"--------------------------, . ,-NA,--TU-,-",R",E=--YOL-=.=''--':.;:J.:;.6...:.1,;.6'"'1I::..1'1l:..;:..=llc...:...;IVlI,;.7'--- REVlEW ARTICLE--------~----------= ..~s

Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate Robert J. Charlson·, James E. Lovelockt, Meinrat O. Andreae* & Stephea G. Warren·

• Deparu.cm of Atmospheric Sciences AK~, UnivenilY of Washin'lon, SUllie, Washinaton 9819S, USA t COOrnM Mill Eapc:rimental Slation. launcntOt1, Cornwall PUS 9RY, UK*Depan.all of Ottano.raphy, Florida Stale UniversilY, Tallahassee, Florida 32306, USA

The major source of cloud-condensation nuclei (CCN) over the oceans appears to be dimethylsulphide, which is produCftl by planktonic algae in sea II/ater and oxidizes in the atmosphere to form a sulphate aerosol. Bect:Juse the rej1ectance (a/~do) ofclouds (and thus the Earth's radiation budget) is sensitive to CCN density, biological regulation of the climate is possible through the effects of temperature and sunlight on phytoplankton population and dimethylsulphide production. To coullteract the warming due to doubling of atmospheric COb an apprOXimate doubling of CCN would be needed.

CLIMATIC influences of the biota are usually thought of in connection with biological release and uptake of CO2 and CH. and the effect of these gases on the infrared radiative propenies of the atmosphere'. However, the atmospheric aerosol also panicipales in the· radiation balance, and Shaw2 has proposed that lhe aerosol produced by the atmospheric oxidation of sulphur psts from the biota may also affect climate. So far the physical and biological aspects of this intriguing hypothesis have nOl been quanlified, bUlthree recent discoveries may make this possible for the remote marine atmosphere. ( II Most species of phytoplankton, ubiquitous in the oceans, excrete dimethylsul­phide (OMS) which escapes to lhe air where it reacts io form

,a sulphale and methane sulphonate (MSA) aerosol. (2) This non-sea-salt sulphale (NSS-SO;-) aerosol is found everywhere in the marine atmospheric boundary layer. (3) Aerosol panicles which act as c1oud-condensalion nuclei (CCN) in the' marine atmosphere are principally, perhaps almost e~clusively; these same NSS-SO;- particles.

I n this paper we show thai emission ofDMS from phytoplank­ton is sufficientlo justify its consideration as the gaseous precur­sor of CCN in the remote and unpolluled marine atmosphere. We re-examine the physical role of the sulphale aerosol in atmospheric radiative transfer, panicularly in clouds, which are responsible for mOSI of the Eanh's albedo, and eSlimale the sensitivity of the Eanh's temperature to changes in the abund­ance of CCN. Finally we examine the geophysiologyJ of the system comprising phytoplll,nklOn, OMS, CCN and clouds, as a pUlative planetary thermostat.

a volcano is of regional imponance only. Large eruplions., on the other hand, which emit enough gaseous sulphur compounds to influence wider areas, are relatively rare events. For this discussion, we shall assume that the contribution of CCN from volcanic sulphur to the global atmosphere ~ proponional to its contribution to the total sulphur flux, that is, 10-20% of the nalural componenl at presenl.

Volatile sulphur compounds are emitted both by tenestrial and marine biOla. The marine emissions are almost exclusively in the form of OMS, whereas the emissions from land are in a varielY of chemical species, including H2S, OMS, methanethiol, CS" COS and others. This difference is related to the biological processes which are responsible for the production of the sul­phur volaliles. MOSI sulphur at the Eanh's 5urface is present as sulphate, which is the thermodynamically stable form ofsulphur in lhe presence of oxygen. Sulphate is reduced by organisms lhrough two mechanisms, 'assimilatory' lind 'dissimilatory' sulphate reduction. The dissimilatory pathway is restricted to sulphate-reducing bacleria in anaerobic environments; due 10

physical and microbial restrictions only a small fraction of the H~S produced by lhis process can escape to the atmosphere'. The products of the assimilatory palhway are a variety of organosulphur compounds, the largest fraction being the amino acids cysleine and melhionine, which are incorporated into proteins. '>

On the cOnlinenlS, the breakdown of organosulphur com­pounds during fermentalive decomposition of organic mailer is probably the most imponant mechanism for the release of sulphur volaliles. Due to the difficulty in obtaining represenlative data from land biomes, lhe magnitude or . ux is still ood.>:

'~~1:::l.. ~~~Jt-' . ate ge' r.Qm about 'J:

will be concluded below), w-e have to consider what processes land surface', ~r abollt 2 mm~ m~:'yr ;"which ~i~~rns out to be ~~-are.responsible-for-lhe·production-of-lhe·volalile sulphur·com-----a-sightly-sll1ll'ller-flux-per unil arerthll'n~i5'em·lt(Cd-fr"{rrTrthe----

pounds 'thal are the precursors of sulphuric acid in the almos­phere. In lhis discussion, we shal.1 ignore the penurbations of lhe almospheric sulphur cycle by manmade /luxes ofS02(moslly from burning of coal and oil) and shall consider only the natural /luxes, which curren·lly represenl about 50% of lhe tolal flux of gaseous sulphur to the atmosphere, and which slill dominate the almospheric sulphur cycle in lhe Soulhern Hemisphere'--.

The only significanl non-biological natural flux is lhe emission of S02 and H2S by volcanoes and fumaroles. This process releases of the order. of 0.4 Tmol yr-', about 10-20% of the total natural nux of gaseous sulphur to the atmosphere4

,5. The emission of sulphur by volcanoes is highl)' variable in space and lime. Consequenlly, the production of sulphate aerosol by the oxidation of volcanic sulphur during the quiescent stage of

oceans (aboul 3 ± 1.5 mmol m-2 yr- t ). Preliminary data from the

lropical regions suggest that there may be a net transpon of biogenic sulphur from lhe marine to the continental atmosphere, consistent wilh a smaller emission flux per unil area on the conlinents·. Coastal wetlands playa surprisingly small role in lhe global sulphur cycle; al present we estimate their contribu­tion to be about 2% of the total gaseous emissions5

. This is because although the emission per unit area may be relatively high, the lotal area of coastal wetlands is qllite small.

In the marine environment, most volatile sulphur is emitted in the form of OMS which is excreted by Iivina planktonic algae. In contrast to microbial decomposition which produces a com­pleli. mixture of volatile sulphur species, algal metabolism yields OMS as the only volatile sulphur species. 1be biological func­

REVEW ARTICt£--- -_. ._NArullli.- '"ot.. ~~!'_'.!!._A!~~..L 1~.7~---­

tion of the production of OMS is still unclear. The substance from which OMS ori,inates, dimethylsuJphonium propionate (OMSP), is imeonant in osmorqulation in a number of phyto­plankton types and also participates in the biochemical cycle of methionine. OMSP is also excreted by al,ae, and its break­down in sea water may release additional amounts of OMS.

Even thoulh OMS is ncreted by phytoplankton. its con­centration in surface sea water. and consequently its emission rate to the atmosphere, is only weakly correlated with the usual measures or phytoplankton activity, for example. chlorophyll concentration or I·C.uptake rate. In f.ct. the calculated IIUlt of OMS from the tropical ~ans. which hive a low primary productivity, is essentially the same per unit area (2.2 mmol m-, yr- l

) as from the much more productive temper­ate oceans (2.4 mmol m-2 yr-')'. Recent evaluations of seasonal

1Ieftects lO• sugest that the annually averaled lIull. from temper:

ate relions may even be considerably lower than the value liven here. Even the hilhly productive coastal and upwelling regions do not suppon much hi,her OMS fluxes (5.7 and 2.8 mmol m-~ yr-'). respectively. based on a larae body of data from several research groups'). It appean that, independent of the rate of primary production, the -warmest, most saline, and most intensely illuminated re,ions of the ~ans hl\'e the highest rate of OMS emission to the atmosphere. This is a key fact to keep in mind when discussinl possible climatic feedback mechanisms.

