Structure of Mars' Atmosphere up to 100 Kilometers from the Entry

Home , Viking 2

would be released as the martian atmo- struments flown on sounding rockets and sphere struck the surface, a phenomenon terrestrial satellites, by about a factor of seen earlier during perigee passes of the x 0.01 103. The low-mass spectra are currently Earth orbital satellites Explorer C, Ex- being analyzed: preliminary results sug- plorer D, and Explorer E. The mass peak 15N 14N 15N 14N 14N 14N gest upper limits for the mixing ratios of at 16 includes contributions from CO2, TERRESTRIAL H2 and He relative to CO2 of about 10-4. CO, 02, and H2O, in addition to a residu- A. 0. NIER al most probably due to 0. A quan- School of Physics and Astronomy, titative analysis is difficult, however, 12 170 University ofMinnesota, Minneapolis since 0 may be expected to react rapidly M. B. McELROY with surfaces in the instrument. It is Center for Earth and Planetary hoped that further laboratory studies will 13c 1O J2c 160, Physics, Harvard University, clarify these matters, and that they might Cambridge, Massachusetts 02138 eventually permit a more quantitative estimate for the concentration of 0. 14N 160 References and Notes 1. A. 0. Nier, W. B. Hanson, A. Seiff, M. B. The densities of NO, as shown in Fig. NOISE 12 _8 McElroy, N. W. Spencer, R. J. Duckett, T. C. 2, were obtained from a detailed analysis D. Knight, W. S. Cook, Science 193, 786 (1976). 30 29 28 A typographical error appears in line 11, column of the peak at mass number 30. Figure 3 MASS (AMU) 1, p. 787. The sentence should read "Peaks at gives a scale representation of a typical masses 18 and 17 are...." Fig. 3. Block diagram representing to scale 2. A. 0. Nier, M. B. McElroy, Y. L. Yung, ibid. spectrum obtained by VL1 at an altitude the relative heights for the mass peaks at 30, 194, 68 (1976). near 130 km. Slightly more than half of 29, and 28 amu in a spectral scan obtained 3. K. Biemann, T. Owen, D. R. Rushneck, A. L. near an altitude of 130 km during the descent LaFleur, D. W. Howarth, ibid., p. 76. the signal at mass 30 may be attributed to 4. M. B. McElroy, Y. L. Yung, A. 0. Nier, ibid., of VL1. The diagram shows the contribution p. 70. 12C18O+ formed by the dissociative ioni- of CO+ (from the fragment due to the dis- 5. M. B. McElroy, ibid. 175, 443 (1972). zation of 12C'80160. After correction for sociation of CO2 in the ion source and from 6. F. D. Colegrove, W. B. Hanson, F. S. Johnson, J. Geophys. Res. 70, 4931 (1965); R. S. Lindzen, 12C18O+ from both CO2 and CO, there re- ambient CO) to the several peaks. The remain- Mesospheric Models and Related Experiments, mains a component whose magnitude sig- der, in the case of the peaks at mass 29 and 28, G. Giocco, Ed. (Reidel, Dordrecht, Netherlands, is attributed to ambient N2. The excess peak 1971), p. 120. nificantly exceeds the noise level of the at mass 30 is attributed to ambient NO. The 7. A. 0. Nier and J. L. Hayden, Int. J. Mass is most rea- Spectrom. Ion Phys. 6, 339 (1971). amplifier. This component diagram also illustrates that both the excess 8. A. 0. Nier, J. H. Hoffman, C. Y. Johnson, J. C. sonably attributed to NO, and it is seen peak at mass 30 and the excess peak at mass Holmes, J. Geophys. Res. 69, 979 (1964); A. 0. in all spectra for which the peak at mass 29 (beyond that expected if the nitrogen has a Nier, W. E. Potter, D. R. Hickman, K. Mauers- terrestrial isotopic composition) are well berger, Radio Sci. 8, 271 (1973); J. B. French, 30 exceeds the amplifier noise back- above the noise level of the amplifier which N. M. Reid, A. 0. Nier, J. L. Hayden, CASI Trans. 5, 77 (1972); J. L. Hayden, A. 0. Nier, J. on June 10, 2014 ground. The data indicate a mixing ratio measured the ion currents. B. French, N. M. Reid, R. J. Duckett, Int. J. for NO relative to CO2 of approximately Mass Spectrom. Ion Phys. 15, 37 (1974). 9. M. B. McElroy, T. Y. Kong, Y. L. Yung, A. 0. 10-4. The density of NO in the upper at- Nier, Science 194, 1295 (1976). Mars is thus use electrometer amplifiers rather than 10. J. Mattauch and R. Herzog, Z. Phys. 89, 786 mosphere of significantly (1934). higher than the density of NO at com- electron multipliers. The background 11. Work at the University of Minnesota and at were in Harvard University was supported under NASA parable levels of Earth's atmosphere. fluctuations of these amplifiers contracts NAS- 1-9697 and NAS-1-10492, respec- It is clear from even a casual in- the range (1 to 5) x 10-'4 amp. The in- tively. A.O.N. is indebted to M. Wade and W. spection ofthe data in Fig. 2 that the mar- strumental sensitivity on Viking was Johnson for help