscopic fossils to reveal the chemical makeupof their preserved cell walls a novel techniquethat has been used to demonstrate the presenceof the molecular signatures of disordered and

graphitic carbon in 650-million-year-old micro-fossils from Kazakhstan and 1878-million-year-old microfossils from the Gunflint Formation inCanada (McHone et al 1999) and (4) synchro-tron x-ray tomography of rock-embedded micro-fossils a new non-intrusive and non-destructivemeans of fossil detection and chemical character-ization that is of special interest for analyses ofmaterials for which only small amounts of sam-ples are available for study such as those plannedto be returned from Mars (Tsapin et al 2000) Asthese and related techniques have been shown al-ready to be highly promising for chemical char-acterization of minute ancient fossils we recom-mend major emphasis be placed on their furtherdevelopment and refinement

New application for establishing biogenicity traceelement biosignatures

Many microbes secrete polymers [extracellularpolmeric substances (EPS)] outside of their cellEPS can form either dense structures such assheaths or capsules that encapsulate cells ormore diffuse slime matrices that bind togethergroups of cells The potential functions of EPS aremany but controlling the chemical microenvi-ronment surrounding cells is among the most im-portant (ie Decho 1990) In microbial commu-nities organisms exist in a common EPS (ldquoslimerdquo)matrix These materials hold great taphonomicimportance because they can control aspects ofthe chemical microenvironments that promoteearly diagenetic mineralization a key factor inmicrobial fossilization (Farmer 1999) Cell wallsand extracellular exopolymers have many nega-tively charged surface sites that bind metalliccations and cation complexes These include an-ionic carboxyl and phosphoryl groups that are re-sponsible for absorbing and concentrating manyof the metallic cations required for growth EPSis known to bind a wide variety of metals in-cluding Pb Sr Zn Cd Co Cu Mn Mg Fe Agand Ni (Decho 1990 and references therein) Inaddition cation binding sites can also catalyze theprecipitation of initially disordered mineralphases (eg Ferris 1997)

While the property of metal binding and min-eralization can promote microbial productivityby providing ready access to high concentrationsof micronutrients required for growth the con-tinued precipitation of minerals can eventuallylead to the fossilization and demise of cells The

CADY ET AL364

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

concentration of these metals above backgroundlevels presents an interesting possibility for thedetection of organisms even after organic mate-rials have been degraded (Farmer 1999 Conradand Nealson 2001) The primary difficulty at pre-sent is to identify techniques capable of accu-rately mapping variations in trace element con-centrations relative to suspected microfossilremains in ancient or extraterrestrial materialsThe mapping of low-abundance trace metals us-ing standard electron and ion microprobe analy-

sis is difficult and presently subject to many un-certainties because of beam damage and poorcounting statistics However recent adaptationsof analytical HRTEM and synchrotron x-ray to-mography (as discussed above) may ultimatelyprovide a useful approach Used in conjunctionwith other techniques trace element mappingcould provide an additional criterion for demon-strating the biogenicity in ancient terrestrial andextraterrestrial materials and we recommendfurther study of this topic

MORPHOLOGICAL BIOSIGNATURES 365

FIG 2 A carbonate globule in specimen 198 from ALH84001 (an external chip of the meteorite known to containterrestrial bacteria) demonstrates the utility of combining a high-resolution FEG-SEM investigation with a ToF-SIMSanalysis (Steele et al 2000) See text for details A Spectrum in the 400ndash600 AMU range that contains predominantpeaks at 441 455 469 533 and 547 AMU B FEG-SEM image of the carbonate globule analyzed C Backscatter SEMimage of the same globule The arrows point to the ldquowaistrdquo of what appears to be two intergrown globules D ToF-SIMS map of the distribution of peaks between 533 to 561 AMU Boxes 1 and 2 in B and D represent the areas wherethe 533 AMU peaks are concentrated E Remains of bacterial cells surrounded by an organic film these features wereonly seen in areas 1 and 2 of B and D The scale bar in BndashD equals 10 mm The scale bar in E equals 1 mm

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

LIMITS OF TECHNOLOGY ACAUTIONARY NOTE

The inability to detect morphological biosigna-tures and prove their biogenicity may not be hin-dered by the lack of technological advances or inadequacies in our knowledge of the processesresponsible for their preservation On the con-trary extraterrestrial objects may lack conclusiveevidence of being formed biologically It isstrongly recommended that in addition to fund-ing projects aimed toward optimizing instru-mentation to detect morphological biosignaturesthat funding be allocated to projects that seek toexpand our knowledge base regarding the pro-cesses by which objects that mimic morphologi-cal biosignatures could form in extinct and pastldquomicrobial ecosystemsrdquo on Mars

SUMMARY OF RECOMMENDATIONS

Our review of the various lines of research re-garding morphological biosignatures leads us torecommend the following