Two reasons can be given for this unupected behaviour. Fint, the concentration of OMS in surface sea wa(tr- depends not only on the rate of OMS production, but also on its. rate of removal. The two dominant removal processes arc ventilation to the atmosphere and the photochemical and microbial break­down of OMS in the water column'~·I.l. The rate of microbial removal of OMS will increase both with the density of bacteria

.which are able to metabolize OMS and, in a nonlinear fashion, with the concentration of OMS in sea water. The density of bacterioplankton is a function of the densities of phytoplankton and of grazing organisms. Due to these diverse and nonlinear relationships between the variables which relulate OMS produc. tion and removal, we cannot expect to lind a simple, linear relationship between phytoplank.ton density and OMS con· centration. Funhermore, the production rate of OMS shows variations over three orden of malnitude among difterent phyto­plankton species (ref. 14 and M.O.A.• unpublished data). It appean that some algal groups, such as the coccolithophorids, which arc most abundant in tropical, oligotrophic waten, have the h.ighest rate of OMS excretion per unit biomass. We· con· c1ude,therefore, that the global input of OM.S into the atmos­phere is much less sensitive'to changes in total primary produc­tion than to variations in phytoplankton speciation. In other words, we can accommodate rather large swings in marine productivity without changinl the climatic eftectsproduced by the sulphur cycle, and on the other hand, we could essentially eliminate the marine input-o ilIlphur to the atmos.2M.re without

~sIiiii" earitl ..-' . - -- ­-. ­eum _a . "2 tl I I 0 ,

levels, however, the fraction of 'MSA produ~ is less lhan 20%, IS suuested both by laboratory uperiments (ref. 19 lind unpub­lished manuscript by I. Barnes, V. Bastian and K. H. Becker, 1986) and by the ratio of 'MSA to NSS-.SO;- observed on marine aerosols from remote re,iont JQ.l'. 'MSA is mostly present as aerosol panicles (M.O.A., unpublished data) and due to its hi,h solubility is also likely to be eftective as CCN.

The oxidation of OMS to OMSO by the 10 radical, which has recently been proposed' (unpublished manuscript b)' I. Barnes II/ all. is probably not imponani as a sink for OMS because the concentrations of OMSO in the atmosphere and in marine rain /\ave been found to be very low compared to those of OMS. MSA and NSS-SO;- (ref. 22 and M.O.A., unpub­lished data). The pnsent evidence thus suuests that SO~ is the major oxidation product of OMS in the unpolluted marine atmosphere. In the presence of cumulus clouds, which are

-abundant in the marine boundary layer, conversion of SO, to NSS-SO~- is rapid and significantly exceeds the rate of dry deposition of SO~ (refs 18,21).

The rates of OMS oll.idation obtained from ,laboratory ell.peri­ments, and the corresponding lifetimes, arc consistent with observed concentrations, the observed flull. out of the atmosphere in rain and the calcuhned Aux into the atmosphere from the ocean surface'·n.... Thus, the lack of alternative sources of sulphur.bearinl lases and the internal consistency of mass­balance calculations suppon the hypothesis that these biologi. cally derived lUes ate the probable main sources of NSS-SO~­in the remote marine boundary layer. \I is therefore reasonable to assume that any change in atmospheric OMS concentration would cause a corresponding change in Nss-soi' concentra­tion and hence in the number of panicles which act as CCN. This could be due to chanles in the number-concentration of panicles. chanles of the mass of water·soluble material in nist­ing panicles, or both .

Efrect on radiative properties or douds Clouds of liquid water droplets form only in the presence of CCN. In the unpolluted marine atmosphere, the concentration of panicles capable of being CCN varies from about 30 to 200 cm-.'. dependinC on aerosol content, supersaturation and meteorological condilions!·'. In Ihis same environment, the total number of submicrometre panicles is often about 200 cm -.1,

such that a significant fraction of them~CCN. Bill!' made observ'at,ions of CCN at Cape Grim, Tasmania, when the total number-concentration of panicles was less than 300 cm -.1.

Results for relative humidities between 100.3 and 101 % (values believed to be-typical of marine douds) showed t~at 40 to 80% of the panicles in that marine setting were active CCN. Thus, changes in NSS-SO~- in marine air arc e~pected to result in clianges in the number.density of cloud dr0r.lets, contrary to the widely held belief supponed by Fletcher' that the atmos­phere always has an overabundance of panicles that coul\:l act. as CCN. Present-day continental air. by contrast. probably is co s· 't etch' . en' if T n

~~ .• _ . ~..:.:.J

.-------.concentration-of-SQ•.:...jn·sea-water-ts-so·larle-{·~9·mmol-kl::'L)--s0Iuble-materials;-such-that-the-marine-N-SS-S0~ ou <r­that it does not limit the emission of OMS.

We can now evaluate the role of sulphur gases in general (and OMS in particular) as sources of NSS-SO;- and of CCN in the unpolluted marine troposphere. Because submicrometre NSS-SO;- panicles are not directly emitted by land or ocean surfaces, they are created in the atmosphere only by chemical reactions. As far as is known, the only significant gaseous sulphur precursors of NSS-SO~- of biololical origin are OMS from the oceans and H2S, OMS and perhaps other sulphVr species from land biota, These gases are oxidized in air, largely by OH"·I6·7S, to form SO~-. At the low NO~ concentrations typical of lhe u~polluted marine troposphere, the oxidation of OMS by NOJ can be considered negliciblel1·lI. In addition to S02 the oxidation of OMS by OH also produces MSA. At low NO.,

qualify. It is obser-ed that sea-salt panicle concentrations at cloud height are typically not more than I cm -J, so that sea-salt itself cannot be the main CCN H .l1 .'I. Water-soluble materials other than sea-salt and NSS-SO;- do ell.ist in marine air but the data are not extensive. Nitrates probably are found on coaner particles, and hence do not contribute substantiallJ' to the number population l

•. Orlanic compounds exist in the sub· micrometer panicles, but their mass-c:>ncentration is probably only a tenth that of NSS-SO;- (W. H. Z!lller, personal com­munication). Their number-concentration and sources arc unknown.

The idea that natural CCN in marine air consist of sulphates has been believed for decades and that they have a widespread gaseous precursor was suggested by Hobbs2'. OMS was iden.,

NAT\JRE VOL. n6 16 AI'lUL 1917 -------REVEWNrnClE----

Relalive numOe'-denslly 01

8 6 4] 2 15 1.0 l I I ':::::: I

oaOI

~:09ss '0~0.02~ ~. 0.8 0006

:> ~......... ·O.oe! 071. "'0010

g 00.12 ~ 0.6 00.14•! "0 0.5 0­2 _ 0.4

• o " ~ 03 ii

•D 0.2 -; >

clOud droplets NINo 06 0.4 OJ 02 0.15

I ,

AlbedO 01 oceln su,llce I o --- 1 I r· 1 • I -- -y---. ----,-. ~---.--. .....--.----+­

05 06 o7 08 09 1.0 12 1.4 16 18 2 0

RelatIve ef 'eclive radius ot droplets , I'" r~ll

ti/ied as Ihe likely source of moSI particles in marine air by Bin ~, al. 5O

, The size dislribution of 5ubmicrometre ,NSS-SO~- as deduced from size-resolved samples)U2 is about righl fo~ activa­tion at supersaturation between 0.1 % and I% which is appropri­ate for marine clouds and is comparable to the values used by Bigg'·. The mass-concentration of NSS-SO;- in the remote marine troposphere is about 0.3 ..g m-) (refs 31, 33) which, if the number-mean radius is 0.07 ..m (which is reasonableT.) yields a total number-population of about 100 cm'- l

, in agree· ment with measured CCN populations. The concentration of NSS-SO~- in remote marine rain waler is around 2-5 x 10-6 M (ref. 34), which ·agrees with a Simple nucleation scavenging calculation with 0.2-0.5 ..g.m-~ of NSS-SO;- aerosol and I g m-J liquid water In the c1oudJs .

