Table 1. Worldwide (Ww) and New YorkCity (NYC) depositions of 9"Sr. The averageratio was 0.15. The 1966 figures are estimates.

ww NYC ~~RatioYear (wMO m

NYC (Ww:(Mc (ic/kn') NYC)

1954 0.35 2.76 0.131955 .60 3.57 .171956 .55 4.43 .121957 .50 4.44 .111958 .90 6.16 .151959 1.13 8.68 .131960 0.38 1.58 .241961 .46 2.43 .191962 1.53 11.07 .141963 2.59 23.87 .111964 1.84 15.85 .121965 1.00 5.53 .181966 0.4 2.3 .17

ing the highest frequency. Plus valuesoccur at least once during all monthsbut September, October, and Novem-ber; in every year, March, April, andMay were all plus. These observationsindicate that in New York City thepeak is indeed a spring phenomenongenerally centered on and most intensein the month of April. Conversely, thefallout is lowest in the autumn, withSeptember having the highest frequencyof minimum values.

Predictions of 90Sr deposition inNew York City are important, partic-ularly in estimation of future human-body burdens (12). The method ofpredicting 90Sr fallout there is based onpredictions of worldwide depositionand the observed relation betweenworldwide and New York City fallout(5). Since the beginning of the cur-rent ban, the stratospheric inventory of90Sr has been decreasing, with an ap-parent half-residence time of about 10months (2). Extrapolation of the mea-sured stratospheric burdens yields the

Table 2. Stratospheric half-residence time(HRT) derived from fallout data for New YorkCity. Figures for 1966 are partially estimated.

Fallout HRTPeriod (nmc/kM2) (months)

First moratoriumDec. 58-Nov. 59 9.16Dec. 59-Nov. 60 1.74 JMay 59-Apr. 60 3.93 9May 60-Apr. 61 1.56 5

Current test banJan. 63-Dec. 63 23.87 20Jan. 64-Dec. 64 15.85 8Jan. 65-Dec. 65 5.53 10Jan. 66-Dec. 66 2.3 1

June 63-May 64 22.221 9June 64-May 65 9.021 9June 65-May 66 3.71

16 JUNE 1967

inventory decrease for any year, whichequals the expected global fallout forthat year, less the change in tropo-spheric inventory. T'hen, on the assump-tion that the ratio of worldwide to NewYork fallout is constant (Table 1), thedeposition value for New York Citycan be calculated. This constant for90Sr is 6.6 mc/km2 in New York forevery megacurie of worldwide fallout.By this method yearly predictions of90Sr fallout from 1966 through 1970are respectively 2.3, 1.1, 0.48, 0.21,and 0.09 mc/km2.

During the course of the year (7)alternate predictions of yearly andmonthly fallouts are computed; the his-torical record of 90Sr fallout in NewYork City is employed, by use of themonthly values relative to the yearlydeposition. For example, the cumulativedeposition to the end of June 1966was 1.71 mc/km2. On the average,67.6 percent of the annual depositionoccurs by July 1; thus 2.5 mc/km2 ispredicted for the year. This is in goodagreement with the original predictionbased upon stratospheric depletion.

In summary: By the middle of 1966the peak of accumulated 90Sr onEarth's surface near New York passed;the deposit is now decaying faster thannew material is falling. The strato-spheric half-residence time of 8 to 10months, based on the rate of 90Sr dep-osition in New York, is in very goodagreement with estimates from directmeasurements.The overpowering variations in the

absolute deposition levels were virtuallyeliminated by a method of normalizingthe monthly data on fallout on a con-tinuing basis; thus seasonal oscillationin fallout is easily observable. Thisspring peak is extremely symmetricaland constant relative to the rest ofthe year; it is independent of the tim-ing and magnitude of atmospheric tests.

Predictions of fallout in New York,based on empirical observations and onthe assumption of no substantial addi-tions of 90Sr to the atmosphere, indi-cate that less than 0.1 mc/km2 willbe deposited during 1970-less than 1percent of deposition during the peakyear, 1963.

H. L. VOLCHOKHealth and Safety Laboratory, U.S.Atomic Energy Commission, New York

References and Notes

1. One of 160 stations in the AEC network.2. Calculated from direct measurements: H. W.

Feely and P. E. Biscaye, 41st Monthly Re-port on Project Stardust (U.S. Dept. of De-fense Atomic Support Agency, 1965).

3. G. A. Welford and J. H. Harley, Health and

Safety Lab. Rept. HASL-42 (1958); Science131, 3416 (1960).