in the computations. tian atmosphere must be mixed to therefore less than that for similar in- 12 November 1976 heights greater than 130 km. This obser- www.sciencemag.org vation implies an eddy diffusion coefficient of at least 5 x 107 cm2 sec-1. One would expect a density for N2 of about 5 x 107 cm-3 at 190 km in VLl, or Structure of Mars' Atmosphere up to 100 Kilometers about 108 cm-3 at 160 km in VL2, if diffu- from the Entry Measurements of Viking 2 sive separation should occur above 130 km. The measured densities at these alti- Abstract. The Viking 2 entry science data on the structure ofMars' atmosphere up Downloaded from tudes are only 2 x 107 cm-3 and 4 x 107 to 100 kilometers define a morning atmosphere with an isothermal region near the cm-3, respectively. The densities as surface; a surface pressure 10 percent greater than that recorded simultaneously at measured for CO are also consistent with the Viking I site, which implies a landing site elevation lower by 2.7 kilometers than the assumption of rapid vertical mixing, the reference ellipsoid; and a thermal structure to 100 kilometers at least qualita- as discussed elsewhere (9). tively consistent with pre-Viking modeling of thermal tides. The temperature profile The upper atmospheric mass spec- exhibits waves whose amplitude grows with altitude, to 25°K at 90 kilometers. trometers on both VL1 and VL2 used These waves are believed to be a consequence of layered vertical oscillations and Mattauch-Herzog geometry (10), which associated heating and cooling by compression and expansion, excited by the daily allowed for simultaneous collection of thermal cycling ofthe planet surface. As is necessary for gravity wave propagation, ions with different masses. Two collect- the atmosphere is stable against convection, except possibly in some very local re- ors were used to measure ions differing gions. Temperature is everywhere appreciably above the carbon dioxide con- in mass by approximately a factor of 7 densation boundary at both landing sites, precluding the occurrence ofcarbon diox- (7). The heavy collector was sensitive to ide hazes in northern summer at latitudes to at least 50°N. Thus, ground level mists the mass range 7 to 49 atomic mass units seen in these latitudes would appear to be condensed water vapor. (amu), whereas the light collector recorded ions in the range 1 to 7 amu. Because The second Viking lander entered the Time (P.D.T.), and landed at 3:58:20 of weight restrictions imposed at an early atmosphere of Mars on 3 September p.m. P.D.T. in Utopia Planitia at 47.66°N stage of the project, we were obliged to 1976 at about 3:49 p.m. Pacific Daylight latitude and 225.78°W longitude at 9:06 1300 SCIENCE, VOL. 194 a.m. Mars local time. During entry and Table 1. Surface elevations at the Viking 2 landing site indicated by measurements of the plan- descent, measurements were made to de- et's radius after landing. fine the structure of the atmosphere be- Radius (km) Landing low about 120 km, as on Viking 1 (1). site Smooth elevation Real-time transmission of the entry data eismooth Measured b - a to the earth, by relay through the orbiter, (a)elsi(b) (km) was prevented by problems with the atti- tude stabilization of the orbiter during Topographic 3385.0 + 0.7(14) 3381.4 + 1.3(15) -3.6 ± 1.5 this period. The data were recorded on Gravitational 3384.3 ± 0.7(14) 3381.5 ± 0.6(16) -2.8 ± 0.9 tape recorders, however, on both the orbiter and the lander, and were trans- to mitted the earth from the orbiter rec- lander radar altimeter below about 130 difference with the measured pressure ord almost exactly 24 hours later. The re- km were also invaluable. difference. sults presented here are based on The first data interpreted after touch- The profile of density for the altitude analysis of the first orbiter replay. down were the temperature and pressure range to 100 km defined by Viking 2 (pri- We present a preliminary analysis, measurements during parachute descent, marily by the accelerometers) is shown which does not include all the correc- shown in Fig. 1. Significant atmospheric in Fig. 2. (The upper altitude limit will ul- tions and refinements that will ultimately temperature data were obtained by the timately extend to about 125 km.) Den- be applied. The primary corrections sensor mounted on footpad 2, from the sities determined independently from omitted here are corrections for effects time ofleg deployment at 3.9 km to retro- stagnation pressure measurements are in- due to atmospheric final winds, correc- rocket firing at 1.5 km. The dynamic cor- cluded as circular symbols, and agree tions for terrain variations