1 Exploit existing high-resolution imaging andspectroscopic techniques for detecting mor-phological biosignatures and develop and op-timize instruments capable of acquiring multi-ple biosignatures from the same morphologicalobject in order to demonstrate the object wasonce a microbial cell or a microbially influ-enced sedimentary structure

Improve the imaging capacity and mass reso-lution for the ToF-SIMS and identify the mech-anisms that produce ToF-SIMS spectral signa-tures

Develop and optimize analytical HRTEM ionmicroprobe Raman spectroscopy synchrotronx-ray tomography and related techniques

2 Exploit instruments capable of detecting fea-tures that distinguish morphological biosig-natures from their abiotic mimics to avoididentifying false-positives for life

Expand the use of analytical HRTEM to studymicrofossils pseudofossils and biogenic andabiogenic sedimentary structures

Develop analytical methods that will allownon-destructive integrated morphological

chemical and mineralogical analysis (includ-ing trace element anomalies) at the nanometerscale

3 Pursue research strategies that improve ourunderstanding of the processes by which mor-phological biosignatures form and become al-tered to maximize the ability to assess the biogenicity and indigenicity of all types ofmorphological fossils

Develop growth process models for stromato-lites and microbialites that include surfaceroughness of microbial mats or biofilms par-ticularly where sedimentation also occurs

Determine the fate of biofilms in modernecosystems the fossilization potential of thesestructures and how these factors relate to es-tablished criteria for demonstrating biogenic-ity

Identify criteria that can be used for demon-strating the biogenicity of bona fide microfossilsat the electron microscopy scale

Determine the nature of phenotypic changes inresponse to environmental perturbations andwhether those changes alter the susceptibilityof a microorganism to fossilization

Develop the use of techniques that can be usedin modern ecosystems and in laboratory ex-periments to document geomicrobiologicalprocesses and fluid geochemistry at the cellu-lar scale

Identify the extrinsic geochemical characteris-tics that affect preservation in environmentsrecognized as analogs for extant and past Mars

Conduct systematic studies regarding the ef-fects of mineral diagenesis (recrystallizationand phase transformations) on taphonomiclosses of biosignatures

Integrate data related to biosignature forma-tion alteration and destruction across a widerange of spatial scales

Determine what effects the selected sample re-turn sterilization protocol may have on the fi-delity of mophological biosignatures

4 Pursue research strategies (such as those inRecommendation 3) that quantify the funda-mental processes that govern the formation ofpseudofossils abiotic stromatolites and non-biologically influenced sedimentary structuresand fabrics that mimic biologically influencedtypes

CADY ET AL366

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

ACKNOWLEDGMENTS

SLC acknowledges NASA Grant NAG 5-9579and NSF Grant EAR-0096354 JWS acknowledgesNASA Grant NAG 5-12357

ABBREVIATIONS

DLA diffusion-limited aggregation EPS ex-tracellular polmeric substances FEG-SEM fieldemission gun scanning electron microscopyHRTEM high-resolution transmission electronmicroscopy SEM scanning electron microscopyTEM transmission electron microscopy TESThermal Emission Spectrometer ToF-SIMS timeof flight secondary ion mass spectrometry

REFERENCES

Allen CC Albert FG Chafetz HS Combie J Gra-ham CR Kieft TL Kivett SJ McKay DS SteeleA Taunton AE Taylor MR Thomas-Keprta KLand Westall F (2000) Microscopic physical biomarkersin carbonate hot springs implications in the search forlife on Mars Icarus 147 49ndash67

Bandfield JL Hamilton VE and Christensen PR(2000) A global view of Martian surface compositionsfrom MGS-TES Science 287 1626ndash1630

Banfield JF Moreau JW Chan CS Welch SA andLittle B (2001) Mineralogical biosignatures and thesearch for life on Mars Astrobiology 1 447ndash465

Beveridge TJ Hughes MN Lee H Leung KT PooleRK Savvaidis I Silver S and Trevors JT (1997)Metal-microbe interactions contemporary approachesAdv Microb Physiol 38 178ndash243

Bradley JP McSween HY Jr and Harvey RP (1998)Epitaxial growth of nanophase magnetite in Martianmeteorite Allan Hills 84001 implications for biogenicmineralization Meteorit Planet Sci 33 765ndash773

Buick R (1990) Microfossil recognition in Archean rocksan appraisal of spheroids and filaments from a 3500MY old chert-barite unit at North Pole Western Aus-tralia Palaios 5 441ndash459

Cady SL (1998) Astrobiology a new frontier for 21st cen-tury paleontologists Palaios 13 95ndash97

Cady SL (2001) Formation and preservation of bona fidemicrofossils In Signs of Life A Report Based on the April2000 Workshop on Life Detection Techniques Committeeon the Origins and Evolution of Life National ResearchCouncil The National Academies Press WashingtonDC pp 149ndash155