Other data also support Ihe Idea thai NSS-SO;- particles

F1&. I Chan,e (.1a) of visible albedo (0.6-..m wavelenatb) al doud lOP caused by chan,in, droplel numbcr-densily N while holdin, wr1ically lnle,raled liquid wiler conlenl (liquid waler .,.Ih, LWP) conllanl. .1G is plolled as a funaion of Ihe albedo of Ille rcference c:Ioud and Ihe effeClive radius ('••. the surfa~­wei,hled mean radius~) of Ihe dropsize dislribulion rclative 10

WI of lhe reference cloud ('~.)' The correspondinl cha~ in top-of-atmosphere albedo (needed for eSlimalin, Ihe effect on Ear1h radialion bud,el) can be oblaieed approdmalely by lIIulli· plyin,lhae values by 0.1. as described in foolnole of Table t. The different albedos for the reference cloud (ver1ical axis) cornspond to diffennt values of LWP. When LWP and Ihe 'shape of the dropsiu distribution are held fiud, the number-dcnsity of doud droplets N is related to ' •• by NINo=('."/~.)-).shown oa the scale at the top of the fi,ure. These calculalions used '~IJ =...m. but the'fi,ure also is approsimately valid for other rcfercnce douds: the plolled ValUfl'of .1G are in error by less Ihan 20% if aJbedo <0.9 and 4.. ,~... Sao ..m (Ihat is, nywhere in Ihe ran,e of ' •• found in real clouds by Hell""). The siu di.tribulions used for lhe calculations are almost monodispcne, broadened just enoulh to averaae over the oscillations in the Mie·scallerin, quantities. However, the calculalions are also valid (or any realilti< siu distributions with the same '••, as Hlnsen and Travis6> showed IhSl the scallenn. propenies o( a cloud Ire controlled essent~lIy

by,••, with very lillIe inftuence from other momenu..ofsize distribu· tion. The.., calculations Issume a direCl solar beam at Ihe Jlobal avera.e zenith anile /10 ~ 600 incidenl on a clO'Ud of sphnical droplets o( pure water, above

,

an ocean surface. The albedo of an ocean surface under a cloud is essentially independent of wavelen.th and aven.es 0.06-0.08 (re(s 67-69); these calculations assumed 0.07 (dashed horizontal line). The computation of phase funClion, sinale.scallerin. albedo and utinClion effiCiency for individual cloud. drOplets used· the Mie prOlram of Wiscombe'o assumin. the re(raCli)'l! indu (or Water ;'·-1.332-1.09+ 10-'j at 0.6 10m wavelen,th". The computation of radiaiive transfer in the c:Ioud used the della-Eddinlton approsimalionll . This leads 10

absolute erron in albedo (for waler-c1ouds al visible wavelenaths and /10 ~ 600 o( 0.00 10 0.03 dependina on cloud oplicallhickncss )

(Fla. 8 of ref. 73). but the error in albedo differences plolled here is much smaller, generally by a faClor of 10.

We now consider the potential effects of NSS-SO~- vari­ations on cloud properties. Changing the size distribution or concentration of the CCN causes the size distribution of cloud

.droplets to change'", which coul6 affect the coalescence and rain production process and possibly the time-averaged cloud cover. However, the effect which is well established (and which we think to be the most significant effect) is that changes in the size distribution of droplets would change the reneaance (albedo) of clouds. This step of the proposed c1imati~ f~back, loop is at present more readily quantified Ihan the other steps, so it is presented in some detail. How the average liquid water contenl of clouds would change is unknown. so we hold it constanl in our analysis.

The liquid water content L(l.m -J), the number-density of droplets N (m-», and Ihe droplel radius, (for a monodisper­

re. ai ,Il -~."" sion) are related by

evaporating in a fashion identical to s 'i nge ,~,~,~"~,~~,:~,,,~;~:;-~= :E~n.,..;,;;,'.~"..,.~ compositions from H2SO. to (NH.hSO., as deduced from lem­peralure-and-humidity-controlled nephelometryT6.7T. The drop­let-nucleating property of particles in marine air is also heal-labile. with CCN disappearing ,at T> 300·C (ref. 36). Finally, Ihe lurnover time of CCN from purely physical data !n the

Tatmosphere has been deduced to be of the order of one dal , which is the same as resuhs from an estimate for turnover time of NSS-SO~-based on mass-concentralion, rainfall concentra­tion and rainfall amounl, as follows. If we take an NSS-SO;­nux, F, of 0.31 m- 2 yr-' as representa!ive-~f remote marine sites~.0.3 ..g m -J as the lypi~1 concentration, C, of NSS-SO~ aerosol, and 3,000 m as its scale heiaht, H, the turnover time, .,. = HC / F - I day.

where p is the densily of water in the appropriate units. Various studies have examined the effect of holding one of Ihese three variables constanl while Ihe olher two chlt(lge:--PaltridgeJ9; Charlock40

, and Somerville and Remer" considered cloud albed9 changes due to a climatic warming which might increase the liquid water content of clouds. They held' fixed so that the increase in cloud albedo was due to increase of N. Bohren" showed thai the increase of albedo would be somewhat less if N was instead held fixed (that is, assuming Ihat CCN conccntra­tion did not change), so that the drople1s increased in size rather than number,

For our hypothesis. we must instead consider the eftea of holding L constant while increasing N, This leads to a decrease

NAT\JRE VOl. '211 I. """-IL I~"!---I£VEWARTICLE

Tallie 1 Climalic ekct allscd by incrnsin. CCN concenlralion over lhe ocean

a Global annllal avera.e clolld cover (ocean arca~ 0111)'1 It bample: effect on surface climale dlle 10 increasin. CCN concenlra· lion N by 30% while holdin. liqllid waler path filed Eanh covefCil

Ocean arca by oceanic For area covefCil by Avcraled Clolld Iypc' conred 1%1 dOllds (%) oceanic slrali(orm over unh's

waler clollds sllrface area Non·overlappcd SI/SC t 25.2 17.6 Non·overlappcd As/ Act 10.8 7.5 Imposed chan.e in N +30% As/ Ac onrlappcd wilh 8.8 6.1 Chin.e in '.* -10%

SI/Sct Chin.e in 0.5-0.7·...m albedo +0.02 Nimbostraills, cumllills, not applicable al TOe.

cllmllionimblls (optically Ihid.; Chanle in 0.5-0.7·...m albedo +0.018 hi.h !Ibedo) al lOA"

Cirrvsl not applicable Chan.e in solar albedo al +0.016 +0.005 (ice) lOA"

TOI~'I cover o( oceanic 4U~ 31.2 Eqllivalenl chan.e in solar -0.7% sluli(orm water c10llds conslanltt (As/ Ac + SI/SC) nOI over­ Chan.e in .Iobal.avera.e -1.3 K I~ppcd wilh cllmllli(orm sllrface lempcralllreU clouds