under the rections shown applied to the measured well with those defined the decelera- flight path, and by small corrections for temperatures, about 1.5°K, are required tion. The curve joins smoothly with para- such instrument effects as response lag for descent velocities on the parachute chute descent data based on and temperature radiative input. [The latter are known between 50 and 60 m/sec, as measured and pressure with an assumed composi- to be small from previous studies (2).] We by the Doppler radar during descent. tion of 0.947 CO2, 0.035 N2, 0.015 Ar, believe that the results presented here The atmosphere is seen to be isothermal and 0.003 02 (see below). The computed substantially represent the atmosphere from 1.5 to 4 km, with a lapse rate of at surface density is 0.0180 Curva- sensed by Viking 2. kg/m3. most 1.3°K/km above 2.5 km. This is in ture and waviness in the profile signify a The instruments and techniques ap- contrast to the lapse rate of 3.7°K/km re- nonisothermal atmosphere. plied to obtain and analyze the data have ported for Viking 1 (1). Integration of the density distribution been described (2, 3). The primary in- This difference between the near-sur- under the assumption ofhydrostatic equi- strument for altitudes above 25 km was a face temperature profiles is a diurnal ef- librium yields the pressure with al- three-axis set of profile accelerometers, accu- fect, which has been discussed theo- titude given in Fig. 3. The pressure pro- rate to within 0.02 percent of reading. retically (4). In the morning profile of Vi- file is less wavy than the density profile, Two sets of pressure and temperature king 2, the near-surface atmosphere has and again goes very smoothly into the di- sensors were provided, one for the entry been cooled overnight by radiation. In rectly sensed pressures in the parachute phase before heat shield jettison and the late afternoon profile of Viking 1, it descent. Altitudes are measured above parachute deployment (at nominally 6 has been warmed by midday radiation the landing site, throughout. km in altitude) and the other for the para- from the surface. From the densities and pressures in chute descent. Altitude data taken by the The atmospheric pressure profile de- Figs. 2 and 3 and the equation of state, fined by parachute-phase sensing from the temperature profile in Fig. 4 is calcu- 1.5 to 4.5 km and its correction for de- lated. An atmospheric mean molecular 8 JETTISON scent AEROSHELL velocity are shown in Fig. IA. Cor- weight of 43.34 was assumed, based on rections range from about 0.2 mbar at the 7 Pmeas. lower altitudes to about 0.5 mbar at 4.5 100 _ E DYNAMIC km. The atmospheric pressure curve ex- PRESSURE trapolates to 7.75 mbar at the surface (5). 90 _ Simultaneously at the Viking 1 site on 80 _ - ,from that date, the pressure was 6.98 mbar, 70 ax and the mean for E so Ps that.sol was 7.13 mbar .X50 - (6). This indicates that the landing site of 40 Viking 2 was 0.96 to 1.20 km below that 30 of Viking 1 (7). The "6. l-mbar reference 20 surface" is 2.7 km above the Viking 2 10 from p&T X\ 0° I I' 1% landing site (8). 10-7 104a 1o5 10-4 10-3 joe-2 lo-' Low surface elevations at the Viking 2 (p, kg/m3) landing site were also indicated by planet Fig. 2. Profile of atmospheric density to 100 radius measurements after landing, by km in altitude. For the acceleration data re- means ofradio tracking and measured duction, laboratory drag coefficients in a CO2 ac- atmosphere at simulated velocities and Rey- z, km DEPLOY celeration due to gravity (3.7307 m/sec2), nolds numbers are used; these are believed ac- SENSOR as shown in Table 1. Elevation measured curate to within 1 percent below about 85 km Fig. 1. Measurements of temperature and at the Viking 1 site was - 1.7 (2). The stagnation pressure analysis is based pressure taken during parachute descent. The to - 2.0 km. Thus, the elevations measured in- on the same velocity history, but is otherwise interval of good data is constrained by de- independent. The accelerometer data will, scent engine firing at 1.5 km and (for temper- dependently are consistent with the when analysis is complete, fill the gap shown ature) by lander leg deployment at 3.9 km. measured pressures, and the elevation from about 5 to 25 km. 17 DECEMBER 1976 1301 100- VIKING 2 (11). The latter are phase-shifted relative 90 \ to the Viking 2 profile-there is a maxi- 80 -