4. H. L. Volchok, Health and Safety Lab. Rept.HASL-174 (1966).

5. , Science 145, 3639 (1964).6. L. D. Y. Ong and F. Y. Durkan, Health

and Safety Lab. Rept. HASL-158 (1965).7. H. L. Volchok, Health and Safety Lab. Rept.

HASL-155 (1965); Rept. HASL-172 (1966).8. L. T. Alexander et al., AEC Report. TID-

6567 (1961); E. P. Hardy et al., AEC Re-port TID-17090 (1962); E. P. Hardy andJ. Rivera, Health and Safety Lab. Rept.HASL-149 (1964).

9. E. P. Hardy and J. Rivera, in Health andSafety Lab. Rept. HASL-173 (1966).

10. M. W. Meyer, V. T. Velassis, J. S. Allen,Health and Safety Lab. Rept. HASL-171(1966).

11. N. G. Stewart, R. N. Crooks, E. M. R.Fisher, Atomic Energy Research Establish-ment Report HPR 2017 (1956); L. B. Lock-hart, Jr., A. W. Paterson, Jr., A. W. Saun-ders, Jr., R. W. Black, Naval Res. Lab. Rept.5528 (1960); P. W. Krey and W. C. Rosa,Jr., HASL Surface Air Sampling ProgramNov. 1964 (1965).

12. J. Rivera, in Health and Safety Lab. Rept.HASL-173 (1966).

27 January 1967; revised 12 April 1967

Jupiter's Atmosphere:Its Structure and Composition

Abstract. Recent laboratory and ob-servational data support the hypothesisthat the composition of Jupiter's at-mosphere is consistent with the rela-tive abundances of the elements foundin the sun. A model based on thisassumption provides a reasonable inter-pretation of abundances of hydrogenand other gases obtained from studiesof various regions of the planet's spec-trum. Two presently unidentified ab-sorptions may be caused by organicmolecules in the Jovian atmosphere.

The composition of Jupiter's atmo-sphere remains an enigma, despite thefact that direct measurement of thehydrogen abundance has been possi-ble for several years (1), while twominor constituents, methane and am-monia, were identified long ago (2).Determinations of the amount of hy-drogen, made by various workers fromobservations of the quadrupole absorp-tions in the infrared, differ widelyamong themselves. The best new valuesalso appear to conflict with independ-ent H2 abundance determinationsfrom analysis of the ultraviolet reflec-tivity of the planet. Deductions con-cerning other possible constituents arestrongly dependent on the resolutionof these dilemmas; as yet no addition-al gases have been discovered (3). Thisreport discusses some preliminary re-sults of our efforts at reevaluatingthese problems. Our study is based ona structural model for the Jovian at-mosphere that appears to be consistent

1489

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from

2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000

WAVELENGTH (ANGSTROMS)

Fig. 1. Ultraviolet geometric albedo for several H2 abundance-surface reflectivitycombinations matching the observed Jovian albedo at 3000 A.

with available observations. To pro-vide a model for the chemical compo-sition, we have started with the hy-pothesis that the elements are presentin solar abundances, as suggested bythe following argument.The recent work of Rank, Fink, and

Wiggins (4) offers the best currentlyavailable values for the H2 abundanceas derived from the quadrupole ab-sorptions; an average value of 240 km-atm is obtained from their determina-tions. Correcting this by a factor of 2for the effective air mass, we obtain

Table 1. Chemical equilibrium concentrationsat 2980K for a model Jovian atmosphere basedon solar abundances.

Assumed Calculated

Reactant* initial equilibriumReactant* abundance abundances(m-atm) (m-atm)

Ha 120.0 x 103 120.0 X 103He 23.5 X 103 23.5 X 103Ne 28.5 28.5Ct 126.0 002 109.5 0N2 11.5 0Sit 7.6 0Mgt 6.0 0St 4.8 0Fet 0.89 0CH4 0 126.0NH8 0 23.0SiH4 0 0SiO2t 0 7.6MgOt 0 6.0H20 0 197.8H2,S 0 3.91FeSt 0 0.89

*Numerous other compounds were consideredin this study but were not formed in significantamounts. tlndicates solid at 298'K.

