The Surface of Mars 1. Cratered Terrains

Home , Dewar (crater), Kepler (Martian crater), Kuiper (Martian crater), Mariner (crater), Ranger 7

VOL. 76, NO. 2 JOURNAL OF GEOPHYSICAL RESEARCH JANUARY 10, 1971

The Surface of Mars

i. Cratered Terrains •

BRUCEC. MURRAY,LAURENCE A. SODERBLOM, ROBERTP. SHARP,AND JAMESA. CUTTS Division of GeologicalSciences California Institute of Technology,Pasadena 91109

Mariner 6 and 7 pictures show that craters are the dominant landform on Mars and that their occurrenceis not correlated uniquely with latitude, elevation, or albedo markings. Two distinct morphological classesare recognized: small bowl-shaped and large flat-bottomed. The former showlittle evidenceof modifications,whereas the latter appear generally more modified than hmar upland cratersof comparablesize. A regionalmaria/uplands dichotomylike the moon has not yet been recognized on Mars. Crater modification on Mars has involved much greater horizontal redistribution of material than in the lunar uplands. It is possiblethat there are erosionalprocesses only infrequently active. Analysisof the natures and fluxesof bodiesthat have probably impacted the moon and Mars leads to the likelihood that most of the large flat-bottomed craters on Mars have survived from the final phasesof planetary accretion. Significant crater modification, however, has taken place more recently on Mars. Inasmuch as the present small bowl-shaped craters evidence little modification, the postaccretion crater-modification processon Mars may have been primarily episodic rather than continuous. The size-frequency distribution of impacting bodiesthat produced the present small Martian bowl-shapedcraters differs from that responsible for post-mare primary impacts on the moon by a marked deficiency of large bodies. Survival of crater topography from the end of planetary accretion would make any hypothetical earthlike phasewith primitive oceansthere unlikely. The traditional view of Mars as an earthlike planetary neighbor in terms of its surface history is not supported by the picture data.

Perhapsthe most excitingresult of the Mariner Mariner 6 and 7 television pictures as they bear 4 television experiment in 1965 was the discovery on the nature of the Martian surface.The present of cratered terrains on Mars, a generally unex- paper deals with the nature and significanceof pected addition to the annual development of cratered terrains. The second is concerned with the frost caps, seasonaldarkenings, and other uncratered terrains [Sharp et al., 1971a]. The presumed earthlike phenomena.The television third paper discussesphotographic observations pictures returned by Mariners 6 and 7 in 1969 relevant to the nature of light and dark markings greatly extended knowledgeof Martian cratered [Cutts et al., 1971], and the fourth describes terrains and showed two new uncratered terrains observationsand implications of surface features as well. Close-up pictures were obtained of the of the south polar cap [Sharp et al., 1971b]. We south polar cap and also of certain prominent shall refer to them as papers 1, 2, 3, and 4, light-dark boundaries. These and other results respectively. of the television experimentswere first discussed The picture data used are principally the in three preliminary reports published shortly 'maximum-discriminability' versions displayed after receipt of the picture data [Leightonet al., in the accompanying papers by Dunne et al. 1969a, b, c]. The objectiveof this paper and of the [1971] in this issue, supplemented by other three companionpapers is to describethe results versions of the near- and far-encounter photog- of subsequentanalysis and interpretation of the raphy. Image-processing techniques are discussedin a separate paper in this issue [Rind- fieis& et al., 1971]. Picture location and notation i Contribution 1891, Division of Geological are summarized in the preceding article in this Sciences, California Institute of Technology, Pasadena. issueby Leighton and Murray. Craters are visible in 52 of the 55 near-en- Copyright • 1971 by the American Geophysical lJnion. counter frames (best resolution 0.3 km) and

313 314 M•URRAY ET AL.

constitute the principal landform observed. At surface processesand history of Mars inferred this early stageof Martian exploration we define from the evidence of the terrains already recog- cratered terrains to be regions of the surface nized will provide a useful framework for more in which craters are the dominant, often the detailed information and knowledge to be ac- only, topographic forms recognizable at the quired by future spacemissions. resolution of the Mariner 6 and 7 pictures. In the following, the geographic distribution Intercrater areas are included in this definition. of cratered terrain is reviewed and possible As better pictures become available, more correlations,with elevation, latitude, and albedo sophisticatedcriteria may haveto be developed are considered.Morphology of local features and to categorizewhat may well be a variety of crater abundances is then treated. Martian }Martian terrains. We feel, nevertheless,that the observations are compared with those of the

4N14

Fig. la. Positions of Mariner 4 photographs4N7 through 4N14 plotted on Mariner 7 far- encounterphotograph 7F76 with a sub-spacecraftlongitude of 199øE.The ringed structure,Nix Olympica, is visible in the northern hemisphere. SURFACEOF MARs--CRATERED TERRAINS 315

4N9

Fig. lb. Crateredterrain displayedin Mariner 4 frames4N7 through4N14, taken across Zephyria,Mare Sirenurn,Mare Cimmerium,and Phaethontis.(See paper 3 for identification of namedfeatures.) Frames 4N8 and 4N13 containprominent light/dark boundaries.Individual frames are about 250 km on a side. Longitude, latitude, and solar elevation angle range from 186øE,13øS, 61 ø for 4N7 through209øE, 42øS, 29 ø for 4N14. Detailsare availablein Leighton ct al. [1967]. lunar-crateredterrains, leading to the conclusion or dark areas or with any particular elevation that many largeMartian craters,like thoseof the range, and it contains two distinct types of lunar uplands,have survivedsince the last phases craters: large fiat-bottomed and small bowl- of planetary accretion. The later history of shaped. In this section we summarize the Mars, however,appears to differ significantly characteristics of Martian cratered terrain and from that of the moon. Finally, the traditional compareit with the crateredterrain of the moon. view of Mars as once having experiencedearth- Geographicdistribution. Craters ranging in like conditions is recvaluated in light of the diameter from a few hundred meters to a few similarities of the Martian cratered terrain to the hundred kilometers are visible in the Mariner 6 lunar uplands. and 7 photographs.If this sampleis representa- tive, cratered terrain (including intercrater OBSERVATIONS OF MARTIAN CRATERED TERRAINS areas) constitutesat least 90% of the Martian Cratered terrain on Mars was first revealed surface. in televisionpictures returned by Mariner 4 in Somevery largecraters are recognizablein 1965. Mariners 6 and 7 extended the observations the far-encounter frames of Mariner 6 and 7, of cratered terrain sufiqciently to demonstrate especiallyin the dark area Syrtis Major (frame that it is the dominant Martian landscape;it is 7F87) and Mare Cimmerium(frame 7F82). The not uniquelycorrelated in occurrencewith light featureNix Olympica,long recognized as a bright 316 MURRAY ET AL. patch from earth-basedobservations, is seenat Figures3 and 4 of paper2 and Figures1 and 2 of higherresolution (frame 7F77) to be a multiple- paper 4). ring structureof maximumdimension exceeding Figureslb, 2, and 3 in this paperand Figure 3 500 kms. Presumablythis is at least the remnant in paper4 are mosaicscomposed of the bestnear- of a very large crater. Circular features almost encounterphotography from Mariners4, 6, and 7. as large are faintly visiblein bright areaswithin They all show predominatelycratered terrain. frames 7F83 and 7F84. The polar cap edge Each of thesemosaics is discussedbriefly in the observed in the late far-encounter frames of both context of the regional associationsof Martian Mariner6 andMariner 7 is partlydelineated by cratered terrain. largec•aters up to 110 km in diameter(see Craters from 4 to 350 km in diameter are