70 - mum in the occultation immersion pro- 60 file at 80 km and a minimum at 70 km-a E 50 _ result, no doubt, of the times of day at N 40 the occultation sites, which were not 30 stated. The further analysis of the Viking 20- 1 thermal structure has not altered its 10 _ o basic form, but has extended it in alti- 10-5 10-4 10-3 10-2 lo-1 10 1 (p, mb) tude and somewhat increased its regulari- ...... 55 ty. .. ._ . V,< Fig. 3. Pressure profile of Mars atmosphere to As required to support gravity waves, 100 km. Above 90 km, the uncertainty in tem- the Viking 2 temperature profile is con- J perature at the experiment threshold also af- 40 60 80K 0 8 16 24 fects pressure by perhaps + 10 percent at 100 vectively stable, probably so even in the AMPLITUDE HOUR OF MAXIMUM km, diminishing to + 1 percent at 80 km. regions of decreasing temperature above TEMPERATURE the wave peaks. The upward propaga- Fig. 5. Vertical profiles of diurnal temperature tion of the diurnal surface heating there- amplitude and phase, after Zurek (10). The data by the upper cases presented here are for northern lati- given atmospheric (1) fore takes place predominantly by radia- tudes at southern summer solstice on a and landed (9) mass spectrometers, used tion, and hence depends quantitatively smooth planet at the prime meridian. Height to select the composition given above. on the dust content of the atmosphere. is in units of 10 km. The time dependence of The temperature profile is character- The temperatures, as for Viking 1, are the temperature is sinusoidal (in the model). ized by a subadiabatic lapse rate of everywhere well above the CO2 con- The amplitude typically varies periodically with altitude, but in one case (6 = 190 lati- 1.80K/km between 5 and 19 km,. above densation boundary. Since this is a morn- tude) does not. which there is a temperature oscillation ing profile, it appears unlikely that CO2 about the mean. This wave appears to hazes form at latitudes to 50°N in north- confirm the presence of the diurnal ther- ern summer. The minimum overnight mal tide modeled for martian atmospher- ground temperature of 180°K also sup- profiles shown are also consistent with ic conditions by Zurek (10). Figure 5, re- ports this conclusion (12). Water ice temperature data obtained by Nier (13) produced from Zurek (10), shows the cor- clouds would form, of course, given suf- from the entry science mass spectrome- respondence. Although none of the cases ficient water vapor, and where morning ter, at altitudes above 120 km. There will presented matches the Viking 2 latitude mists are seen in these summer latitudes be further efforts to improve analysis of and season, the similarity in the wave it is presumed that they are water. the threshold region, possibly by joint structure is evident. More detailed com- The two curves shown above 80 km in consideration ofdensities from the accel- parisons should indicate the changes re- Fa'ig. 4 are for two assumed initial temper- erometer and the mass spectrometer to quired in the model to achieve quan- altures at 100 km. These initial temper- determine pressure at 120 km without as- titative correspondence. attures were selected from an analysis of sumption. Vertical wavelengths (altitude differ- thie data to 110 km, in which starting tem- A computation was performed to test ences from peak to peak) range from 17 eratures from 100° to 170°K converge to the sensitivity of the temperature profile to 23 km in the data. Theoretical 1:35° + 5°K at 100 km. All of these cases to uncertainties in the lander drag wavelengths range from 22 to 24 km. The cnverge within ± 1°K at 80 km. The coefficient and angle ofattack at the high- physical interpretation of the temper- est altitudes (above 90 km), for which the ature maxima and minima is that the Reynolds numbers are very low and the fluid, oscillating vertically in the ther- 00 VIKING 2 aerodynamics not as well defined. An mally driven gravity waves, is alternate- 90 - angle of attack of 11.20 was, in this test ly heated by compression and cooled by 80 - calculation, imposed throughout the re- expansion. The oscillations occur in sev- U70 gion above 60 km, instead of permitting eral layers, whose thickness is the appar- f012 _ ( from a. angle of attack to be defined by the ratio ent wavelength in temperature, with of lateral to axial acceleration, and the phase alternation between layers, while 50 - drag coefficient corresponding to 11.2° the gravity wave propagation has a major 40 - Adiabat was imposed, instead of allowing the horizontal component. Radiative trans- 30 from p, & T, computer program to select drag fer acts to reduce the temperature differ- 20 _ \, t/ coefficient from tables based on angle of ences To generate the ob- vertically. 