Cady SL and Farmer JD (1996) Fossilization processesin siliceous thermal springs trends in preservationalong thermal gradients In Evolution of HydrothermalEcosystems on Earth (and Mars) Ciba Foundation Sym-posium 202 edited by GR Bock and JA Goode JohnWiley and Sons Chichester UK pp 150ndash173

Christensen PR Bandfield JL Clark RN Edgett KS

Hamilton VE Hoefen T Kieffer HH Kuzmin ROLane MD Malin MC Morris RV Pearl JC Pear-son R Roush TL Ruff SW and Smith MD (2000)Detection of crystalline hematite mineralization onMars by the Thermal Emission Spectrometer evidencefor near-surface water J Geophys Res 105 9623ndash9642

Committee on the Origins and Evolution of Life NationalResearch Council (2002) The Quarantine and Certificationof Martian Samples The National Academies PressWashington DC

Conrad PG and Nealson KH (2001) A non-Earthcen-tric approach to life detection Astrobiology 1 15ndash24

Decho AW (1990) Microbial exopolymer secretions inocean environments their role(s) in food webs and ma-rine processes Oceanogr Marine Biol Annu Rev 28 73ndash153

De Leeuw JW and Largeau C (1993) A review of macro-molecular organic compounds that comprise living or-ganisms and their role in kerogen coal and petroleumformation In Organic Geochemistry Principles and Ap-plications edited by MH Engel and SA MackoPlenum Press New York pp 23ndash72

Des Marais DJ Beard B and Canfield D (2003) Stableisotopes In Biosignatures for Mars Exploration edited byJF Kerridge NASA publication SP-XXX US Govern-ment Printing Office Washington DC (in press)

Farmer JD (1999) Taphonomic modes in microbial fos-silization In Proceedings of the Workshop on Size Limits ofVery Small Organisms Space Studies Board NationalResearch Council National Academies Press Wash-ington DC pp 94ndash102

Farmer JD and Des Marais DJ (1999) Exploring for arecord of ancient Martian life J Geophys Res 10426977ndash26995

Ferris FG (1997) Formation of authigenic minerals bybacteria In Biological-Mineralogical Interactions editedby JM McIntosh and LA Groat Mineralogical Asso-ciation of Canada Ottawa pp 187ndash208

Ferris FG Schultze S Witten T Fyfe WS and Bev-eridge TJ (1989) Metal interactions with microbialbiofilms in acidic and neutral pH environments ApplEnviron Microbiol 55 1249ndash1257

Fortin D Ferris FG and Beveridge TJ (1997) Surface-mediated mineral development by bacteria In Reviewsin Mineralogy Vol 35 Geomicrobiology Interactions Be-tween Microbes and Minerals edited by J Banfield andKH Nealson Mineralogical Society of America Wash-ington DC pp 161ndash180

Friedmann EI Hua MS and Ocampo-Friedmann R(1988) Cryptoendolithic lichen and cyanobacterial com-munities of the Ross Desert Antarctica Polarforschung58 251ndash259

Grotzinger JP and Chan K (1999) Stromatolite mor-phogenesis [abstract] GSA [Abs] 30(7) 393

Grotzinger JP and Rothman DR (1996) An abioticmodel for stromatolite morphogenesis Nature 383423ndash425

Hesse R (1989) Silica diagenesis origin of inorganic andreplacement cherts Earth Sci Rev 26 253ndash284

Hofmann HJ (1973) Stromatolites characteristics andutility Earth Sci Rev 9 348ndash349

MORPHOLOGICAL BIOSIGNATURES 367

Hofmann HJ (1987) Precambrian biostratigraphyGeosci Canada 14 135ndash154

House CH Schopf JW McKeegan KD Coath CDHarrison TM and Stetter KO (2000) Carbon isotopecomposition of individual Precambrian microfossilsGeology 28 707ndash710

Jolliff BL Moersch J Knoll A Morris RV ArvidsonRE Gilmore M Greeley R Herkenhoff K Mc-Sween H and Squyres SW (2000) Remotely-sensedgeology from lander-based to orbital perspectives re-sults of FIDO Rover field tests In Concepts and Ap-proaches for Mars Exploration Lunar and Planetary In-stitute Houston pp 164ndash165

Jorgensen BB and Des Marais DJ (1990) The diffusiveboundary layer of sediments oxygen microgradientsover a microbial mat Limnol Oceanogr 35 1343ndash1355

Kardar M Parisi G and Zhang Y (1986) Dynamic scal-ing of growing interfaces Phys Rev Lett 56 456ndash472

Knoll AH (1985) Exceptional preservation of photosyn-thetic organisms in silicified carbonates and silicifiedpeats Philos Trans R Soc Lond [Biol] 311 111ndash122