Fig. 2. Area of mosaicof frames6N9 through6N23, shownin outlineon the far-encounter frame7F67, i•mluding the dark areasMeridiani Sinus and SabaeusSinus and the light area l)eucalionisRegio. (Geographic names are locatedin paper3.) Solarelevation angles, which rangefrom 52ø downto 3 ø are shownat the lowerright-hand corner of the B frames.The outline of thesenarrow-angle frames is shownin whiteon the mosaic of A frames,along with latitude and lo•gitudegrid li•tes. Individual A framesare of the orderof 1000km in the longdirection; B framesare l/10 scaleof A frames.One degreeof latitude on Mars is 59 km. Maximum discriminability versionshave beenused that greatlyaccentuate topographic detail but distortand suppresssome albedo marking, especially the outlines of Meridiani Sinus. The crater counts usedi•t Figre'es4, 5, and6 arederived from frames 6N16 to 6N23and refer principally to the I)eucalionis Regio area. SURFACEOF •/•ARS--CRATEREDTERRAINS 317 318 M•RR^¾ ET AL.

visible in Mariner 4 frames 4N5 through 4N14, By way of comparison, on the moon two distinct as is shown in Figures l a and lb. A prominent kinds of cratered terrain are evident that bear a- light/dark boundary,the northernboundary of simple relationship to height and light and dark Mare Sirenurn, is included in the pair of frames markings. On the moon, large, highly-modified 4N?/4NS. Crateredtopography does not appear craters are restricted to the uplands. Mare to vary significantly across this boundary. formation presumably obscuredmost of the older The near-encounter frames from Mariner 6 preexisting craters in the maria basins. So, nearly all encompasscratered terrain except for whereas the lunar uplands are rough, elevated, small areas of chaotic terrain (see Figure 2 of bright terrains containing numerous old fiat- paper 2•. The mosaic of Figure 2 covering the floored craters, the lunar maria are dark, low in equatorial areas Meridiani Sinus and Deucalionis elevation, smooth, and essentially devoid of Regio exhibits cratered terrain especially well old highly modified craters. Mariner 6 and 7 and illustrates again that the general character of photographs do not reveal a similar regional cratered terrain can remain unchanged across associationof Martian craters. First, both light light/dark boundaries. and dark areas of Martian cratered terrain The near-encounter frames of Mariner ?, contain large fiat-bottomed craters generally included in the mosaic of Figure 3, also display in comparable abundance. The lunar correlation principally cratered terrains except for the of large degraded craters with high albedo and featureless floor of Hellas. These frames show the uplands is not evident on Mars. Hence, the presence of craters at southern midlatitudes characteristic associationof crater form, eleva- (20ø to 45øS). tion, and albedo denoted by the term 'mare' on The mosaicin paper 4 (Figure 3) demonstrates the moon, is apparently absent on Mars, at the occurrence of numerous craters under the least in those areas photographed in the near- south polar cap. Thus, cratersare abundant in the encountersof Mariners 4, 6, and 7. Hellas, the equatorial, midlatitude, and south polar regions large circular area of featurelessterrain discussed of Mars, evidencing no marked latitudinal in paper 2 of this series, might conceivably dependence of cratered terrains. The continua- represent a mare-like surfaceeven though it is of tion of cratersacross the Noachis/Hellespontushigh albedo. However, the marked absence of boundary (Figure 3) again demonstrates the any visible impact features would require it to be absenceof any unique correlation between light extremelyyoung. A continuousprocess modifying and dark areas and cratered terrain. the floor of Hellas seemsmore likely. The distribution of cratered terrain has been Topographic relief. A striking difference examined for possiblecorrelation with elevation. between the Martian and lunar surfaces is the A limited quantity of reliable data on regional lower relief of the walls of large craters (diameter elevation has been obtained from ground-based > 15 km) and of the intercrater areas on Mars. radar observations[Goldstein et al., 1970; Lincoln This was first recognizedfrom the Mariner 4 TV Laboratory, 1970] and from equivalent width experiment [Leightonet al., 1965], although in variations in C02 absorption bands measuredby that instance large spurious background light the onboardinfrared spectrometersof Mariners 6 levels originating from some unknown optical and 7 [Herr et al., 1970]. Cratered terrain occurs degradation weakened the credibility of this over a wide range of elevations;no unique cor- conclusion[Young, 1969]. The Mariner 6 and 7 relation with height is evident. camera systems were free of such anomalous In summary, the cratered terrain of Mars is effects;the gentlenessof Martian topographyis the most extensive terrain on the planet. It again indicated. occurs over a wide range of latitudes and eleva- Several specificlines of evidenceare available. tions and is not confinedto either light or dark As is evident in Figure 7a, the horizontal width areas or to a particular elevation or latitude of walls of large fiat-bottomed craters (diameter range. > 15 km) on Mars is generally several times Although variations in the character of the smaller than the width of walls of similar size cratered terrain do occur [Cutts et al., 1971], craters on the moon. Thus the relief across such they are not clearly defined,and the relationships walls must be much less than on the moon, with elevation, albedo,and latitude are complex. provided that the wall slopeson Mars are not 319