1490

a value of 120 km-atm for the JovianH2 (5). Although there is still a largeamount of uncertainty (± 50 percent)in this number, owing to the scatterin the equivalent width data, it is thebest value currently available. The ob-served Jovian methane abundance is150 m-atm (6). If one assumes thatall of the carbon in Jupiter's atmos-phere is present as methane, an at-mospheric C/H ratio of 0.63 x 10-3is obtained. The C/H ratio in the solaratmosphere (0.52 X 10-3) is in goodagreement with this value, consider-ing the uncertainties involved. Onthe strength of this agreement, wehave assumed that all the elements arepresent in solar abundances, but in theform of fully reduced compounds. Thisassumption gains support from its suc-cessful adoption in recent work on theplanet's interior (7). Furthermore, heli-um must be included in relative solarconcentration in order to increase themean molecular weight to 2.3, a valuewithin the range determined from theobservation of a stellar occultation (8).

Considering only elements with theabundance of iron or greater, one ob-tains the composition given in the sec-ond column of Table 1 (9). In orderfor this compositional model to repre-sent a physically meaningful situation,the mixture must be chemically stable.We have determined the conditionsof chemical equilibrium (10) for anappropriate group of reactants, usingthe solar compositional model for theinitial conditions. The results showedthat the second column in Table 1does not in fact represent an equili-

brium mixture; SiH4 is unstable in thepresence of water, becoming SiO2 de-spite the large excess of hydrogen. Theabundance of H2S is reduced because,at these temperatures, the iron com-bines with sulfur to form FeS [a re-sult pointed out by Urey several yearsago (11)]. Since the solar abundanceof sulfur is considerably higher thanthat of iron, however, a large amountof H2S remains. The revised atmos-pheric composition is presented in thethird column of Table 1, and formsa basis for comparison with the re-sults of observations. The ultravioletregion of the spectrum will be con-sidered first.The ultraviolet geometric albedo for

Jupiter was calculated by solving theequation of radiative transfer for asimplified Jovian model atmosphere.The model is comprised of one scatter-ing gas (which does not absorb) dis-tributed in hydrostatic equilibriumabove an opaque Lambert surface (ref-erence level) of reflectivity A(A), andof an arbitrary number of gaseous ab-sorbers (nonscattering) which are eith-er mixed or in diffusive equilibriumwith the scatterer. Since H2 is morethan five times as abundant as He (solarcomposition), and scatters ten timesmore effectively, the one scatterer ap-proximation should be valid to withina few percent. Isothermalcy is also as-sumed. With these assumptions, thegeometric albedo P(X) can be expressed(at opposition) as

p = 7Ts J'E3 [T8 U ± 2 T2 UJNa du

+ 2A E4 [LIwhere

in which

L (X) = [Tr + Y TaJ

T. X)2af °) n.X47r

is the scatterer optical thickness andra(A) = 2kaXa' is the absorber opticalthickness; ou,(X) is the Rayleigh scatter-ing cross section per molecule, D0 isthe Rayleigh phase function for back-scattering, nO is Loschmidt's number,ka(X) is the gaseous absorption coeffi-cient, X.0 and Xa0 are the scattererand absorber abundances, respectively,above the reference level, and E3[ ]and E4[ ] are exponential integrals. Inthe above definitions, A indicates wave-length-dependent quantities. The sum-mation is taken over the individualabsorbing constituents; Ma = 1 ifthe absorber is mixed with the scat-

SCIENCE, VOL. 156

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from

terer, while MAO! = in_( /ii if it is dis-tributed in diffusive equilibrium. Heremna is the absorber molecular weightand m- is the mean molecular weightof the atmosphere.

If all absorbers are mixed with thescatterer, the expression for the albedoreduces to

P:=:[1I-L-+ (2A - L EE[L])The observational data used in this

study consist of spectra (2000 to 3000A) of Jupiter obtained from rocketsby Stecher (12) and subsequently con-firmed (above 2300 A) by Evans (13).Stecher analyzed his observations interms of simple Rayleigh scattering bymolecular hydrogen (12), and obtaineda best fit with the data for an H.,abundance of 10.2 km-atm and zerosurface reflectivity.