Knoll AH and Semikhatov MA (1998) The genesis andtime distribution of two distinctive Proterozoic stro-matolitic microstructures Palaios 13 407ndash421

McHone JF Kudrysvtsev AK Agresti DG WdowiakTJ and Killgore M (1999) Raman imagery of martianmeteorites [abstract 1896] In Lunar and Planetary Sci-ence XXX [book on CD-ROM] Lunar and Planetary In-stitute Houston

McKay DS Gibson EK Jr Thomas-Keprta KL ValiH Romanek CS Clemett SJ Chillier XDF Maech-ling CR and Zare RN (1996) Search for past life onMars possible relic biogenic activity in Martian mete-orite ALH84001 Science 273 924ndash930

Moersch JE Farmer JD and Baldridge A (2001)Searching for evaporites on Mars remote sensing of ter-restrial analogs for putative martian paleolake basins[abstract] GSA [Abs] 33(6) A403

Pope M and Grotzinger JP (1999) Controls on fabric de-velopment and morphology of tufas and stromatolitesuppermost Pethei Group (18 Ga) Great Slave LakeNorthwest Canada In Carbonate Sedimentation and Dia-genesis in the Evolving Precambrian World Special Publi-cation 67 Society of Economic Paleontologists and Min-eralogists Tulsa OK pp 103ndash121

Schopf JW (1999) Fossils and pseudofossils lessons fromthe hunt for early life on Earth In Size Limits of VerySmall Microorganisms Proceedings of a Workshop Na-tional Academy Press Washington pp 88ndash93

Schopf JW and Walter MR (1983) Archean microfos-sils new evidence of ancient microbes In Earthrsquos Earli-est Biosphere Its Origin and Evolution edited by JWSchopf Princeton University Press Princeton NJ pp214ndash239

Semikhatov MA Gebelein CD Cloud P AwramikSM and Benmore WC (1979) Stromatolite morpho-genesismdashprogress and problems Can J Earth Sci 19992ndash1015

Steele A Goddard DT Stapleton D Toporski JKWSummers D Martill D and Beech IB (1998) The na-

ture of artefact in the imaging of the martian meteoriteALH84001 [abstract] GSA [Abs] 30(7) 150

Steele A Goddard DT Stapleton D Toporski JKWPeters V Bassinger V Sharples G Wynn-WilliamsDD and McKay DS (2000) Investigations into an un-known organism on the martian meteorite Allan Hills84001 Meteorit Planet Sci 35 237ndash241

Steele A Rummel J Baross J and Schopf JW (2003)Biological considerations of biosignature research forMars In Biosignatures for Mars Exploration edited by JFKerridge NASA publication SP-XXX US GovernmentPrinting Office Washington DC (in press)

Summons RE (1993) Biogeochemical cycles a review offundamental aspects of organic matter formationpreservation and composition In Organic GeochemistryPrinciples and Applications edited by MH Engel andSA Macko Plenum Press New York pp 3ndash21

Summons RE Albrecht P Chang S McDonald Gand Moldowan JM (2003) Molecular biosignatures InBiosignatures for Mars Exploration edited by JF Ker-ridge NASA publication SP-XXX US GovernmentPrinting Office Washington DC (in press)

Toporski JKW Steele A Westall F Avci R MartillDM and McKay DS (2001) Morphological and spec-tral investigation of exceptionally well preserved bac-terial biofilms from the Oligocene Enspel formationGermany Geochim Cosmochim Acta 66 1773ndash1791

Tsapin AI Storrie-Lombardy MC and Nealson KH(2000) Using X-ray computer tomography (CT) forimaging biological systems inside rocks Origins LifeEvol Biosphere 30 388

Walter MR ed (1976) Developments in SedimentologyVol 20 Stromatolites Elsevier New York

Walter MR Des Marais D Farmer JD and HinmanNW (1998) Lithofacies and biofacies of mid-Paleozoicthermal spring deposits in the Drummond BasinQueensland Australia Palaios 11 497ndash518

Westall F (1997) The influence of cell wall compositionon the fossilization of bacteria and the implications forthe search for early life forms In Astronomical and Bio-chemical Origins and the Search for Life in the Universeedited by C Cosmovici S Bowyer and D WerthimerEditori Compositrici Bologna pp 491ndash504

Westall F (1999) The nature of fossil bacteria a guide tothe search for extraterrestrial life J Geophys Res 10416437ndash16451

Wynn-Williams DD Edwards HGM and Garcia-Pichel F (1999) Functional biomolecules of Antarcticstromatolitic and endolithic cyanobacterial communi-ties Eur J Phycol 34 381ndash391

Address reprint requests toDr Sherry L Cady

Department of GeologyPortland State University

1721 SW Broadway 17 Cramer HallPortland OR 97201

E-mail cadyspdxedu

CADY ET AL368