correspondinglygreater. It is clear that such is highly modified from their presumed initial not the case because the conspicuousshadows appearance as impact craters. Rims are not that would be cast by such steep walls are not detectable or greatly subdued, central peaks are found. In addition, slopes of walls of large fiat- rare, and other impact-associatedfeatures, such bottomed craters can be estimated from oblique as secondarycrater swarmsand cjccta blankets, views independently of the lighting conditions or have been greatly modified, usually beyond photometric properties of the surface. This is recognition. Some of these craters, here termed done by assuming that these large craters are vestigialor ghost,have been so greatly modified approximately radially symmetric and by com- that the relief of their walls is only faintly paring the projected widths of the near and far visible. Other large flat-bottomed craters are walls measured on a photograph. If the slopesare much more easily recognized. These two states low, these apparent widths are nearly the same of preservation are well shown in frame 6N16. over a wide range of viewing angles. If the We do not see a clear gradation between these slopesare great (i.e., the order of the emission two types of flat-bottomed crater; they may angle), then the relative apparent widths vary reflect a complex episodicphase of early Martian rapidly with emission angle. Such tests were history. made on two large fiat-bottomed craters in the Small bowl-shaped craters represent the northeast corner of 6N16 (emission angle majority of cratersobserved so far with diameters •43ø). Such estimates give slopesof the order of below 10 to 15 kin. These smaller craters exhibit 10ø; certainly they are not more than 20 ø. almost all the associated impact phenomena Finally, low average slopes on Mars are sug- found with lunar primary craters of similar gestedby the rather narrow width of the marginal size. The associatedslump blocksand secondary zone of the south polar cap. As is discussedmore crater swarms are not visible, as expected, as fully in paper 4, the latitudinal width of that these would be below the limit of resolution. zone is probably controlled by the magnitude of The small bowl-shaped craters appear to be of crater slopes;average slopesof the order of 5 ø or uniform morphology, indicative either of very less on scale of several kilometers are indicated little or very uniform modification. along the polar-cap edge. In summary, the A parallel is found in comparing the crater relief of large fiat-bottomed craters (diameter populations on Mars with those on the lunar > 15 km) on Mars is less, perhaps several times uplands. Most of the large craters in the lunar less, than that found on the moon, and the uplands also have suffered significant modifica- average slopes of large crater walls and of tion (Figure 7a), as a result of which their surroundingterrain also may be less than those ejccta blankets are no longer recognizable and found in the lunar uplands. Small bowl-shaped secondary crater swarms have been obliterated. craters, however, may or may not differ signifi- Many exhibit smooth, flat floors. Some have cantly in form from fresh craters of the same lost rims and central peaks as well. Also present diameter on the moon; that analysis is not yet in the lunar uplands are abundant fresh small complete. craters (Figure 7b), which have sufferedlittle, if Morphology of craters and associatedfeatures. any, modification. Thus, both the lunar upland Two distinct types of craters are distinguished and the Martian surfacesgenerally display two in the areas viewed by Mariners 6 and 7 and are families of craters: abundant, fiat-flooredcraters referred to here as small bowl-shapedand large that have been significantly modified, and flat-bottomed craters. No gradation between smallercraters that appear relatively unmodified. these two types has been observedso far; they However, on Mars both classesshow less varia- are discussed separately here. A few ringed tion in the degree of modification. structures (6N20) and a considerablenumber of Important differences do exist between the polygonal craters (6N13) have been recognized. lunar uplands and the Martian terrains. The This latter similarity to the moon may have Martian large flat-bottomed craters are less structural implications. numerous and more highly modified than those Large flat-bottomed craters seen in Mariner observed on the moon. The intercrater areas on 6 and 7 photographs range in diameter from Mars are much smoother. Although the large about 15 km to severalhundred km. They are flat-floored craters both on Mars and on the moon 320 MuRR^¾

must have sufferedperiods of intensemodification ,,iO4 ' I ' I before the formation of the small fresh craters, the intensity or period, or both, of suchmodifica- tion on Mars, not only must have been greater, but also must have occurred principally after 6N,48,20AND22 the formation of most of the large craters. Other- 4000 wise, some large fresh craters would be easily recognizable in the Mariner pictures, just as Copernicus,Tycho, and Kepler are conspicuous on the moon. Martian cratered terrains also display a 400 z variety of positive and negative local features <• .1- other than craters. Included are sinuous channels ? and ridges visible in many B-camera frames 6 N ,'17, '19 AND 21 AVERAGED (see, for example, 6N16, 6N18, and 6N20). A 4O series of subparallel short linear markings, in some instances composing polygonal patterns similar to those of the moon, is visible in the SBC dark area, SabaeusSinus, in the northern parts < - LF of 6N19 and 6N21. , I , I Certain diagnosticlunar and terrestrial surface .,f 4 ,1o ,1oo CRATER DIAMETER D (KM) features were looked for but not recognized in the Mariner pictures. Especially conspicuousis Fig. 4. Crater abundancesof Deucalionis Regio. the absence of fresh large craters and their The cumulative crater-diameter relationship is plotted by using the numerical data of Table 1. associated rays and secondary swarms. Such The vertical axis is a logarithmic scale of crater craters, like Tycho and Copernicus,are among diameter. The range of crater diameters for small the more striking features observedon the moon. bowl-shaped and large fiat-bottomed craters, Sinuousrilles, flow fronts, and partially flooded respectively, is indicated approximately by arrows craters that characterize the lunar maria have overlapping in the range of 10-15 km. Error bars have been derived on the basis of the expected not been observed. Finally, regional terrestrial statistical uncertainty in the actual number of featuresresembling folded mountains, islandarcs, craters counted. Thus, the errors are largest for continental/oceanicplates, and rift valleys are the largest (and therefore fewest) craters counted not recognizedon Mars. in either A or B frames. Size-frequencydistribution of Martian craters. Cratered terrain on the moon traditionally has any genuine geographic variation or small been described according to the number of solar-elevation angle effect. craters per unit area as a function of diameter. It would be desirable to have size-frequency In Figure 4 the cumulative number of cratersper data separately for the large flat-bottomed and unit area larger than a given diameter is plotted small bowl-shaped populations. Unfortunately, against that diameter for the Martian area the resolution of the A frames is inadequate to Deucalionis Regio. The relevant data are listed confidently separate the two classesin the im- in Table 1. portant 5-15 kilometer range, and the areal Figure 5 shows the crater counts for the coverageof the B frames, which do possessad- individual A and B frames that are averaged in equate resolution, is insufficient to provide ad- Figure 4. Figure 5 demonstrates two aspects of equate crater counts in the same interval, as the data: (1) major geographicvariations in the shown by the error bars in Figure 4. Thus we density of large fiat-bottomed craters are not have included both populations in Figure 4 and present within Deucalionis Regio, nor are the merely note on it the approximate size ranges of apparent densities correlated with either solar- the two classes. elevation angle or A-camera filter, and (2) the MARTIAN SURFACE PROCESSES AND HISTORY statistical noise in the measurements of small bowl-shaped craters is large enough to mask The observed characteristics of the Martian SURFACE OF MARS--CRATERED TERRAINS 321