• Only Ihe areas covered by low SlnillS and slralOCllllllllus (SI/Sc) clouds and middle alloslralus and altocllmulus IAs/ Ad are considered. bec~use the ch~n.e in albedo is smaller (or convective clouds and (or nimboslralus becallse Ihey are Ihicter and may have albedos .realer Ihan 0.8.

f The zon~1 aver~~ ~mOllnl ((ractional areal covera~) of SI/Sc over Ihe oceans varies (rom 18% al low lalilude to SO% al hi.h lalilude·' and anr~.es 34% (or Ihe world ocean. Ilver".e or all (our KUOtlS (n:f. 6\ and addilional dala from S.G.W. " aL. ill preparalion). The daylime (sunlil) ~mOllnl is close 10 Ihe diurnal aver~.e amount. When St/Sc is observed over lhe ocean, As/ Ac is also prcscnl above il aboul 52% of the lime·'. The ~mounl.when·pn:Knl or oceanic As/ Ac is SO% (S.G.W. " ilL. in pnparalion), so SO% )( 52% = 26'1'. of the SI/Sc amounl is overlapped by As/ Ac. Non·overlappcd St/Sc Ihus coven 34% )( (1.0-0.26) z 25.2"1. of lhe ocean an:a or 17.6"1. of Ihe Eanh's surface.

t As! Ac coven 22.4% of the ocelln durin. the daytime (ref. 61 and S.G.W. " aL, in preparalion I bUI only:,,"0.8"1. of Ihe ocean is covered by As/ Ac which docs nOI overlap SI/Sc. cumuills. or cumulonimbus (usin. a procedure parallcllo Ihal used in Ihe SI/Sc analysis abovel.

§ We -ssume Ih'll Ihis Iwo-Iayer cloud is slill optically Ihin enollih for iii albedo 10 be len Ihan 0.8. Ihal is, in Ihe reiion of Fi•. I where ~a is inscnsili\le to G.

nA ch~n8e in albedo o( low or middle water·clouds may be muled al TOA when Ihose clouds are panially hidden by hi.her ice clouds (cirrvs). Here we assume Ihal cirrus is Ihin enoulh Ih~1 the chan.e in planelary albedo due 10 Ihe waler-c1ollds is Ihe same as if cirrus were absent.

e If cumulus is also incillded. 'Ihe lOla I an:al covera.e of clouds whose albedo is sensilive 10 CCN concenlralion chanles from 44.8% 1056.6% . • From Fi•. I. •• Because of lIbsorplion in Ihe 0.6-... m band of ozone above Ihe cloud, Ihe visible-channel albedo al TOA is smaller Ihan al TOC, by a factor

of aboul 0.9 (ref. 47). A further f~ctot of aboul 0.9 is nceded 10 conven visible-channel TOA albedo 10 sol..r TOA albedo"'·". because cloud albedo is lower in Ihe nur-infrared Ihan in Ihe visible. The same facton apply 10 A...

ff The .Iobal avcrale planelary albedo is now 0.30 (ref. 64), so II chana~ in planelary albedo ~a cauleS Ihe same chanle in Ihe amount of solar ener8Y abSOrbed by Ihe Eanh-almo5phere syslem as would a fractional chanae iri solar conSlanl of'( 1.0-0j),h. .

H From Tllble I of ref. 49: .

in mean radius. which causes an increase in lOlaI surface area sola; zenith anile if the r~ults arc npressed as a functi"on of of droplels in lhe cloud and thus an incrCllse of doud albedo. lhe reference albedo ralher than of Ihe cloud thickness. Figure The study of this e!fect was pioneered by Twomey". His Figs I shows that the Cloud albedo is most sensitive to N at Il = 0.5 12.5 and 12.6 show Ihe albedo, G, al cloud lOP at visible but that All is ~sentiaily independent of Q for 0.3 <'Q < 0.8, for wavelengths as a function of cloud thickness for a reference small changes in NINo. Fiaure I was calclliated for eRective plane-parallel cloud with unifonn droplet radius , = 8 jLm droplet radius of the reference cloud r~« = 8 jLm, but it isapproxi­('fairly clean maritime conditions'), as well as Ihe .Ibedo which mately valid' also for any value of r~« in the range 4--500 jLm resulted if N was multiplied or divided by 8, causing r to (sec Fig. I legend). decrease or increase respectively by a factor of 2, (Much of Real clouds arc not homogeneous. and the albedo of a cloud Twomey's subsequent work on this topic eumined the codlpet- with a horizontally inhomolleneous distribulion of droplets is ing e!fecls on cloud albedo due to absorption of sunlillhl by always ,less than thai of II hypothetical homOlleneous cloVd...··). dark aerosol panicles and the increased number of droplets. because·-the albedo-versus-optical-depth function is nonlinear, Only the laller eRect is considered here. because H~SO. and its concave downwards. We therefore compare observed~loudS to 3mmonium salls arc transparent in the solar spectrum.) modelled clouds not by comparing their optical depths t rather However, his calculalions apparently assumed an overhead sun by comparing their albedos. (The changes AQ wi I also be

If!l~i':~~~~rue~h~l.ffi~!l~~n I~~~ .. nd ...1:1 lIe,-u erl in urfa. d>:- -s<!-rnew~at s "a)~r:fC?r ..n:i"h~m2l~neou~·c1ou_d:thJlnfora"plane-­II: e ~ eI!~ • :iU Ii) • a-- 'a'~i ill ~. ~ <, ""'ticti1i .e'ia1&l.' '" . fie'tdill"h'nt:":t- ;;e

Twomey's calculations using the global average zeni! angle c oud elements- are' a jn~the range 0.3-0:8, as can be seen fro;;;" . --. 80 = 60· (which causes the cloud albedo 10 increase relative 10 Fig. I.) , '8;; =0°) and-assumipg"arrocean.urface-a.the-fuwerbourrdary; ----'fhe"1op;of"atmos phere -(-TE>*t-albedo (or' planetary '-albedo) for many diRerent values of N. Figure I shows Ihe change in measured by satellites in Ihe 0.5-0.7 jLm wavelength channel ~Ibedo Aa from that of Twomey's reference cloud, caused by over marine stratus and stralocumut.'us (St/Sc) varies in Ihe variation or NINo in Ihe range 8 10 1/8, where No is Ihe range 0.25-0.65 (ref. 46). Channel albedo al TOA' is usually number-densily of droplets in the reference cloud. Differenl smaller Ihan channel albedo al top-of·c1oud (TOC) by a factor lhicknesses oflhe rderence cloud are represented on the venical of about 0.9 (ref. 47 and S.G.W. and W. J. Wiscombe, personal axis by their IIlbedos. The information in Twomey's Fig. 12.6 is cominunication), because ozone above the cloud absorbs radi­therefore con~ained 'in the right and left edges or Fig. I here, ation at 0.6 jLm, so most marine St/Sc probably have visible which also shows the effect of any smaller changes in N. Our TOC albedos, G, in the middle rellion of Fig. I where Aa is results agree with those ofTwomey where they overlap, showina insensitive 10 Q for small relative changes of N. that the change in albedo due to changing N is not sensitive to Table I illustrates the use of Fia. \. StratiforTfi water clouds

NAllJRI' VOL. J:6 I. APRIL 19K7 ---REVlEWARTlClf---------------­

~ '.nh·,......

