NEWSANP V?FW$
Theearliestmemoriesof life on Earth
John M. Hayes
How old is life on Earth? lts chemical traces may have been recognizedin rocks fiom Greenlandthat are more
than 3.8 billion years old, at least 300 million ybars older than any previousevidence.
Irsre is Greenlandicfor'ancient thi[g'.
The ltsaq G[eiss Complex,a geological
region in southern West Greenland. is
rndeedancient. And. importanlly.
it conraitrsthe metamorphosedremainsof the
Eanh s oldest-knownsediments,providine e\idencethat therewassomethinglike
.rn ocean 3.85 billion years (Gyr) ago.
L nrii now it has seemedthat thesesedimenlsmight be relicsof a prebioticworld,
bul Lheanicle on page55 of this issuer
may well removethat option. Specifically,
new analysesof the tracesof organiccarbon presentin the_sedimentschangeour
estimatesof its ''C/''C ratio. The new
value is within the rangecharacteristicof
biological debris from sedimelts 2.7 to
3.5Cyr old.
In modem oceans, both calbonates
and organic moleculesare derived from
COj through precipitationand photoslnthesis. Isotopic effects associated
with these reaction pathwayslead to an
isotopic difference, or ftactionation, of
about 25 parts per thousand,with the
organicmaterialsbeing depletedin r3C
relative to the carbonates.A similar
isotopic signal persists,with some variations',back throughtime to rocksthat are
3.5Gyr old.
The new finding seemsto e)'tendthat
record to the very bottom of our planet's
sedimentarypile, cruciallyaltedngearlier
viewsof theseoldestsedimentsand leaving almost no time betweenthe end of
the 'late healy bombardment'of bodies
within the inner Solar Systemby giant
meteorites and the first appeararce of
life. Two,earlier paperupublishedin this
joumal"" conclude that these asteroid
impactsmight have sterilizedthe planet
as recentlyas 3.3 Glr ago.It could follow
that the 3.85-Gyr-old Itsaq organic
material de ves from biochemical
processesthat developed with breathtaking rapidity after the last large impact,
or eventhat it is a product of a biota that
was wiped out, then supplantedby our
ancestors,
Can this all be true?
Isotopic fractionations result frorn
physical phenomena associated with
chemical reactions.The task now is to
determinewhetherthe observedfractionations are due to biochemicalprocesses,
other low-temperature chemical reactions, or some post-depositionaleffect.
The magnitudeof the observedfractionations favou$ the biochemicalaltematrye.
NATURE. VOL384 . 7 NOVEIVIBER
1996
but does not lit perfectly.About half of
the new analysesi[dicate fractionations
larger than any in the next-oldestorganic
mate al, found in Australianrocks about
3.5 Gyr old, which also contain bacte um-like bodies and laminated sedimentary structures that resemble algal
mats.Thesefeaturesare not considereda
sure sign of life, but the carbon-isotope
differences
are roughlyequalfo rhosein
All previous analysesof such ancient
mate als have dealt onlv with bulk
organic carbon,which showsall too well
the hazardsof an earlv birth. Anv sediment that sits around for long is lilef to
be buried by still further sedimentation
and thendraggedinto rhe crust'stectonic
traffic. The Itsaq rocks were churned
downwards within a billion years ard
heatedto 500'C al pressures
of 5.000ar-
A chip off the old block.A geologist'shammerleans againstthe oldest-knownsedimentsin the
world, on the bleak Akilia island off the Greenlandcoast. Despite metamorphism,they show
signs of ancientlife.
still younger rocks that have compelling
evidencefor biology, so many investigatols have proposedthat life was present
3.5 C1,rago and lhat il was controlling
the redox chemistryof carbonat Earth's
surface.
The new Itsaqdatawere gather€dusing
an ion-microprobe mass spectrometer
that can selectivelyanalysefeatures as
small as 20 pm. That selectivityhas been
used to examine carbon occluded in
apatitegains (seeFig. I on page 56).
Apatite, a phosphatemineral, is known
for its biological associations,but there
have been no previousmeasurementsof
the isotopic composition of organic
carbon closely associatedwith it. Interpretatiolrs of the new data may evolve as
Doints of comDarisonbecome available.