crater terrains have been compared with those of the uplands and maria of the moon. This TABLE 1. Martian Crater Abundances (I)eucalionis Regio Area) section examines the implications of this com- parisonfor Martian history and surfaceprocesses. Diameter Cumulative Review of lunar impact history. A valuable Interval, Number Cumulative Number/10 6 approach to understanding the evolution of the km Counted Number km 2 lunar surface has been the analysis of crater size- frequencydistributions. These crater populations Average of Wide-Angle Frames 6N17, 6N19, and 6N21 record the energy-frequency distributions of < 4 1 3 579 + 24 25o 4- 10.3 impacting bodies (i.e., the mass-frequencydis- 4 2- 45 7 572 + 23 247 4- 10.2 tribution for given impact velocities) as well as 4 6- 5 0 34 538 + 23 233 4- 10.0 the net effectivenessof crater-removal processes. 5 1- 5 5 46 492 + 22 213 4- 96 Analysis of surfacesof differingage can elucidate 5 6- 6 0 35 457 + 21 198 4- 93 6 1- 7 0 58 399 + 20 172 4- 86 40 4 7 1- 8 0 25 374 + 19 162 4- 84 8 1- 9 0 28 346 + 18 150 4- 8 1 9 1- 10 0 34 312 4- 17 136 4- 77 10 1- 11 0 20 292 4- 17 1 126 4- 74 11 1- 12 0 16 276 4- 166 119 4- 72 12 1- 13 0 13 263 4- 16 2 114 4- 70 13 1- 15 0 31 232 4- 15 2 100 4- 66 4000 15 1- 18 0 29 203 4- 14 2 88 4- 62 18 1- 20 22 181 4- 13 5 78 4- 58 21- 24 33 148 4- 12 2 64 4- 52 25- 30 37 111 4- 106 48 4- 45 31- 35 29 82 4- 9 1 35 4- 38 36- 40 12 70 4- 84 30 4- 36 400 -- 41- 45 13 57 4- 75 25 4- 33 4000 46- 50 16 41 4- 64 18 4- 28 ' i 51- 60 17 24 4- 49 104- 2 1 61- 70 19 54- 22 2.24- 1 70 5 0 A vetage of Narrow-A ngle Frames 6N18, 6N20, and 6N22 400 < 0 55 11 113 4- 10.6 5600 4- 530 '•. 0 56- 0 65 3 110 4- 10.5 55OO 4- 52O "•xx6N6N 2549 _ 0 66- 0 75 16 94 4- 97 4700 4- 480 0 76- 0 85 10 84 4- 92 4200 4- 450 0 86- 0 95 12 72 4- 85 3600 4- 420 0 96- 1 05 10 62 4- 79 3100 4- 390 1 06- 1 15 4 58 4- 76 2900 4- 380 40 1 16- 1 25 8 5O 4- 71 2500 4- 356 i 26- 1 4o 8 42 4- 65 2100 4- 320 1 41- 1 70 12 30 4- 55 1500 4- 270 1 71- 2 15 8 22 4- 47 1100 4- 230 2 16- 2 30 6 164- 40 800 4- 200 2314- 2 7o 2 144- 37 700 4- 190 2 71- 35 2 12 4- 35 6OO 4- 170 • '10 4OO 3.6- 45 3 94- 30 450 4- 150 CRATER DIAMETER D (KM) 4.6- 57 3 64- 25 300 4- 120 Fig. 5. Plots of crater abundances similar to 22 5.8- 78 1 54- 250 4- 11o those in Figure 4 are presented for individual 7.9- 9.9 44- 20 200 4- lOO Mariner frames. The narrow-angle B frames (top) 1 4 10- 20 2 24- lOO 4- 70 show considerablevariation in the large craters, 21- 27 1 1 14- 50 4- 50 mainly becauseof the limited number of craters; 27 i 0 it is also possible that some minor geographic variations in crater abundances are included. Neither A nor B frames show any correlation with solar elevation angle, thus ruling out a potential source of systematic error. 322 MURRAY ET AL. temporal variations in impact fluxes and in maria of different age, the form of the distri- surface processes.The size-frequency distribu- bution remains nearly unchanged[Kuiper et al., tiens of lunar craters and the impact-flux his- 1966; Trask, 1.966].For.example, Trask found tories implied for the earth-moon environment that the exponentsof the powerfunctions repre- are reviewed here as background for the dis- senting the size-frequencydistributions of craters cussionof Mars. at the Ranger 7 and 8 sites are very nearly the Figure 6 shows typical crater size-frequency same,whereas the total number of cratersin this distributions for lunar surfacesof differing age: size range (diameter >3 km) differ by about a two maria and two upland regions. The crater factor of 3. Thus through the period of the evolu- populations on the younger surfaces,the maria, tion of the lunar maria the form of the mass-fre- are less complicated by historical variation in quency distribution of impacting bodieshas been impact flux, and it is appropriateto begin the dis- constant. cussionthere. Craters of diameter larger than It is possible to predict the population of about 3 km are generally attributed to primary secondary impact craters to be expected with impacts [Shoemaker,1965]. Such craters appear this primary distribution. By studying the sharp and fresh, with raised rims, rays, ejecta secondary populations of nuclear craters and blankets, and surrounding secondaries;in the large lunar primaries, Shoemaker [1965] and larger size range, they exhibit central peaks. Brinkmann [1966] predicted that the secondary Although the magnitude of this primary crater distribution will have a greater ratio of small to distribution may vary significantly between large cratersand shouldexceed the primary popu-

•,0 5 ,

MARS COMPARED WITH MARS COMPARED WITH ' LUNAR MARIA LUNAR UPLANDS

o ,104 _ •-RANGER'•nT • MARE TRANOUIL- n.' o LITA TIS - SOUTHPOLAR ' Lu REGION

LU '1000 -- - 'z.

r• . Z

•- •00 -- RANGER•IT - • MARE LU _ TSIOLKOVSKY REGION •- CO GNI TUM '