10 Condensation\\ \ from Tpara. attack. Changes in drag coefficient up to served temperature amplitudes adiabati- boundary I 8 percent were thus introduced. Because cally at the six peaks below 80 km 80 120 \,160 200 240 of compensating effects, the changes in requires compression ratios from 1.26 to T, K temperature profile were small, nowhere 0.80 in pure CO2. For these six peaks, Fiig. 4. Temperature structure of the atmo- greater than 20K. Hence, current uncer- the vertical motion amplitudes need be sphere seen by Viking 2. Above 29 km, tem- tainties in the aerodynamics at extremely perature is derived from acceleration data; be- only 0.6 to 2.1 km. The uppermost peak, low that, from measured temperatures and high altitudes apparently do not affect if read to a mean temperature of 134°K, priessures. The three sets of independent the results presented. requires a motion amplitude of 6.6 km. measurements, including those on the para- ALVIN SEIFF The wave structure was also seen in chute, are highly self-consistent. The limited DONN B. KIRK isothermal region near the ground gives way the Viking 1 preliminary temperature to a region of subadiabatic lapse with what is Space Science Division, profile (1) and in the profiles derived apparently the diurnal thermal wave superim- NASA Ames Research Center, from the occultation of s Geminorum posed. Moffett Field, California 94035 1302 SCIENCE, VOL. 194 Refereafti sa Notes 10. R. W. Zurek, J. Atmos. Sci. 33, 321 (1976). netic, relatively pure mineral. This was 11. J. L. Elliot, R. G. French, E. Dunham, P. J. 1. A. 0. Nier, W. B. Hanson, A. Seiff, M. B. Gierasch, J. Veverka, C. Church, C. Sagan, considered most likely to be magnetite, McElroy, N. W. Spencer, R. J. Duckett, T. C. Science, in press. D. Knight, W. S. Cook, Science 193,786 (1976). 12. H. H. Kieffer, S. C. Chase, Jr., D. D. Miner, F. and by comparison with laboratory tests 2. A. Seiff, Space Sci. Instrum., in press. D. Palluconi, G. Munch, G. Neugebauer, T. Z. on terrestrial materials its abundance 3. A. 0. Nier, W. B. Hanson, M. B. McElroy, A. Martin, ibid. 193, 780 (1976). Seiff, N. W. Spencer, Icarus 16, 74 (1972). 13. A. 0. Nier, personal communication. was estimated as 3 to 7 percent by 4. P. Gierasch and R. Goody, Planet. Space Sci. 14. E. J. Christensen, J. Geophys. Res. 80, 2909 weight. 16, 615 (1968). (1975). 5. Thp fact that the atmospheric pressure curve 15. Radio tracking radius (personal communication On sol 40, a high resolution, direct ppSAs near the measured values during descent from E. Euler of the Viking Flight Team). view image of the backhoe was received, ergine operation is probably fortuitous. 16. Radius from measured acceleration due to grav- 6. Viking Meteorology Team, personal communi- ity (personal communication from E. J. Christen- after a cumulative total of 13 insertions cation. sen). 7. An order of magnitude estimate shows that the 17. The authors extend their appreciation to S. into the surface. This image indicates pressure difference between the sites cannot be Cook of the Viking Entry Science team and to that slightly more material was held on accounted for by dynamics. Pressure differ- E. Euler, F. Hopper, R. Dupree, and others of ences due to measured wind velocities are on the Viking Flight Team who provided invaluable sol 40 than on sol 28, and that some milli- the order of 10-3 mbar. support. Thanks also are due to many colleagues meter-sized fragments were still adher- 8. It is now clear that this reference level cannot be at Ames Research Center, in particular S. Ro- identified with a specified atmospheric pressure, gallo for computer programming support and to ing on sol 40; since pressure varies seasonally [S. L. Hess et S. C. Sommer for support during the mission On sol further of the back- al., Science 194, 78 (1976)]. By chance, pressure operations period. The authors are further in- 41, images at the reference ellipsoid was 6.14 mbar (from debted to C. B. Leovy for helpful discussions on hoe were received that were taken dur- Viking I data) on 3 September 1976. the thermal tidal processes. 9. T. Owen and K. Biemann, Science 193, 803 ing a special sequence in which insertion (1976). 