005

lK

"ider.d, Ihan 0.8. d." .nd . t ,~nlil)

c tirnc6 :!.

pptd by

.ered by

re ~" is

(cirrus). II. , S6.6%.

a factor I albedo

of solar

tion of Filure

4l =0.5 l.8, for leetive Iproxi· OO ....m

cloud ,lets is d4'·')

I ,

linear, udsto rather Iso be plane. /feient ~. I from' . bedo) lannel ;n lhe sually faClor rsollal

radi­'isible ~a is

louds

Fla. 2 Conceptual dialram of. possible climatic feedback loop. The rtCIanlles art measurable quantities••nd the ovals art prCM:CS­ses linkinathe rectanaln. 11Ie 'ian (+ at' -) in the oval indica•• 'he efleet of a posilive chanae of the quantity in the prtctdina rtCIanale on .hal in lhe lucccedinartClanalc, (ollowina Kello"'I" nota,ion. The mOS\ uncenain link in the loop ilthe deet of cloud albedo on OMS emission; its sian would have to be positive in

order to rcauJa.e the climate.

thaI probably have visible TOC albedos in the range 0.3-0.8 cover about 45% of the ocean (Table (a). As an example (Table I b), we increase N by 30% over the ocean only (because OMS cannot compete with anthropolenic sulphur over land), which causes 'elf to decruse by 10%. From Fig. I, this causes visible­channel albedo at TOC to incruse by 0.02, which causes in increase of 0.016 in planetary albedo averaged over the solar spectrum above these marine clouds, or 0.005 averaled over the entire Eanh. (This increase in whole-Eanh albedo is smaller than the value 0.006 shown in Fig.6b of Twomey tl al"· for this example, because we assume the changes in planetary albedo to apply only to the ocean areas.)

If none of the climatic feedbacks causes cloud albedo to change, the increase in planetary albedo- of 0.005 is equivalent to a decrease of the solar constant by 0.7% in a climate model, which causes a decrease of 1.3 K in global mean surface tem­perature T, when water-vapour and snow-albedo feedbacks (both positive) are accounted for (Table 1 of ref. 49). This reduction in T, caused 'by reducing the eftective radius of cloud

_droplets by 10% everywhere over the world ocean is about one-third as larle as the increase in r. predicted for a doubling of atmospheric COl (ref. I).

Of course we do not know the relationship between a change in aerosol concentration over the ocean and the resulting chanle in cloud-droplet concentration N. Theory and experiments (equations 9-1 and 13-41 of ref. 23) indicate that N is propor­ti~nal }O rCC11l~~, :-..,O.8.-:SClliog.£-=.J Jo~!!!l j~'.

~h~."'Sj:fa • e I~liQ~.me1tta "e. ffcS1if.!1!fO!.liii1~ . "- ' . (Jlsc't,-ssea:.rbove IS mucif'ila'rgen in tlie"<lif€"ct 'radiative e"~ci of non-nucleated aerosol. Using a mass extinction coefficient of

loa. or 101" radiation to. ce

'_10" 01 •••.,........ pirtle".

+

Ocoon

so the increase in planetary albedo from aerosol alone is 9 x 10-' x (1 -0.64) = 3.2 x 10-' for Ihe ocean or 2.3 x 10·' for the whole Eanh; thai is, only about 5% as large as the cloud-mediated increase. ___

The example we chose for illustrati~ 30% change in N) is actually relatively small comparea to observed variations. CCN concentrations over the remote ocean can vary with susan and time of day, and fiom one cloud to another, by an order of magnitude or moreD

; thus even the extreme left and right side~ of Fil. I may be applicable for some models of climatic chan&e.

This analysis of the climatic eftect of changinl the CCN population has assumed that the radiative properties of douds will change only in the solar spectrum, not in the therma" infrared (beyond 4 II-m wavelength). Changinl the size of droplets will nOI significantly aftect the thermal-infrared emissivity of ll'!ost water clouds. because they a.re in eftect optically semi-jnbite and are nearly black bodies al those wavelengthsS!. Cirrus is the only type of cloud which is normally thin enough for its emissivity to be sensitive to optical thickness, bUlthe ice-particle sizes are unlikely to be aftected by variations in the concentration of CCN because ice nuclei are normally much rarer and from difterent origins than CCN.

Global climate and OMS emission

we do nol understand quantitatively the relationship of source strength of OMS to CCN number concentration. It seems likely

LO.-.m!.g::.'J.oLl~SS=Sf)~:--(.r.e( ..5O'),.a.colurnn-mas~UO~g..m:.:,---..that.inc[eased.D.MS./Iuxes-would-increase-the-CGN-population, and a ratio of backscallering to total scaueringof9.2,the averale total backscauering optical depth, .5t>..p, of aerosol panicles is 3 x 10- 3

, smaller than the Rayleigh backscallering optical depth due to air itself which is -0.1. This empirically derived quantity is similar to the estimates of Shaw2

. A 30% increase in the number of panicles (over the oceans only) with n(1 change in their mean size would increase Shop by 9 x 10-'. increasing the planetary al!bedo over dark surfaces by the same amount. but having no effect in areas where clouds are present. The averale cloud cover overthe oceans is 64% (S.G.W. tl aL, in preparation)

based for example on the observation in polluted air that gas-to­particle conversion produces new parlicles'~:

When looking for feedbacks which link the sea-to-air mass flux F of OMS to global climate (Fig. 2), we can consider the variables which make up tne flux equation:

F=A·k·tl.c

We can change either the total ocean surface area (A) available for gas ellchange, the transfer velocity (k l. or the concentration gradient across the air/sea interface (tl.c). Because the ocean is

·.' 1M -..:.N:.:;A:.:l\I..:::.:::It=_E--=V-=OL=-.;):.:I.:.:.~~16::....:.:AI1l::.=I=_L_'I..:;: :.:.7==-----------------------f£VEWARTJCl£ ..

hi,hly oversaturated (by at leut three orders of mqaitude) PoIysipltonla ~ntained u much u IS% of their dry weilht as relative to the atmosphere, the concentration lradient ICTm is the sulphur betaine, OMSP. Until recently it hu been difticult essentially identical to the OMS concentration in surface sea to envisa,e any leophysioloaicallink between OMS production water. in Ihe oceans or on the shoreline and Ihe need of land-bued

A 10% decrease in ice-free--occan'-surface area accompaniecr--ecosystemSlor sufpliur.-whYitiould all allarcommuniiYonti~e---the last Ilacial maximum 51 • Ourin.an ice a.e, the areas covered ocean make the utrava.ant altruistic aeslure ofproducinl OMS by ice on the continents and oceans increase, and there is some for the benefit of, amonl other thinp. elephants and airaffes? uposure of continentai shelf now under water. The deausc in One possible Inswer is that the biosynlhesis of OMS be,an as ocean arel durin. In ice a.e could lead to a dr0l' in the Ilobal I local·..ctivity thlt g."Cw tu become an unconscious benefit for flux of OMS to the atmosphere. Climatic chanllcs would liso the system. aIIect the wind field over the ocean (perhaps oaly sliabUr"1 With Iliac it seems likely that the 'local problem that led to and thereby the transfer velocity II.. The ma.nitude of this effect the biosynthesis of OMSP was salt slress. Ali a result of tidal on CCN would depend on the re:lativeratcs of loss of OMS to movement .Ipe are left ClI.posed Ind dryinl on Ihe belch at the Itmosphere Ind in the water column. lelSt twice daily. Neutral solutcs, such as Ilycerol or dimethyl