• -

.4 4 40 '100 'I dO '•00 CRATER DIAMETER D (KM) Fig. 6. The Deucalionis Regio data of Figure 4 are compared with crater abundances of the lunar maria (left) and the uplands (right). The Ranger 7 and 8 data are from Trask [1966].The uplands data were compiledfor this paper from Orbiter 4 frame 88, taken in the vicinity of the south pole, and Orbiter 3 frame 121 from the lunar backside near Tsiolkovsky, chosenbecause they appear to be free of effects of Mare formation. 8UIIFACE OF MAIls--CIIATERED TEIIItAINS 323 lation for craters of diameter smaller than about the moon. (We shall ignorethe possiblecomplex- 1 kin. The steep part of the maria curves, for ities ill the actual events that may have con- craters smaller than 3 kin, is compatible with stituted the terminal phasesof accretionand the these predictions (Figure 6). The distribution immediate period thereafter. We presume a of large telescopically observable secondaries smooth and ral•id decline of cratering rates at supports this hypothesis [Shoemaker, 1965]. the end of accretion of both the moon and Mars.) Below diameters of about 100 meters the rate To return to the significanceof the Martian of increaseof lunar crater density with decreasing crater abundances, a comparison of size-fre- diameter falls to a lower value than expected for quency distributions of craters in Deucalionis a secondarypopulation. Craters in this size class Regio and in the lunar maria and uplandsis also rangefrom fresh and pristine to nearly obliterated presentedin Figure 6. Martian cratered terrain shallow irregular depressions.This distribution and the lunar uplandsboth displaylarge subdued of morphologies has led observers to conclude craters and small bowl-shaped craters. The over- that these smaller craters form a steady-state all forms of the distributions are similar. How- population in which craters are formed and ever, the size-frequencydistribution of the Mar- destroyed at the same rate [Moore, 1964; Shoe- tian small bowl-shaped craters is significantly maker, 1965]. Soderblom[1970] has shown that different from that of comparably sized craters the relationship between the steady-state popu- on the moon. The implications of these similar- lation and the secondary production curve can ities and differences are considered in the next be derived analytically from a model of down- two sections. slope transport of material by impacts that Age of largefiat-bottomed Martian craters. The produce negligible individual topographic age of large fiat-bottomed Martian craters can changes. be estimatedby extrapolationof lunar cratering In summary, the various segmentsof the crater history. Lunar inpact fluxescan be extrapolated frequency curves for craters in the lunar maria to Mars on some basis, and the Martian surface are related in recognizable ways. Primaries larger ages obtained by scaling with known surface than a few kilometers produce abundant secon- ageson the moon. This was first attempted by daries smaller than 1 kin. The erosive action of Anders and Arnold [1965]. A more complete small secondaries (diameter •10 meters)on analysis was performed by Hartmann [1966] other large secondariesproduces a steady popu- using Mariner 4 crater counts. Hartmann's lation in which craters are generated and ob- analysis indicated that these Martian craters are literated at the same rate. at least 4 d- 2 b.y. in age. We argue here that Interpretation of the size-frequency relation- Hartmann's result is a significant underestimate ship of craters in the lunar uplands is complicated of the age of thesecraters when reviewed in light by superpositionof the later impact history re- of current information. corded in the lunar maria on earlier cratering, Hartmann assumed that both the lunar and the which produced the very numerouslarge highly- Martian impact fluxes were entirely asteroidal modified craters. Under the assumption of con- and used a previous estimate by Anders [1964] stant flux rates throughout the history of the of asteroidal fluxes at Mars higher by a factor of moon, the relative densities of large craters 25. This factor of 25 is probably grossly over- (% 20 km in diameter) on the ,•4-b.y.-old mare estimated, as indicated by a variety of reasons. surfacein Mare Tranquilitatis [Albeeet al., 1970; First, Wetherill [1968] has shown by dynamical Silver, 1970] and in the lunar uplands would arguments that the probability of Mars being imply an age for the uplands greater than 50 b.y. impacted by an asteroidal object whose orbit This unrealistic age estimate, coupled with the was perturbed by Mars is very low. This con- observed severe modification of the large upland sideration significantly reduces the figure of a craters leads to two conclusions:(1) the dense factor of 25. Second, the assumptionthat the population of large craters in the lunar uplands impacting debris would preferentially impact records a period during which the cratering rate Mars rather than the moon by a factor of 25 is was orders of magnitude greater than has oc- not consistent with meteorite observations. curred since the formation of the maria, and (2) Studies of meteorite orbits from observations of the uplands represent the accretional surface of meteor trails have shown that most of the debris 324 •URRAY ET AL.