In order to confirm Stecher's re-sults, we recalculated the albedo, us-ing the formulation given above andrequiring exact agreement with the ob-servations at 3000 A where no absorp-tion is expected. Under these condi-tions, we obtain a best fit with thedata for 12 km-atm of H., and 7 per-cent surface reflectivity. This fit is madedirectly to Stecher's data points andessentially confirms his conclusions. Re-sults for abundance-reflectivity combi-nations giving exact agreement at 3000A are illustrated in Fig. 1. The maxi-mum H., abundance allowed in theabsence of absorption at 3000 A is 16km-atm, as determined for zero reflec-tivity, while a lower abundance limitof about 6 km-atm exists within thereported error limits.The "best fit" curve in Fig. I (heavy

line) suggests the presence of threeabsorption features previously pointedout and discussed by Stecher. Theseare the dip from 2400 to 2800 A cen-tered near 2600 A, the gradual drop-off below 2400 A, and the absence ofmeasurable reflectivity below about2100 A (Stechers' instrument couldhave recorded radiation below thiswavelength if the intensity were highenough). Our formulation for the al-bedo, in addition to permitting con-firmation of Stecher's results, now putsus in a position to identify and deter-mine abundances of the atmosphericconstituents responsible for these ab:sorptions.

Of the atmospheric constituents ex-pected to be present on the basis ofsolar abundance considerations (seeTable 1), only NH3 and H.,S are im-16 JUNE 1967

0

LUJC-)

m

LLJ

oLUJ

(5D

2000 2200 2400 2600 2800

WAVELENGTH (ANGSTROMS)

Fig. 2. Ultraviolet geometric albedo for 12-km-atm of H., 7 percenttivity. and various amounts of NH2 (in cm-atm) mixed with the H2.

portant absorbers in the region 2000to 3000 A. Figure 2 shows the resultsof albedo calculations (14) for a rangeof abLindances of NH. mixed with theatmosphere above the reference level(reflectivity = 7 percent) defined by 12km-atm of H.. An abundance of 230cm-atm of NH3 would be expected,assuming solar composition and nochange in composition with height. Thisresult is completely incompatible withthe observations. Although none of thecalculated curves fit the data well inthe region below 2400 A, an upperlimit to the NH3 abundance of only0.007 cm-atm can be set by assumingthat NH3 is the only absorber and byaccepting Stecher's data points without

0.60 0.0

0

XX

L !. 0.2LU

2000 2200 2400 2600

WAVELENGTH (ANGSTROMS)

Fig. 3. Ultraviolet geometric albedo for 12 km-atm oftivity, and various amounts of H2S (in cm-atm) mixed

3000

surface reflec-

the associated error. Within the error

limits, abundances up to 0.03 cm-atmof NH3 are permitted. It is worthpointing out here that in Stecher's (15)original presentation of these observa-tions, he gave a much lower valuefor the albedo at 2100 A. The earlierdata show much better agreement withthe calculated albedo curves for am-monia absorption and suggest an abun-dance of -0.05 cm-atm of NH3 mixedwith the atmosphere.The results of a similar treatment

for H.,S are shown in Fig. 3. Thesolar abundance prediction, as in thecase for NH.3, is more than three or-ders of magnitude above the upperlimit determined from the calculations.

3000

H2, 7 percentwith the H2.

surface reflec-

1491

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from

The upper limit to- the H2S abun-dance is 0.003 cm-atm for the bestfit with the data points; it can be in-creased to 0.02 cm-atm within the er-ror of the observations. While the H2Sprofiles show a better fit to Stecher'scurrent data (12) than the NH3 pro-files, the opposite is true if the originaldata (15) are used. The quality of thedata prevents a better definition of therelative abundance of the two gases.This problem will remain unresolveduntil better spectra are obtained show-ing the NH3 band structure and defin-ing the profile with greater certainty.To summarize, we find that (i) either

NH3 or H2S (or a combination of thetwo) can account for the absorptionbetween 2100 and 2400 A, (ii) theabundances of these gases permitted bythe observations are orders of magni-tude below the solar abundance predic-tions, and (iii) neither gas can accountfor the absorption feature near 2600A or the absence of measurable re-flectivity below 2100 A (16).We now find ourselves in the posi-

tion of having contradicted our origi-nal solar-abundance hypothesis. In thecase of ammonia, this dilemma can beresolved by considering the tempera-ture profile in the atmosphere of Jupi-ter. Wildt (2) pointed out many yearsago that the cloud deck observed onJupiter is probably composed of con-densed ammonia. An upper limit forthe ammonia abundance at any levelin the atmosphere is thus set by theammonia vapor pressure correspondingto the temperature of that level. A com-parable restriction on the hydrogenabundance can be obtained from therecent work of Trafton (17), whohas shown that hydrogen can producea greenhouse effect in the atmosphereof Jupiter. The atmosphere thus as-sumes a temperature gradient depend-ent on the amount of hydrogen abovea given level. We have gone one stepfurther and added an isothermal strat-osphere (18) in radiative equilibriumwith the lower atmosphere (Fig. 4).We can now use this model to de-

termine upper limits on the densitygradient of ammonia. Ammonia tem-perature-density profiles for three val-ues of the saturation coefficient a

(cr = 1 represents 100 percent relativehumidity) are shown in Fig. 4. Inspec-tion of the figure indicates that for asaturated atmosphere [the usual as-sumption following Kuiper (6)], theamount of ammonia expected abovethe reference level corresponding to

1492

the ultraviolet observations is only0.006 cm-atm. This is in excellentagreement with the maximum amountpermissible on the basis of the ultra-violet data (0.007 cm-atm).