Empi~cally, we find t~at the la~est ~Ult of OMS coma from sulpho~~de Ire w~1I known to Iprot~ct ..ainst tbe ad~enc effe~ Ihe troptcal and equatonal oceans. This sUllests thlt the most of freezlft. or drylftl. The mechanism of Ihe protective effect IS

important climatic role of OMS is 10 contribute to c1cvated simply the solvenl effect for salt ofthcsc involatile compounds'9. cloud albedo over the wlrmest ocean re.ions, and thus to reduce Amon. solutes able to protect cells .,ainst'desiccation arc the the inpul of heal into :he low·latitude OttaDS. A coolina of Ihe beUliDes. These compounds IllhoUlh ionic are chu,e-neutral; oceans or I reduction in area of Ihe tropical seas could Ihus Ihe ne,ative and positive moielies Ire on Ihe same molecule. lead 10 I smaller OMS f1Ult, providing a stabilizing MJllive OMSP, (CH)hS··CH2CH~COO- la thionium beUline), is feedback. widely distributed amona marine Ilaae and serves 10 protect

On the other hand, the sCl·to·air /IUlt of OMS Ind con·, them a.ainst dryin. or increased salinily. Thus Ihe requirement sequenlly the libedo of marine clouds could si,nificantly man.e of an osmore.ulalory subsunce by inlenidal or,anisms may IS a resull of ecoloJical chan,es which ·...ould favour phylo- have been Ihe ori"ina! reason for OMSP synlhesis. planklon species wilh ,large' OMS oUll'iulnlles over Ihose wilh We 'bej'in to sc'e"aposSlble jeophysiolojicallink belween the low oulpUI rales. or vice vena. At Ihis lime we hivE only local self inlerest of sall-slress prevention and Ihe: ,Iobalsulphur eumined a few species for Iheir OMS emission rates. AJlhouah cycle. The accidenlal by-product of OMSP production is its we know that there are larle inlerspecific differences, we cannol decomposilion product OMS. This compound or its aerosol yel relale Ihem 10 ph)'loplanklon luonomy. In particular, we oltidllion products will move inland from Ihe shore and deposil are not yet able to use the dala on phyl9planklon speciation sulphur over Ihe land surface downwind of the ocean. The land durin. Ihe leo'iogical past, which· has been oblained by Ihe lends to be depleled of sulphur and Ihe supply of Ihis nutrienl CLIMAP programS), 10 predict the OMS /lux durin, periods of e1emenl from Ihe ocean would increase productivity and Ihe ,llcillion. As I resull. we Ire left wilh an incomplete slory; we rale of wealhenn, and so lead 10 I retum flow of nUlrienls 10 are convinced thallhe emission of OMS inlo Ihe marine almos- Ihe'ocean ecosyslems. What seems a naive allruism is in fact an phere pllys a crucial climatic role. but we cannot yet define unconscious self-interest. Sulphur froen OMS can Iravel fanher precisely Ihe processes which regulale Ihe rate of OMS emission. than the sea-sail aerosol because several steps are involved in

Ihe conversion of .aseous OMS to aerosol sulphale; also Ihe GeopbysiolOl)' and homoeostasis reSulling aerosol panicles are smaller and so have mucl! lon,er Gaia theory} suggests Ihat in order to maintain Ihermosusis on lifelimes. Earth, CO2 is conlinuously and increasingiy pumped from the A large proponion of the current bi05ynthesis of OMSP is in atmosphere. There is a conslant input rrom tectonic processes, the open oceans dislanl from the land surfaces. Is this OMSP and Ihe long-Ierm sink is Ihe burial' of carbonale rock in Ihe also made for the relief of sail stress. or is it a redundam sediments. The sink for CO2 is almost wholly biolo.ically deter- mechanism kept in action because of ,Iacial epochs when Ihe mined; withoul life, CO2 would rise 10 an abundance of well sea or pan of it was sallier? Inlerglacials have occupied only over 1% by volume. Lovelock and WhilfieldH observed thai if one lenth of the lime during the current series of glacialions; climate regulalion docs lake place by pumping of CO2 lhen the Ihis may be 100 short a period for the devolution of OMSP mechanism is now close 10 the Iimil of ils capacilY looperale. biosynthesis. Alternalively, il may be thaI production of OMSP Almospheric CO2 has been reduced from aboUI 30% of the in the open ocean has a differenl .cophysiolOlical basis from almosphere al Ihe start of 'life 10 the presenl 300 p.p.m.v.• a that in the continental shelf regions and one Ihal is unconnected factor of 1,000. It was suuesled Ihallhe decrease of CO2 through with salinity as such. ils declining greenhouse offect had compensated for Ihe We have seen how the local self·inleresl of shoreline algae monOlonic ,increase of solar luminosity and so the c1imlte had could lead to Ihe mutual sharing of sulphur and nutrients wilh remained consUlnt and suiled for life. It cannol be much more land-based organisms. The evolUlion of a link belween ocean

ccd withoul seri0!isly i pa!{u, he,;Br !I1!I~~IH~,~.t';'~c0,~~d roduction could ~~~.,hap.ee?~~)~ it;!! if.~,~. --~~ : , . '. ~SJ 0;111:1 {\ nlHlltru .,.~'"

"'.il~£~inintmu, .p:p·.m;v~; of-Ole:: ~ahationM. ,: some ulent vu neral:i e'to solar u ra-violet: Cloud formation OMS emission, Ihrough its elltct on ;~e planetary albedo, with rainfall would return nitrogen to the ocean and also serve

shares wilh CO2"pumping a coolin, tenaency. So if the OMS as a sunshade. If either or bOlh of Ihese effecu were significant production increases with lemperature and/ or solar irradiance, for the health of phyloplankton then species that emitted OMS Ihe sign or its c1imltic effect would be in the right direction 10 might be at an advantage. offset what seems, for the biota, to be an ucessive solar flult, Is the sulphur cycle also involved in global climate control?

How could the local activity of species living in the ocean, Evapotranspiralion is known to modify Ihe climate of foresls evolve 10 serve in the altruism of planetary regulalion? The . in Ihe humid Iropics. The additional cloud cover that comes biogeochemical cycles of carbon and of sulphur are intimately from the vast water vapour flult of Ihe trees increases Ihe linked and appear to be connected wilh the regulalion of redolt planetary albedo and funher cools the surface. The emission of potenlial in bolh oxic and anoltic ecosYSlems~6.j? The firsl OMS from the oceans seems 10 let similarly. Through its aerosol indication of a .eophy.ioI08~cal role fo'r sulphur came from Ihe oltidation products it alters the properties of clouds, increasing observations of Challenger>- thai marine allae emilled OMS. Ihe albedo of the oceanic re.ions and hence of the greater pan He found Ihat some species of shoreline al,ae of Ihe order of the planet. The link betw~en the biola and c'limate in both

66\

of these processes of cloud formation could be a mechanism abundance. We also need to undel'5tand the relationship between for climate control, the clouds servin. as do white daisies in the OMS concentration in the air and the CCN population, 'Oaisyworld" model of Gaian climate regulatio/l60

• through the intervening aerosol physical processes. Knowing Lastly, OMS is the principal component of the present bio· how the area of cloud cover is inlluenced by CCN is also

---leoehemieal..ftull-of-sulphuMo-the-atmosphere:-8utiti:nmrlhr-lmportant. -.-.----~~.~----'. '.,­only one; some sulphur is emitted as H~S, COS and CS~. COS, .Nonetheless, the role of the CCN population in controlling both from direct emission and as an oxidlltion product of CS~, albedo, the production of CCN in marine air by the oxidatiJn is sta!>le in the troposphere long enough to be an important of OMS from the biota and the sensitivity of the Earth's temper· source of sulphur to the strlltosphere. COS in the stratosphere ture to the CC N pojWlation seem to be ·established. Although would be oxidized and produce a sulphate aerosol there. Such we do not understand the details of the climatic feedback, it stratospheric aerosols scatter sunlight back to space and lead seems that CCN from biogenic OMS currently act to cool the to II cooler climate. The biological variation of COS output is Earth. It is possible that the Earth's climate has been mediated therefore another possible .eophysiological means of regulating in the past (for instance, that this feedback has helped to c1imllte. counteract the increasing luminosity of the Sun and/or that it

has already counteracted the inftuence of the ~ecent increase in Future rese.-reh needs CO~ a'ld other 'greenhouse' gases). However, the data required We propose that sulphate aerosols derived froln the sulphur to demonstrate the latter ellect have not been and are not now gases produced by the marine biotll an: important determinants being acquired. of cloud albedo and, liS a consequence, the climate. It also seems The portion of this work done in the USA was supported in likely that the rate of OMS emission from the oceans is affected part by NSF grants ATM·82·IS337, ATM·83·18028, OCE·83· by the climate, thus c1osin'lI feedback loop. 15733 and ATM·84-07137. The computations were done at the