impacting;the earth at present is in highly eccen- with or postdating the formation of these large tric orbits with aphelion distances of 4 to 5 AU craters. Both showa collectionof younger,fresher [Wetherill, 1969]. Objects in such eccentric orbits craters that have accumulated since the end of would have almost equal likelihood of striking major surfacemodification. Thus, not only are the Mars and the moon. Furthermore, recent evi- Martian crateredterrains unquestionablyancient, dence of decreasing impact fluxes during the but they mimic to some extent the sequenceof period of lunar formation (L. A. Soderblom and events recorded on the lunar uplands. We feel it is L. A. Lebofsky, unpublisheddata, 1970) suggests probable, therefore, that both the Martian cra- that a significant fraction of the lunar crater tered terrain and the lunar uplands record the population was produced by cometary impact, final stagesof planetary accretion. or at least by a sourceother than asteroidal. An Differencesin post-accretionalhistories. Major asteroidal impact flux would tend to increase modification of large fiat-bottomed craters on with time [Kuiper, 1950; •pik, 1951]. Further, both the moon and Mars must have occurred be- as will be shown in the next section, the distri- fore formation of the present relatively unmod- butions of objects forming the present small ified small crater populations on those surfaces. craters (diameter • 10 km) on Mars and on the Small bowl-shaped craters on Mars appear com- moon have been differefit. Hence this factor of parable in degreeof modificationto cratersof the 25 must be a significant over-estimate, since samesize (diameters 1-10 km) on the lunar maria. impact fluxes on Mars and on the moon cannot A comparison of these two distributions should both have been entirely asteroidal. The true expose differences and similarities in the size ratio may be closer to unity, since objects in spectra of impacting objects that have created cometary orbits have almost equal probability them. of striking the earth and Mars. Finally, the The size-frequencydistribution of primaries large fiat-bottomed Martian craters have been in the lunar maria can be written severelymodified, and it is likely that many are N = AD -1'7 no longer rocognizablein the Mariner pictures. Hence, crater counts are on the low side com- where N is the number of craters per unit area pared to counts on the moon. All these arguments with diameters larger than D and A is a constant. tend to increase the Hartmann-model age esti- Thus, for example, the ratio of the numbers of mate to something far exceeding the age of craters larger than 3 km to the number larger the solar system. This, of course, is under an than 30 km for the lunar maria is about 50. assumption of constant flux. The implication is, From Figure 4, the number of small bowl-shaped then, that these large fiat-bottomed craters on craters larger than 3 km in each A frame (area Mars were formed, as were those in the lunar •0.6 X 106 km '•) is about 400. Hence, if the uplands, during the early history of the planet proportion of 3- to 30-km Martian bowl-shaped under very great fluxes. As in the case of the craters were the same as the proportion on the moon, we associate such conditions with the moon, one would expect to see about 8 fresh final phasesof planetary accretion. bowl-shaped craters larger than 30 km in each A second set of arguments also suggeststhat A frame. Or, stated differently, 1 in 5 of the the large fiat-bottomed craters on the moon and large fiat-bottomed craters in Deucalionis Regio Mars record analogousstages of planetary evo- that are larger than 30 km in diameter should lution. The sequenceof early eventson Mars and be as fresh and sharp as the small bowl-shaped on the moon were evidently similar. Both sur- craters and should appear similar to Kepler and faces show an old, now-subdued population of Copernicuson the moon. Of the approximately numerous large craters that are ancient by any 100 large fiat-floored craters in DeucalionisRegio scaling, although the Martian population has with diameter greaterthan 30 km, noneresembles been more severely modified and no longer dis- the large youthful lunar craters. plays the saturation that is seen in the lunar In order to explain this evident deficiency,one uplands. The size-frequencydistributions of the might presume that large Copernicus-likecraters lunar and Martian large craters are similar in on Mars rapidly assume the morphology of both amplitude and form. Both show evidenceof large fiat-floored craters. It is indeed conceivable major epochs of surface modification concurrent that the floors could quickly becomefiat. It is S•TRr^c• or M^Rs--C•^T•D T•^•s 325

difficult, however, to imagine how rims, ejecta averages over abundant comets and later aster- blankets, and secondariescould be obliterated oids and Mars averagesonly over the later aster- without destroying the presently observedsmall oids). Both kinds of effects may be responsible bowl-shapedcraters. It doesnot seem likely that for the differences in the observed distributions. the large craters of a contemporaneouspopu- The absence of Copernicus-like craters in lation could be severly altered without even more Deucalionis Regio indicates still another impor- severly modifying the smaller ones. Thus, the tant difference between lunar and Martian his- distribution of impacting bodies that formed the tory. Copernicus,and Kepler, for example, pre- presently observedsmall bowl-shapedcraters on sumably were formed on the moon by either Mars is deficient, by comparison to the lunar cometary or asteroidal impact. If cometary, ap- distribution, in objects that would producecraters proximately the same number of such craters like Copernicus and Kepler. That is to say, the shouldhave been formed on Mars as on the moon, postulated Copernicus-type craters and the since comets are typically in highly eccentric presentsmall bowl-shapedcraters are not part of orbits. If Copernicusand Kepler are of asteroidal the same population (not that there never were origin, there may have been a greater number of any Copernicus-typeevents). suchimpacts on Mars than on the moon [Anders Another observation supporting the conclusion and Arnold, 1965; Witting et al., 1965; Baldwin, that recent Martian fluxes recordedby the small 1965]. Either way, such craters must have bowl-shaped craters were deficient in objects formed on Mars in at least as great a number as capable of forming a Copernicus-typecrater is are now evident in the lunar maria and at about the lack of associatedsecondary craters. If the the sametime. Thus, 10-20 of the approximately Martian primary distribution were similar to the 250 large flat-bottomed craters in frames 6N17, lunar one, there would be a secondary popu- 6N19, and 6N21 must be no older than Mare lation much like that represented by the steep Tranquilitatis (•4 X 10• yrs). Yet all appear to sections of the crater-frequency relations for have suffered about the same degree of modifi- Mare Tranquilitatis and Oceanus Procellarum cation, or we should recognizethe younger frac- (Figure 6). The secondarydistribution produced tion easily. Hence major modification of the by the observedsmall bowl-shapedcraters, how- large Martian craters must have occurredafter ever, shouldexceed the density of primary craters the period of mare formation on the moon, i.e., only for crater diameters less than 100 meters within the last 3 to 4 b.y. which is below the limit of resolution in Mariner Martian crater-modification processes. Mar- 1969 pictures. The absenceof not only Coper- tian crater morphologiesreflect crater-modifi- nicus-sized fresh Martian craters, but of their cation processes;therefore, they can supply associated secondaries as well, reinforces the insight into the nature of those processes.Small important conclusion that the distributions of bowl-shapedcraters on Mars all display about impacting objects that produced the present the samedegree of preservation.Those observed Martian small bowl-shaped craters and the post- appear fresh, having sufferedonly minor mod- maria lunar craters have been distinctly dif- ification,if any. However,any earliergenerations ferent. of Martian bowl-shaped craters have all been Two possibleexplanations of these differences modified beyond recognition. Two possible in accumulated impacts can be suggested.First, explanationsfor these relationshipsmerit exam- the populations of impacting bodies might be ination: (1) episodicsurges of crater formation composedof several families of objects (e.g., as- coupled with continuousmodification, and (2) teroids and comets) present in different relative episodesof catastrophic crater removal super- abundancesat Mars and at the moon. A second, imposed on an approximately continuous rate alternative possibility is that the ratio of com- of crater formation. It is difficult from present etary to asteroidal impacts has not been sig- information to evaluate confidently the relative nificantly different in the vicinities of Mars and importance of these two possibilities.However, the moon at any particular time but that this the first appearsto us lesslikely for two reasons. ratio has been changing over time. Thus the two Studies of the chemical nature and cosmic-ray surfaces now record accumulations over different exposure ages of terrestrial meteorite falls time scales (this could result, e.g., in moon [Anders,1964] suggestthat as many as 5 to 20 326 Mmm^¾ ET AL.