The problem of the H2S abundanceis not so simple. If we adopt the tem-perature-density profile of Fig. 4, it isnot completely certain that this gaswould condense. Wildt (19) has sug-gested that H2S would not be stablebecause it would be rapidly disinte-grated by hydrogen atoms, leading tothe precipitation of solid sulfur. Itmay be, of course, that Jupiter isdeficient in sulfur compared with thesun, but this conclusion is present-ly premature. It would be extremely in-teresting if this turned out to be thecase, however, since the abundanceratio Fe/ S is greater than 1 in boththe earth's crust and in meteorites(see 9).

10

lo

lm-

tJC:z

m

l4

lI

l0

--I

12 KMAT

tu

MAXIMUMDEPTH OFlop& RADIATION

Y 11 I\_MINIMUM DEPTH OF| X O10 RADIATION

/ / ABUNDANCE

w J UPPER LIMIT FOR MIXED NH3| (MAXIMUM OBSERVATIONAL ERROR)

U)

/\ UPPER LIMIT FOR MIXED NH3. _ lll (EBEST FIT TO OBSERVATIONS)

80 l00 2MP 160 lo 200

TEMPERATURE ('io

Fig. 4. Ammonia vapor abundance in theJovian atmosphere for several values ofthe saturation coefficient. The HRabun-dance profile is taken from Trafton (17).

We still have to account for the ab-sorption feature at 2600 A and theabsence of measurable reflectivity be-low 2100 A. As an alternative to H2S,Stecher (12) suggested that the for-bidden photodissociation of hydrogenmight lead to an absorption near 2600A. We have investigated this possibil-ity and found that absorption near2610 A could result from forbiddentransitions from the fourth vibrationallevel of the ground state (XlYg+) tothe overlying repulsive triplet state(b3y +). At the temperature of Jupi-ter's atmosphere, however, the popula-tion of the fourth vibrational level isdown many orders of magnitude com-pared with the zero level. Using theH2 absorption coefficient derived byStecher (and confirmed in our work)for the 2600 A feature, one wouldpredict an extremely strong absorptionfrom the zero level, and even thetransition from the third vibrationallevel could easily be observed in thelaboratory, contrary to experience. Itthus appears that H2 is also an unlike-ly candidate.We have exami-ned a -long list of other

molecules. It appears improbable thata diatomic molecule could be respon-sible, the only possibility being CS,which is unlikely to be present. Tri-atomic molecules such as HCN, PH3,CH",, and so forth, and the higherhydrocarbons C2H2, C2H4, C2H6 werealso considered, again with no success.Contamination by terrestrial ozone wasalso rejected as a possible explanation.If we restrict ourselves to the relative-ly abundant elements in the Jovian at-mosphere, we find that only rathercomplex molecules produce strong ab-sorptions in this region. The richestsource of such molecules is found froma consideration of carbon chemistry,which suggests benzene as an obviousexample (20). We have examined thispossibility rather carefully. Benzene ex-hibits both an absorption feature at2600 A and a sharp absorption in-crease at 2100 A, but a detailed fitof the absorption profile to the obser-vations does not provide a very con-vincing match. The amount of gasneeded (0.04 cm-atm) is also ratherhigh considering the low vapor pres-sure of benzene at the relevant tem-peratures. (Benzene crystals wouldhave a very similar spectrum to that ofthe vapor, however.) It is interestingthat several aromatic hydrocarbons ex-hibit both a weak dip at 2600 A anda strong absorption edge at 2100 A;