There are significant gaps in our knowledge of this proposed National Center for. Atmospheric Research. We thank Marcia feedback system. Most importantly, we need to understand the Baker, Keith Bigg, Robert Chatfield, Ian Galbally, Dean Hegg, climatic factors affecting OMS·emission. Because some species Ann Henderson.Sellers, Kendal McGuffie, Henning.Rodhe and produce much more OMS than others, we must include the Starley Thompson for commenting on an early draft, and Antony necessary understanding of controls on 'PhytoplanKton species- Clarke ror discussions.

I. J)id:in~IH', M. I:. Ii; Ci ..:t'wnc. R. J. I\'.,.,rr _,It. IIN- 115 Ilo.n:hl.� ,\7. W"llllCt. J. M.1t. H\Jbft~. P. v. A'".IJJJtItc'nc Snnt<-t. 'I" 1",rud"C"M\' S.."~." 21~ (Academic, •. Sh...... G. n,"~,"- (1~,u.~f' s.. ~1_.1U.l II~M.\I. Ne ...· Y\>r~. ''''771.� ,l. Lmdock. J. E. Hld/.."'m. rtwr. S.....·. • 7, .,~~~.\q7 c IYKhI .'K. Huhb:o.. P. V.. U.rri"qn. H. A. lloI'IinM)n. E. Sor"n' Ill. 'iOQ-915 [IQ741.� .t. AndrcolC'. ~I. O. in Th,' ~inlfi1'(.rll.·"ur,,' ('n~"••of .'iftl,,,r IJptd N",...,.,. '" Iltf' 1I"",,,,c' .\"1. P'.. ltnd,c, (j. W. Q. 11 R. "''',. 5ar ·1.... M'Q5·1N9 (19~) .�

"'".u~..,,· led, (i.lIm'l"lIY. J. N.. Chul~<1. M.. L. And,.n~. M. 0.1;: R,odh~. 1t.1 ~-~~ oto. ("h.:lrlock. T. P. TrllKA ).4. !4~-~~ {J9.2J.� I Reid~t. UordreC'hl. 1~1t.' I. otl. Somtrvillr. R. C J.• Remer. L. A. J.ll'~.I"J...u ",%6I-967~ (198-41.�

~. Ant..lrt"..~. ~f. 0, in nc' R",/f' ,,! "'t..y.#:'U"'WU'll't ill (;..o."'·ffut"ul (:I'..:i"t~ led. AU::lI·MrnArd. 4.=!. IIohrrn. C'. f. J. ll'Ntf'A.'·J. Rn.. _. ~K67 (19I~I.

P.I .\.\1-.\4\2 tRotidC'l. Dordr«ht. 19K61.� 4.\. T.omC'y, S. A,,,.u~nr Arm.'OJJ IEI~'i~r, AIm.Icr4am. 19771. h. {·ulli.... (· f.l tii,...."..ttler. M. M. AI"""" 1:",,,,,"_ ..... t~.'·1~71111"'~11 ...... Welctl. fl. M. in "'jJ"" C"rrft!r'l'Jtt'tlH\l AI".,I,Jtntc It.dl41to11 SO,.-S071 AtnC'rican Mtlcorololi· 7. Andr~ ..e. ~f" O. in .'I1I..,,';"al ~'Q'-': S""*",,J.,r, lnh (ohen. roo (,nlenhtll/, M.. "". &:. QI' Socicl)'. RoIlon. 198.\1.

Ibl\"t~N'ln" H. 0.1 ot,,~.c~ lli":,,. Nc.. Yntk. 1'IX"1. 4~. Gliftriel. P.• lo...~;oy, S.• Au~in. G.• Sc'twn.ztr, D. in Sulll Co"i~""a Oft Nm().JfM~ric

It. Andn:,.e" ~t. O. " Andr~:.e. T. W. J. ."'''''.n. H.. ,- huhmlu~uJ. RQJUUHtft ~.\O-H~ tAmeriuI\ Mftcor~oJic.1 Socie,." lotion. 19"1. ~,V,.ir .. \;amunh)·, A.• AndrC':.lr. M. n.• h·cn.on. R. L. L,,,.,,,oL O'·,.QII',.r.... ~~·7U 11~1t~1 .&6. ('a..k1e)'. J. A. Jr ... r>a~·ic,. R. J.• ''''O.l, Sn. .0. IO.2!~ 10J,. '1916).

,u. TumC'r. S. ~t. It. lj,~. P. S. J QlIfto:\., ('JNoM. L ~~J. ~.\~ rI"'.~ I.� 47. Wi,;C'OmDc. W. J.• Weldl. Ro. M. A HIli. w. D. J. .'lfIClJ.. Sci 41. IJJ6-IJH (IUotl. II. K;,.cr,. T. S.. (·linc:. J. n.. Lqmmnn. Ro. fl .• Klrily·HutoC'n. S. M. 1. ~"yJt, ....... Rf". lin IhC' Pfe~l. 4X. T.fMn~)'. S. A. Pi~Pln .. ,. M.• Wolfe. T. L r.... M&. lS6-l66 11",,1.� 1:. And,.u~" ~t. O. L,,,.,,,,I. n.·f'w,.''J,.... '~.-':11l1IW~1 ot9. Wtther.ld, R. T.• ManaM. S. J. a'mllL Sri. n. 2044-:!OS9 119"'l.� I.l. Km"hlcoC'Om~. p. It. Shnocrr. I). Moll'. (.Jtnn. I'..\.ot.\ •.\~.\ tlWltt. ~O. W• .,oner. A. P. COl aL NQr,," )61. J:?O.I21 11976).� lot. ,'"dru~·. M. 0 .. H"rn"rd. W. R. It" Amnliolb.J. M. Fmc.. 11•.11. t ~u<:khnlm))5. 14"17-17711"8.\1. ~ I. Plltridle. G. w. It. PI.II. (. M.... R.Jialit-r P'rvrtJU.' ill M"rtJI'OInc..'· 'I"" ("/i".afoIOJJ' 191�I". ~"il.i. Ie M:tkC'r. I'. Do, S:.\",c. (', M. 4: 8f1'itC'nNch. L 1'.1111. J...ltf'",- J;lmol. I', M7·"'~ 4,".b~\·iC'f. Amwcrd~"'. 1976 I.�

11"Itt.1 1� ~~. M'MurT)·. P. H .... Wils\>n. J. C. J.•f'UJIIt.r.l, Rr1.. "101.~I(),rlI9'UI. 16. "·in. 1-." (jrulljt"OIn. I) It. ~in(eld. J. H. J. fr'1piI1'1. R", ljn 'he rre:>o).I. ~.\. MclnIYf~. A. r' QL Sn',.,.0' 191. 1131-11 ..... 11'761.� I'; I-rrd... M.. J.. (·h:tlliC'ld. R. H. 4 Andrnc. M. O. N.,.."})I. .'iJot_.'ilh I r"PW'lI. ~. Rryiln. K.• Srw1miln. M. J. J. ..,-opItyJ. It,~ ... 11"'·1161" 1198h� Ill: AnJrr;a\,. M. 0.'" lit J......,..41·(. H"," J~.~l-I~WlI)'*tt~). ~.~. Lo\·clod. J. E.• Wltilfidd. M. NQI,,'r 1t6. S61-S6) fl'lII2l.� 1"1. T.\ndotll. G. S .. lturfllWO•. J. P,. SchnC'idC'r. 'c". It. Moon~:"l. G. K. ("1t"Ift. Fftn. Lrn UI; ~ G~rr~I50. R. M. It. lrnnan. A. Pr-r. 114'1... AC'Gl Sri. U.S.A 'TI. 46!2·.t656 (19811.�

,u•.\.oWof, ,J~MtJl. 57. Holl .. nd. H. I>. Th" (·lIr".jnll F~'OI""';lIIt U/IJw Al~'''''QItJOccoaru (PoncC'lon UnivC'(1it)' ~II SOiIt/man.1.. S.• S:.mlc'.Il.1... Pr~~r\l. J. M.Alib. R.G. J. QU"/I\, ('It",.. 4. ~~1-~40( 1~1. Prt..~. Wlt4!. :1.' AndrC':"c. M. 0.. h~rrc,hcim. If., Andrr:JC'. T. W.. Knll_ MI. A.. R"trll. T. S.1t. MC'rrill. J T. ~H. ("h"lIenlC'r, f. Adt·. E":.I-mrH... U. 42'9-ot91 t I'Sfj.

J. £1',..0\. ("~·ftl. ri" IhC' prC':>oll/. ~9. Lonlock. J. ~. HUlC'lIifft" hiflf'h.I'1. At"la II, ~.lI ••W I "nl. Itar.c~. (i. Mo. Ii. l:In,. R. t:. Grrlpftn. Hf'''' Lrll. I.'..... -.~I 119Ktl1. Ml. W.t,Oft. A. J.• lov~lock, J. E. Tt"I/v.J JSa. 2I-4~2" 119&JI.

:'.\. Prul"f':ochC'r. H. )t. &:. ~'rH. J. I>. •\Iien",...'·....·, u(CIIHHb ud f1"r.rc",i,,,,imt' Rrrdc'l, I>ordrrchl. 61. W"rrC'n. S. G .• Hahn. c.. London. J. in Fill' C_fr""f'C'." Ar".cuJtltc',;C' RQ"j"tilJlI .\ IJ·J 14 1"'7MI. IAmC'flC'Un MtierolOilcal Soci.n'l. RcKlon. 191'\1.

~01 N•.,. L K. "\1.1"'. Ht"l. le. 1t;:·1Ul (1~M(lt. nJ.. WilffCII, S. Go, H~hn. ( .• Lon40n. J. J. (7~",- ."J.. Mf'tC"Iw. u.. 6"_667 (l91h ,,, HC'I("h~r.~. H. Th.. PfI.nin ,,(It.,ltrln_"1 If!Ilt'am'''rid,C' lJni\'C'nIlY PrO-'. 1%~1. IIJ. Shlnt" K. P.• HrndC'nGn·xllC'N, A. " StinlG. A. O. J' It.. ".,t. Sot'. 11•• 1170·117'9 (19841.

Kt~hlC'r. H. Tralt\. ,,'QNI4ltn s,1C'. JL II S~ (1')61. M. Strrhcn.... G .. L.• (".m~II.G. G.• ~o~C'r.Hur.T. ijj. J"oTMY:c. Rco.l,16. 97J9-"60( 1981).� Hot'ln,. P. V Q. JI H. "'C:1. ',Su,c',"', :'6-..\.!71 (111711. b.~. H.n...C'n. J. t:, • T,..VI). L D. S{r4H'r St"L R,,~. 16., S:d:'&IO 1197otl.� M.dh. L F. Ihc'\l\. Uni\·. W;r.,hln,IOft, xotUle i I_KI M. H~." D. .-\. J. al"'OL Sri '';1. .1cr.l-4UO (I'KIIl.� (,h:.rl-.(ln. M.. J.. (,h.mc,d~,. W. It. KI'C')'" D. in 1M ~"Irlu-"'t<'MI ('In.". nf Sulr..r o"d 67. Rricllrh. M. 6: R:amanalh.;an. V. 1. afIPI- Mn 11. 11~1171 119t121.�

;� '''',,,.. "'rtkl'{tIlf'rr t GallowOI\. J. N., Ch,bon. R. J .• AJldrC':ar, M. hIt Koodr.U)'tv, K. RQJilJtH", I" ,"" A"n,upJI..'Y (Ac:adtmic, New York. 1%9).'i ' . ~..~ *·,r-r.:.?~d. .. h\I. P'oIynC', fl.. E. J O/".,,'\, Sci. 1", 9SV·970 (19711 .

.. _'Q,._~..1I"-'l.QIl'~~_J·.6pp1..0p1_)J..I)(ll..J.lll9 __._·_.. ._. __• -_. --­..U91lO1.l'\.I'7.IUI1~-!UCK~ {=I.r:;:::..lI.~"='z.-'=..u.:IQ:.u.l- -'­

71. H;alr:, G. M.... QuC'rry, M. R. A",t Opl" IL "'~·56J \ IY7J I.

S.I\on·. D. L, I:. I'rll'rcfu. J M, G.·,"';'n.ll.r1. Lru'. 1~1I7.t:IO ll"'tl.:l. 7:. Jo:oo<ph, J. H .• Wi:o.comM. W. J" , WC'innu,n. J. A. J. Q'lfttlj. ScL n.. ~ot!i2·~4~q [l97bl. Pro,,,,,,,ro. J. M .. S"'\\lir. 1>. l.. Nrc" fl.. T.• I>U\'.... fl.. A. I; Merrill. J. J. ~C·'l(lJII". R,.\, ... 7.\. Kin ... M. D.• HlIuh\-.u·dh"n J. QI".U.l. ScL O. 7t14.801 (I9~6).

tu~XA-IU~W. 11"'''~1. 701. ~C'110". W. W. in Ctimufr uj'IJw A,..,U" lcds WCOnCOf. G .• 8o.1il\1. S. A.I 111·116 .\... G:llIo.... ~·. J. N. in nu: Hi,,~.·JJdu:'"II'.1 ().... I,•• "l StlIJ..,. QItJ Nl,,,-.r" in Iltt" H,.m,," I(jC'nphpiClil In.'liIUh,''. Fl.Iirhankt.. AIII"h, 197""�

''',,"u~nrrf'lcd, (.j~lIo ..'... ), J. N .• '("h;,rbun, R. J.• AA4rc;,c, M. 0.' Rodhc. H.J l.a,\~ 17" H. Gr.riJrl. T. L G<""f'h.n. RI·.'. Lf'1I. 6., .\:'9·).\1 i 197'1.� IMC'idC'l. Oa,dru:hc. 19ltSI. 711. n."-:r. A. I).• Ahh4ui';l. N. t'. • Conn. O. S. 1. ~ropJt.fJ. Rl."l. tJ (in litc f'ro)l.�

,,\~, Ch.. rl:-.on. It. J " Rodht. It. NlJlu,c' 195. hJC.l-6K~ t 1"I!.~1. 77. ("O\'C'n, n. S. 1. 1f11Jf'lIp. Rr.'. (Iuhmilledl.� .\tl T,,·umt.\. S. A. J u'n"n. S'''I.1K. .\77 (1"'711. 7lt. Ndld. A.. f'1 ~/. NUlMrt 195. ~~0-.:'1.\ (J"'l<11.