,qJ[nougnIne age oI Ine samplesanq tne
tantalizingresultscommandinterest and
prominentpublication,the noveltyshould
alsoinsoirecaution.
mospheres.From the point of view of
anyoneseekingto reconstructdetails of
their initial. low-temperature environmert, they'vebeenbadly overcooked.Althoughsomepreviouslyanalysedsamples
containasmuch as2.5per centcarbon',it
looked more like graphitethan biological
debds.
During graphitization,mobile chemical
producls wheezeor ooze trom the geological pressurecooker wherever a vent
can be found. lf carbonatemineralsare
present,their carbon is thro\r'n into the
chemicalmelde.In lhe process.isotopic
evidenceis first altered (as some carbon
depletedin IrC is lost) then destroyed(as
carbonwith metathe residueexchanges
morphic fluidsTt).For this reason,results
of earlier analysesof the Itsaq matedals
were set cside Js lhey were received.
becausethey showed the damagedone
by graphitization. When the authom
examined bulk rather than apatite2t
NEwsAtrtDvlEws
OPTICS
associated carbon, their results agreed
'r-ith tho6eof Dreviousanalyses.
What is missing is alr understanding of
\*hr the aDatite sbould act as a safeJeoosit vault. Unlil that is known. other
ouestionswill claiE attention. First. why
;e dr" t'c/"c ratios so variable and, in
half of the cases,so low? Biologicalfractionatiotr is usually very consistent,and
ratios like those in the lower half of the
preseDtdistribution do not appear for
more thad a billion years,when they are
rhought to be producedby methale-consumingbacteria.Theseorganismsrequire
orygen, which fits well with other indicato$ of increasingoxygenatiol 2.7 Gyr
ago,but not 3.85Glr ago.Second,is there
a plausible non-biologicalcausefor the
fractionation,either an isotopeeffect associatedwith one of the infinite number
prebioticchemicalreacof uninvestigated
tions, or some obscurepost-depositional
effect Decuiiarto microscopicinclusio$
in apaiite? These are worries and longshots.It's most likely that the new results
are indeed evidencefor life 3.85 billion
yearsago.
The use of an ion microprobe to
measure carbon isotopes is lrew. The
probe works by firing caesium-ionbullets
into a polished chip to break up crystals
and molecules at its surface. Negative
ions (C- in this case) are extracted
and accelerated,and the isotopes are
separatedby a mass
electromagnetically
spectrometer.It is assumed that any
instrumental effects biasing the ratio observedfrom the thin films of carbonwithin
the mineral matrices can be removed by
comparisonwith a referencesample of
pure gaphite. Testsindicating consistency
with results of conventional analysesare
imDressive.but more documentationof
the probe's suitabilityfor carbon-isotope
analyseswill be welcomed.As soon as that
is available, curators of collections of
martial meleo les and of ancientmicrofossilsshould start getting in line.
!
X-raylensesnearreality
JeromeHastinEs
beamsprohibitssuchapplications.
The existingoptics fall into a few categories:reflectingopticssuch as grazingincidenceX-ray mirrors; tapered capillaries; and vadous typesof Fresnelzone
plates.Refractivelenses,routinelyusedin
visibleJightoptics,hadbeenthoughtto be
impracticalat X-ray wavelengths.In the
X-ray regime, the difference in the index
of reftaction between air and ordinary
materials is tiny - typically of the order
of 10- - which appearedto present an
insurmountableproblem. To maximize
- the effect, heavy-element
'ff mate als were orooosed.
but they have higi absorption. Light elements do
not, but with their low
refractive-indexdecrement
they would have long focal
length and thus provide
little practical focusing of
the X-ray source.
Snigirevand colleagues'
have now proposed and
demonstrateda compound
refractivelens for hard (14
keV) X-rays that overcomes the difficulties of
A seriesof
earlierschemes.