SCIENCE. VOL. 156

_7 . . . . ... /,,r-in't I - X I

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from

unfortunately there is no a priori wayof distinguishing among these mole-cules. Again, the great current needis for observations at higher resolu-tion in order that the fine structure ofthe 2600-A feature becomes availablefor analysis. The implication that somecomplex molecule must be responsibleis of very great interest since its identi-fication may lead to an understandingof the colors observed in the cloudlayer and could provide a firm basisfor assessing the possible biologicalimportance of Jupiter.We may now ask whether this gen-

eral picture is in agreement with ob-servations made at other wavelengths.The temperature of 129°K determinedfrom observations in the 10-fM region(21) is consistent with the presence ofsaturated ammonia above the clouds,as Kuiper (6) has pointed out. A quali-tative explanation of the observed limbdarkening at these wavelengths is alsoafforded by this model. Since the 10-ptammonia bands are saturated, the in-creased air mass toward the limb willlead to lower measured temperaturesbecause the emitting layer will be ata higher altitude. The observed AT of30 to 5°K is in fair agreement witha prediction of 3°K from Fig. 4 (forsec z = 2).

The hydrogen and methane abun-dances we have accepted as our stand-ards (see Table 1) were determinedin the red and near infrared (0.6 to0.9 t). Ammonia has also been ob-served in this region, leading to a der-ivation of 7 to 10 m-atm for themean abundance. Thus we are obvious-ly "seeing" to deeper levels in theatmosphere at these wavelengths thanin the ultraviolet, but while the hy-drogen abundance has increased byonly an order of magnitude, thatof ammonia is more than 104 timesgreater than the amount derived fromthe ultraviolet observations. The dif-ference is due to the fact that theammonia abundance is following thesaturation vapor curve, while thedensity of the hydrogen is increasinghydrostatically, as is evident fromFig. 4.

For this region of the spec-trum, it is not possible to clearlydefine an atmospheric level correspond-ing to a given amount of gas becauseof the presence of the clouds. Thescattering due to the cloud particleswill complicate the process of lineformation, rendering the determinationof abundances rather uncertain. Experi-

16 JUNE 1967

ence with the spectrum of Venus (22)(which exhibits a similar behavior)strongly suggests that in the hydrogenabsorption lines, the incoming radia-tion penetrates to lower levels than inthe adjacent continuum, that is, thehydrogen abundance determinationprobably corresponds to an opticalpath extending well into the cloudlayer. In addition, the condensationof ammonia must be considered whenattempts are made to compare the hy-drogen and ammonia abundances.

In this kind of atmosphere, absorp-tion bands of moderate strength suchas occur in the 1.0- to 2.5-a regionwill correspond to much smalleramounts of gas than the very weakbands we have just discussed. Thus itis not surprising that investigators (23)studying this part of the spectrum de-rive much lower abundances for thevarious constituents we have men-tioned.To summarize these results, we

would say that there appears to be noinconsistency between the ultravioletand infrared observations. The differ-ences in derived values of abundances(most notably that of hydrogen) sim-ply correspond to different levels ofpenetration (or effective optical paths)into the atmosphere. This means thatprecise determinations of abundanceratios must be made with lines ofsimilar intensity in the same region ofthe spectrum, while absolute abun-dances have relatively little meaning.It is clear, however, that a more de-tailed model atmosphere encompassingthe region of the clouds and takingaccount of the internal heat sourcewhich now seems definitely present(24) must be developed to explain allof the observations in detail. The iden-tification of the agent(s) responsible forthe absorption at 2600 A and below2100 A remains a mystery, althoughsome kind of aromatic hydrocarbonseems to be the most probable ab-sorber.

With the possible exception of theH1,S abundance, we have found nodisagreement with our assumption thatthe composition of Jupiter's atmos-phere is basically similar to that ofthe sun. It has not proved possibleto critically test this assumption withavailable data, however, and addition-al observations are highly desirable.With the imminent availability of

spacecraft capable of missions to Jupi-ter, it is refreshing to discover that theplanet grows simultaneously more in-

teresting and more mysterious as it issubjected to additional investigation.It has often been postulated that theatmosphere of Jupiter represents (incomposition) the matter from whichthe solar system was formed. The ap-parent similarity we have pointed outhere between the atmospheres of Jupi-ter and the sun supports this view. Theadditional possibility that complex or-ganic molecules are forming in this"primitive" environment has profoundimplications for studies of the originof life on earth. It seems reasonableto conclude that missions to Jupiterpromise to be among the most reward-ing efforts of the space program.