fainilies of meteoritic fragments have been pro- would probably eventually obliterate the large ducedby collisionsin the asteroidbelt in the last craters as well. Infrequent episodes of crater few hundred million years. Theoretical analyses removal, however, might serve to remove large of the lifetime of thesefragments [Wetherill, 1967s numbers of small craters without wearing down Hartmann, 1968] indicate that the half-life for the la.rge fiat-bottomed craters significantly. such families is of the order of l0 s to 10 9 years. Finally, in view of the unusual and recent surface Thus the lifetimes are much longer than the modificationprocesses implied by featurelessand intervals between the production events. Itcrice chaotic terrains (see paper 2), we prefer to place it seemsunlikely that such surgesof impacts of the emphasis of explanation on hypothetical asteroidal objects have occurred on Mars (al- episodesof crater removal rather than on hypo- though this restriction may not apply to cometary thetical episodesof formation. objects). Overlap of differing families would lead The nature of Martian surface processescan to a distribution in the state of preservation of be further evaluated in the light of another small bowl-shaped craters. Secondly, a contin- significant observation. As illustrated in Figure uous erosionprocess capable of rapidly removing 7a, intercrater areas in Martian cratered terrains traces of all preexisting small bowl-shaped craters are much smoother than in the lunar uplands

Fig. 7a. Lunar simulation of Mariner A frame. To illustrate the morphology of the lunar uplands at the same surfacescale of the Mariner 6 and 7 wide-angleframes, a portion of medium-resolution frame 130 of the Lunar Orbiter Mission 4 has been enlarged so that the scale of the left,- hand sideof the frame approximatesthat of the accompanyingMariner frames 6N17, 6N19, and 6N21 (upper right,,lower left, andlower right, respectively).Because of the smallerradius and therefore greater curvature, of the moon, complete simulation of wide-angle frames is not possible. The solarelevation angle in the lunar frame rangesfrom 50ø to 14ø , thereby coveringmost of the rangeof the three Mariner frames.The contrastof the lunar frame has beenartificially modified to resemblethat of the Mariner frames.The lunar area is in the vicinity of the south pole, an upland region of the moon that has not, been severely affected by adjacent,mare formation and filling. SURFACE OF MARS--CRATERED TERRAINS 327

Fig. 7b. Lunar simulation of Mariner B frame. A similar procedure to that of Figure 7a has been carried out to match the surfacescale of Mariner narrow-angle frames 6N18, 6N20, and 6N22. The solar elevation angle of the lunar area is 22 ø and those of the Mariner frames are 38 ø, 26 ø, and 14ø, respectively. The lunar simulation is taken from Orbiter 4 high-resolution frame 107 (longitude 6ø, latitude -42ø).