N refractive lenses has a
The EuropeanSynchrotron
RadiationFacilityin Grenoble, focal length shorter than a
France.compoundrefractivelenses may ioon make its singe lensby a factor ofN,
and, in aluminium,absorpintenseX-raybeamsmuchmoreeffective.
tion is slight for X-ray
lens,the X-ray equivalentof conventional wavelengthsbelow one ingslrom, so a
opticsfor visiblelight. On page49 of this seriesof tens to hundredsol lensescan
issues,Anatoly Snigirev and colleagues be used.The authorsmade their lens by
descdbethe first signi{icantstep towards simply drilling small round holes in an
that goal. This shouldallow us to reach a aluminium block (see Figs 1 and 3 on
submicrometrefocus with much greater pages49 and 50,respectively).
This 'crude' technology permitted a
efficienrythan before.
The desirein a multitude of scientific demonstrationof their ideasalld showed
disciplinesfor intensesubmicrome'.reX- the potentialfor future optics.With more
John M. Hayes is at the Woods Hole ray beamshas driven the constructionof sophisticatedfabricationtechniques,one
Oceanographic lnstitution, MS8, 360 new electron-storage-ng X-ray sources can imagineextendingtheselensesfrom
Woods Hole Road, Woods Hole, Mass- in Europe (the European Synchrotron cylindersto spheresto providefocusingin
Radiation Facility in Grenoble, France), two dimensions.Such lensesshouldthen
achusefts 02543 7539, USA.
the United States(the AdvancedPhoton be commonplaceat synchrotronsources,
Source in Argonne, Illinois) and Japan and the full potential of the X-ray beams
1. Nutman,A, P, Mccregor,V R.. Friend,C, L. R..
(the Sp ng-8Projectin Harima). The obBennet!,V c, & Kinny,P. D. Precantu.Res 74,I-39
(1996)
jective is to perform the various 'standard'
2. Mojzsis,S. J, er at Natu.e384, 55-59 {1996)X-ray analytical techniques of diffraction
3. Hayes,l. M- in Far,life on Eaflh (ed. Benerson.
S ) 22G-236 (ColumbiaUniu Press,NewYo ,
and spectroscopyon individual particles
1994).
or grams.
4. lvahe.. K. A. & Stevenson, O.I Nature 33\ 6!2-6!4
(1988)
these measurementscould provide
5. Sleep,N. H,, Zahnre,K.l-, Kasung,l. F. & Morowitz,
information about the atomic and elec- Jerome Hastings is at the Brookhaven
H, J, N€tu.e 3/r2, 139-142 (1949).
6. Hayes,l. M , Kapran.t. R. & wedeklng.K. w in t e
tronic structure of systemsof chemical, National Laboratory, 75 Brookhaven
E th s Eadiest Biosphere: fts OnEinand Evalution
biological and technological interest, Avenue, Building 725, Upton, New York
(ed.S.hopf, J. W) 9}134 (Pri.cetonUnM Press,
ranging from the characterization of an 77973, USA.
1983).
7. Welekire, ^, W & Haves,J,M,n adva..es tn Ogaac
individual grain in a commercial catalyst
@ochemistry (ed. Bjoroy. M,) 546-553 {Wile, New
to measudngthe strain in the metallic
r. Broelbdch.D. H. & Thiel,D..,. Ree Sci.lnsrum,66,
Yo.k,€43),
2059-2063 (1995).
a. schidlowski,M. &Ahaon, P.in Ea y otganicEwtuno^
interconnections of large integrated
2.
a',8. etal, ReL S.i.lnstm,64, 22a7-22a9 r1995).
(eds schidrowski,M., Gotubic,s., Kimbeney,[4, M.,
-rudilger,
circuits. But the limited ability of avail- 3. Snigr.ev,A" Kohn,v, snlgllM,l. & Lengeler,B. Na.ue
[4cKndy,D. V, &
P A.r1.7 175 so'i.ger,
l
$4,49_51(1996). i
Berlln,1992),
able X-ray optics to focus the X-ray
NATURE. VOL384 . 7 NOVEMBER
1996
X-nev sourceshave already evolved from
the original sealed vacuum tubes into
sophisticatedelectron-storagerings, and
free-electron X-ray lasers may be just
beyond the horizon. These advances
promiseintense,highlyparallel and tuneable photon beams.To focusand manipulale this radiation, a variety of X-ray
optical devices have been developed,
ranging from concentrators,in the foml
of taDeredcaoillaries'. to Fresnel zorc
plates'.But a dreamof X-ray opticiansfor
some time has been to nake a refractive