JACK A. GREENSPANTOBIAS OWEN

Astro Scienices Ceniter,IIT Research Institute,Chicago, Illinois 60616

References and Notes

1. F. R. Zabriskie, Astron. J. 67, 168 (1962).2. R. Wildt. Veriffenitl. Sternw. Gottinigen 2,

No. 22, 171 (1932).3. T. C. Owens, Astrophys. J. 142, 782 (1965),

V. I. Moroz, A strotn. Zh. 43, 597 (1966).4. D. H. Rank, U. Fink, T. A. Wiggins, Astro-

phys. J. 143, 980 (1966).5. The effective air mass is simply the equiva-

lent optical path in the planet's atmospheretraversed by the radiation forming the spec-trum in uinits of a single radial traversal. Inthe case of Jupiter, there appears to be littleif any enhancement of absorption toward theedge of the disk, hence a value of 2 (a sim-ple in and out pathl) seems appropriate. Theabundances arc measured in km-atm wherel km-atm = 2.69 X 102" molecules/cm2(standard temperature and pressure), 1 m-atm = 2.69 X 10'' moleculles cm2, and soforth.

6. G. P. Kuiper, in Attospheres of the Earthand Planiets, G. P. Kuiper, Ed. (Univ. ofChicago Press, Chicago, 1952), p. 306.

7. P. J. Peebles, AstrophYs. J. 140, 328 (1964).8. W. A. Baulm and A. D. Code, Astron. J.

58, 108 (1953). If we add sufficient heliumto bring the molecular weight up to the pre-ferred value of 3.3 given by Baum andCode, we must assume that a large amountof the original hydrogen was lost. In thiscase, it is difficult to see how the C/Hratio can be as small as it is, unless a largeamount of carbon is tied up in some othercompound. The chemical equilibrium calcu-lations (mentioned later in our paper) sug-gest that no other carbon compound is likelyto form in any significant abundance. SinceBaum and Code actually determined the scaleheight rather than the molecular weight, amore reasonable way of achieving a betterfit with their data would seem to be thehypothesis that the local temperature waslower than their assumed value (86°K).

9. L. H. Aller, The Abundance of the Elements(Interscience, New York, 1961); S. Biswasand C. E. Fichtel, Space Sci. Rev. 4, 709(1965).

10. R. Snow, IIT Research Institute ReportC929-3, Chicago (1963). We thank Dr.Snow for many helpful discussions regard-ing the use of this program for our partic-ular purposes.

11. H. C. Urey, Geochim. Cosinochim. Acta 1,209 (1951).

12. T. P. Stecher, Astrophys. J. 142, 1186 (1965).Stecher also considered the effects of Ray-leigh scattering by helium in Jupiter's atmo-sphere. Those results, however, are not ofinterest here.

13. D. C. Evans, Goddard Space Flight CenterReport X-613-66-172 (1966); A.Boggess III, R. Scolnik, Astron. J. 70, 321(1965).

1493

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from

14. These calculationis were performed with NH:absorption coefficients given by B. A. Thomp-son, P. Harteck, and R. R. Reeves, Jr. [J.Geophys. Res. 68, 6431 (1963)]; H2S absorp-tion coefficients given by J. O. Sullivan andA. C. Holland [NASA Contractor Rep. 371,150 (1966)]; and H' scattering cross sectionsfrom A. Dalgarno and D. A. Williams [Astro-phiys. J. 136, 690 (1962)]. The NH:- andH S data were degraded to a resolution of70 A to correspond to the observational data.

15. T. P. Stecher. Anin. Astrophys. 28, 788 (1965).16. For reasonable NH:- abtundances, the calci-

lated albedo increases rapidly below 2000 A.This increase continues with decreasing wave-length down to about 1450 A. Below thiswavelength the albedo again decreases.

17. L. Trafton, AstrophYs. J. 140, 1340 (1964).We thank R. Danielson for calling our at-tention to the efficacy of this approach ininterpreting the atmospheric structure. Daniel-son and Savage (in preparation) have car-ried this type of analysis much further, buttheir purposes differ somewhat from oturs.

18. An isothermal stratosphere compatible witlTrafton's greenhotuse model must havetemperature below 95°K. We have assumedan effective temperatture of 105°K for thetipper atmosphere of Jtupiter [D. Taylor.Icarus 4. T62 (1965)] and derived an ise-

thermal stratospheric temperature of 88°K.19. R. Wildt, Astrophys. J. 86, 321 (1937).20. We thank Professor G. Herzberg for a dis-

cussion of the identification problem whichemphasized this possibility.

21. R. L. Wildey, B. C. Murray, J. A. West-phal, J. Geophys. Res. 70, 3711 (1966).

22. J. W. Chamberlain, Astrophys. J. 141, 1184(1965).

23. G. P. Kuiper, Memn. Soc. Roy. Sci. Liege,9, 365 (1964); V. I. Moroz, Astron. Zhi. 43,579 (1966); R. E. Danielson, Astrophrs. J.143, 949 (1966).