18 October 1976 of the backhoe into the surface was fol- lowed by imaging, by way of the magnifying mirror, the material held on the front of the backhoe both before and af- Viking Magnetic Properties Investigation: Further Results ter vibration. Considerably more material was held on the magnets before than Abstract. The amounts ofmagnetic pdrticles held on the reference test chart and after vibration of the collector head, in- backhoe magnets on lander 2 and lander 1 are comparable, indicating the presence dicating that much non- or weakly mag- of an estimated 3 to 7 percent by weight ofrelatively pure, strongly magnetic parti- netic material entrained with the strongly cles in the soil at the lander 2 sampling site. Preliminary spectrophotometric analy- magnetic particles was purged by this vi- sis ofthe material held on the backhoe magnets on lander I indicates that its reflec- bration. tance characteristics are indistinguishable from material within a sampling trench A list of images of the RTC and back- with which it has been compared. The material on the RTC magnet shows a different hoe magnets received since sol 34 is giv- spectrum, but it is suspected that the difference is the result ofa reflectance contribu- en in Table 1. [For a list of previous im- tion from the magnesium metal covering on the magnet. It is argued that the results ages received, see (2).] indicate the presence, now or originally, ofmagnetite, which may be titaniferous. Results from Viking 2: The RTC magnet. A low resolution image made in the The aim ofthe Viking magnetic proper- and the front by way of a 4 x magnifying survey mode on Viking 2 (VL2) on sol 0, ties investigation (1) is to estimate the mirror. The RTC magnet is a single array in which no clear bull's-eye pattern can abundance anid composition of magnetic equivalent to a strong backhoe array. be discerned, and a high resolution im- minerals in the martian surface material Resultsfrom Viking ). Results from Vi- age on sol 6 were received before any at the Viking landing sites. Two magnet king 1 (VL1), based on images received sampling activity had occurred (Fig. la). arrays in the surface sampler backhoe en- up to sol 34, have already been described In the sol 6 image there was significantly ter the surface material during sample ac- (2) and are summarized here. more material on the magnet than was quisition and attract and hold any mag- After initially attracting some particles observed on the Viking 1 magnet at the netic particles present. Another array out of the dust cloud raised by the termi- nearest corresponding time (sol 3). This mounted on one of the imaging reference nal descent engines during the final could have been the result of a relatively test charts (RTC) on top of each lander stages of the landing, the magnet on the larger or denser dust cloud which might attracts magnetic particles already in, or RTC attracted considerably more materi- have been raised during the landing of raised into, the Martian atmosphere. al during the mission, indicated by a VL2 because of an extra burn of the ter- Details of the magnet arrays have been well-defined bull's-eye pattern of parti- minal descent engines just before touch- given by Hargraves et al. (2). Briefly, cles over the magnet array. The chief down. It is believed that the material was each array consists of a cylindrical mag- source of this material is believed to be not seen in the sol 0-image of the magnet net surrounded by, and separated from, dust raised by surface sampling activity because of its small size and the lack of a ring magnet. The overall dimensions of and subsequent distribution of the soil contrast. A comparison of the material each array are 1.8 cm diameter and 0.3 samples to the three analytical experi- held on the RTC magnets on sol 15 (VL1) cm thick, and the central and ring mag- ments. There may also be a contribution and sol 13 (VL2) shows somewhat less nets are oppositely magnetized along from magnetic particles attracted from material held on the latter than the their axes. Two arrays are placed in the the martian atmosphere. former. On both landers there was sam- backhoe in such a way that their surfaces The backhoe magnets attracted a con- pling activity on sol 8, but because of an are respectively 0.5 mm and 3.0 mm be- siderable quantity of relatively low al- anomaly on the surface sampler only one low the backhoe surface on one side and bedo magnetic particles from the soil dur- delivery had occurred on VL2 as op- 3.0 mm and 0.5 mm below on the other; ing surface sampling activity, including posed to four on VL1. This difference thus, at each face of the backhoe there one or two small fragments 2 to 3 mm probably explains the above observa- are two levels of attractive force, in the across. The existence of well-defined tions. approximate ratio of 12: 1. Images of bull's-eye patterns of similar extent on A further comparison of the RTC mag- the backhoe are periodically taken by the both strong and weak magnets indicated nets on the two landers can be made for lander cameras, the back face directly the presence in the soil ofa strongly mag- sols 31 (VLI) and 33 (VL2). After a simi- 17 DECEMBER 1976 1303