24. F. Low, Lowell Obsertatori' Biull. No. 128,6, No. 9 (1965).

25. We thank C. A. Stone for encouraging thisresearch and for obtaining suLpport for thisstLidy from the in-house research committeeof the IIT Research Institute. J. Woodmanand F. Narin provided valtuable assistancewith the albedo and vapor abundance calcu-lations. The albedo calculations employed inthis paper were originally published by oneof us (J.G.) in a thesis, "Ultraviolet Absorp-tion Studies in Synthetic Atmospheres," Illi-nois Institute of Technology (1966). Thepresent report is Publication No. 198. Spec-troscopy Laboratory. Illinois Institute ofTechnology, Chicaco.

19 December 1966

for 3 hours at room temperature. Itwas necessary to cut each of the juve-niles as short as possible to facilitatethe entry of the fixative. After beingwashed and dehydrated each juvenilewas embedded fiat in Epon 812 (5).Each larva was then removed in apiece of Epon about 1.5 mm't and at-tached to a larger block for cutting.Sections were cut with glass kniveson an LKB Ultratome, mounted ongrids, and stained with Reynolds leadcitrate (6). The sections were examined

Modified Cilia in Sensory Organs of

Juvenile Stages of a Parasitic Nematode

Abstract. LElectron mllicroscopica1l stuidies resealed iie preselice of denidriticnerve processes in sen tor-y or-gans. of tile thir-d- acnd follrtli-stage jliveniles ofHaemonchus contortus, svlicl/i con t((iiled strullctures r'esemlblinig modified cili(l.Witli few exceptions., tle ollter' circle of fiberS confiVSted of ten doubletts, and inp1lace of typical ci('1(-like cenitrial fihers wtere smial nmicrontbile(1i1s or 1'esicles vaIryingin nillnl)ber from z-er-o to file.

Nematodes, especially the free-livingforms, have sensory organs in the formof bristles and papillae, which are

thought to be tactile. In parasitic nc-

Fig. 1. Slightly oblique section of cervicalpapilla of third-stage juvenile showingcilia-like structure and fibers extendingtowards nerve axon at bottom of picture(X 31,680).

1494

imatodes these structures are either ab-sent (although the terminal branchesof papillary nerves are retained) orare present as small protuberances orpits connected to terminal branches ofthe papillary nerves. The amphids. welldeveloped in free-living nematodes butgreatly reduced in parasitic forms, arethought to act as chemoreceptors. Inparasitic forms the amphids may alsohave a secretory function (1).

Very few details are available con-cerning the ultrastructure of these sen-sory organs. Hope (2) described sen-sory organs in Thoralcostolnal californi-CllS, which contained structures thatcould be modified cilia, but he had noevidence for the presence of centralfibers or basal bodies. Roggen et al.(3) described sensory organs in Xiphe-nieia index as consisting, in part, ofa dendritic nerve process having thestructure of a cilium. The number offibers in the cilia, however, varied (forexample, 9 + 2, 9 + 4, 8 + 2, 8 + 4).

Third- and fourth-stage juveniles ofHaemonchus contortus, after a 10- to15-minute treatment at room tempera-ture in 10-3N iodine to straighten them,were fixed in Zetterqvist's fixative (4)

Fig. 2. Closely packed dendritic processesin the amphid near the aperture in third-stage iuvenile ( X 39.900).

Fig. 3. Dendritic processes near the bot-tom of the amphid pouch showing ciliasurrounded by highly vacuolated spaces.Doublet pattern clearly defined. Third-stage juvenile (x 26,040).

SCIENCE, VOL. 156

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from

Jupiter's Atmosphere: Its Structure and CompositionJack A. Greenspan and Tobias Owen

DOI: 10.1126/science.156.3781.1489 (3781), 1489-1494.156Science 

ARTICLE TOOLS http://science.sciencemag.org/content/156/3781/1489

REFERENCES

http://science.sciencemag.org/content/156/3781/1489#BIBLThis article cites 22 articles, 0 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

trademark of AAAS. is a registeredScienceAdvancement of Science, 1200 New York Avenue NW, Washington, DC 20005. The title

(print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for theScience

Advancement of Science. No claim to original U.S. Government Works.Copyright © 1967 The Authors, some rights reserved; exclusive licensee American Association for the

on Decem

ber 9, 2020

http://science.sciencemag.org/

Dow

nloaded from