Both are thought by us to represent original been fundamentally different from those oper- accretionary surfaces, but the Martian surface ative on the moon. has undergonea much greater degreeof leveling A natural inclination to call on solely aeolian and horizontal redistribution of material. This processesto effect a regional redistribution of processhas causedthe disappearanceof elevated materials on Mars is complicatedby the following crater rims and ejecta sheets,yet has permitted considerations.Very effective physical or chem- the survival of what is apparently primary relief ical weathering of consolidated rock is required along crater walls (see 6N18, Figure 7b). Local to reduce original relief of hundredsof meters or processesmodifying lunar terrains •re impact kilometers completely into dust and sand-sized fragmentation and downslope transport by particles that could then be effectively trans- sliding, rolling, and impact ejecta, These proc- ported by the Martian winds. By comparison, cessestend to smooth out local topographic the dominant fragmenting processon the moon irregularities.If the Martian modificationprocess results from impact of high-velocity particles. of redistributingmaterial dependedprincipally As pointed out previously [Leiqhtonet al., 1967], on local impact events,as on the moon, it would the effectivenessof this processon Mars would seem that local relief along crater walls should be greatly reduced by the thin atmosphere. have been smoothed out before the intercrater Inasmuch as liquid water cannot exist in a free areas were leveled. Hence, the processeswhich state on Mars owing to the low surfacepressures have modified the surfaceof Mars may have [Inqer8oll, 1970], and that thermal shattering is 328 MURRAY ET AL. most likely ineffective [Ryan, 1962], the mech- Kuiper, 1952, pp. 3, 343-444; Shirk et al., 1965; anism of effective weathering on Mars remains National Academy o[ Sciences, 1966a, 1966b; unknown. Shklovskii and Sagan, 1966]. The pictures from Episodic crater obliteration might be related Mariners 4, 6, and 7 now provide a concrete to occasionalbut extreme changesin the nature framework for Martian surface history. The of the Martian atmosphere. Thus the Martian survival of large Martian craters for, at the very terrain, like that of much of the western United least, severalbillion years, and indeedmost prob- States where the odd thunderstorm is the major ably from the end of planetary accretion, places eroding event, may display the effects of ero- strong constraintson the surfacehistory of Mars sional processesthat are inactive during average relevant to the possibilityof any earthlike period. conditions. Unlike earth, neither crustal deformation nor Summary. Comparison of regional associ- atmospheric erosion has been sufficient to ations, morphologies, and size-frequency dis- destroy these ancient Martian topographic tributions of craters on Mars and on the moon features. Thus there is, for Mars as well as suggests the following conclusions concerning for the lunar uplands, the implication of a Martian surfacehistory and processes: stable crust and the absence of an earthlike 1. The processof mare formation that has aqueousatmosphere with great erosionalcapabil- ities. generated unsaturated and relatively unmod- ifiedcratered te)rain on the moon subsequent to It is difficult to be precise as to exactly how accretion is not evident on Mars. 'lunarlike' or 'non-earthlike' Mars must be in 2. Many of the large fiat-bottomed craters order to satisfy present observations. At the on Mars probably have survived since the final very least, it can be argued that the most pro- stagesof accretion of that planet. bableearly history of the surfaceof Mars is much like that of the moon. Two kinds of evidence from 3. The large craters on both Mars and the moon record intense modification. However, the Mariner 6 and 7 picture analysis support this lack of large fresh Martian craters, comparable conclusion. The first pertains to the magnitude to Kepler and Copernicuson the moon, implies of and responseto crustal deformation. Like the that the substantial large-crater modification on lunar uplands, the Martian cratered terrain has Mars took place during the last 3-4 b.y. after been exceedinglystable for cons, undisturbed by major modification had ended on the moon. earthlike tectonic activity. Further, the struc- However, no significant modification process tural properties of the lunar and Martian crusts appears to have been operative during the inter- must be similar to account for the corresponding val of the accumulation of the present small development of polygonalization of large craters bowl-shapedcraters. on both surfaces. The second kind of evidence 4. The average distributions of impacting supporting a 'lunarlike' Mars is the strikingly bodies that formed the currently visible small similar sequence of early events recorded on craters on Mars and of the lunar maria have been both bodies. The Martian cratered terrain and different. the lunar uplands preserveremnants of old crater 5. Modification of the surface on Mars has populationsthat are similar in morphology,areal been accompaniedby much greater horizontal density, and size-frequency distribution, and redistribution of material than has occurred on that are highly modified. Both surfacesrecord a the moon. stage of initial high-impact flux followed by intense surface modification that grossly de- COMPARISON WITH THE SURFACE EVOLUTION graded craters and ceased before the accu- OF EARTH AND Moon mulation of most of the presently visible small There has been a traditional interest in Mars as bowl-shaped craters. These similarities strongly a possibleabode of simple life forms. For many suggest that the near-surface crustal environ- decadespreceding spacecraft investigation, Mars ment of Mars has been lunarlike since its ear- was viewed as possibly having experiencedan liest history. There is no present evidenceof any earthlike phase with aqueous atmosphere and terrestrial phase or processes.Thus, there is possibly primitive oceans in which simple life little more basisfor postulationof ancient oceans forms might have evolved [de Vaucouleurs,1950; on Mars than on the moon. Brief earthlike pc- SURFACE OF MARS--CRATERED TERRAINS 329 riods cannot be rigorously ruled out for either Institute of Technology. Cutts has been partly body, but suchevents are lesslikely for moon and supported by NASA-105-69836 and Soderblom by NGL-05-002-003. Mars than for Venus, or perhaps even Mercury whose surface history is not so constrained by REFERENCES observations as that of Mars. Albee, A. L., 1). S. Burnett, A. A. Chodos, O. J. The importance of this kind of comparative Eugster, J. C. Huneke, I). A. Papanastassiou, planetology argument lies in its implication for F. A. Podosek, G. Price Russ II, H. C. Sanz, planetary exploration generally. The reduced F. Tera, and G. J. Wasserburg, Ages, irradiation probability of Martian oceans, implied by the history, and chemical compositionof lunar rocks from the Sea of Tranquillity, Science,167, 463- presence of primordial cratered terrains there, 466, 1970. surely means a significant reduction in the prob- Anders, E., Origin, age, and composition of mete- ability that life could have originated on Mars orites, Space Sci. Rev., 3, 583-714, 1964. in an earthlike manner. The basis for emphasis Anders, E., and J. Arnold, Age of craters on Mars, on Mars as the prime target in the search for Science, 149, 1494-1496, 1965. Baldwin, R. B., Mars: An estimate of the age of extraterrestrial life seemsto be weakenedby the its surface, Science,149, 1498-1499, 1965. strong resemblanceof its surface to the uplands Brinkmann, R. T., Lunar crater distributions from of the moon. Rather, Mars should be accorded Ranger VI photographs, J. Geophys. Res., 71, priority geologically as exhibiting important 340-342, 1966. Cutts, J. A., R. P. Sharp, B. A. Smith, B.C. Murray, aspects of planetary evolution unknown else- and L. A. Soderblom, The surface of Mars, 3, where in this solar system, such as the replace- Light and dark markings, J. Geophys.Res., 76, ment of cratered terrains by younger uncratered this issue, 1971. terrains, as will be discussedin paper 2. Only on de Vaucouleurs, G., The Planet Mars, pp. 71- Mars can there be found a carbon dioxide frost 72, Faber and Faber, London, 1950. Dunne, J. A., W. D. Stromberg, R. Ruiz, A. Collins, cap with all its associated phenomena. There and T. E. Thorpe, Maximum discriminability are erosionalprocesses and modification episodes versions of the near-encounter frames, J. Geophys. recorded on the Martian surface unlike those Res., 76, this issue, 1971. yet known on any other body. The unraveling Goldstein, R. M., W. G. Melbourne, G. A. Morris, G. S. Downs, and D. A. O'Handley, Preliminary of these characteristically Martian processesand radar results of Mars, Radio Sci., 5, 475-478, the associatedsurface history provides a com- 1970. pelling and rewarding intellectual focus for Hartmann, W. K., Martian cratering, Icarus, 5, explorationof that planet. 565-576, 1966. Hartmann, W. K., and A. C. Hartmann, Asteroid Acknowledgments. We are deeply indebted to collisions,Icarus, 8, 361-381, 1968. all personswhose combined efforts made the Mariner Herr, K. C., D. Horn, J. M. McAfee, and G. C. 1969 flights to Mars a success. With respect to Pimentel, Martian topography from Mariner 6 the series of four articles on Martian surface features and 7 infrared spectra, Astrophys. J., 75(8), published herein, we specifically acknowledge the 883-894, 1970. valuable aid of the following: G. E. Danielson, Ingersoll, A. P., Mars: Occurrence of liquid water, S. A. Collins, J. J. van der Woude, T. C. Rind- Science,168, 972-973, 1970. fieisch, J. A. Dunne, R. C. Dewar, and Patricia Kuiper, G. P., On the origin of asteroids, Astronom. Conklin, all of the California Institute of Tech- J., 55, 164, 1950. nology and JPL. Our colleagues of the Mariner Kuiper, G. P., The Atmospheresof the Earth and TV team, M. E. Davies, A. H. Herriman, N.H. Planets, pp. 343-344 and 404, University of Horowitz, C. B. Leovy, B. A. Smith, and A. T. Chicago Press, 1952. Young have provided counsel and information. Kuiper, G. P., R. G. Strom, and R. S. LePoole, Without the leadership and unending efforts of Interpretation of the Ranger records, in Ranger 8 Principal Investigator R. B. Leighton, the TV and 9, 2, Experimenters' Analyses and Interpre- project would never have succeeded. Robert A. tations, Jet Propul. Lab. Tech. Rep. 32-800, 35- West of the California Institute of Technology 248, 1966. assistedin preparation of crater counts. The present Leighton, R. B., B.C. Murray, R. P. Sharp, J. D. paper has been substantially revised from its Allen, and R. K. Sloan, Mariner 4 photography original form to take into account the careful and of Mars: Initial results, Science, 1J9, 627-630, constructive criticism received from a number of 1965. reviewers, and we wish to acknowledge their Leighton, R. B., B.C. Murray, R. P. Sharp, J. D. contribution. Allen, and R. K. Sloan, Mariner 4 pictures of The participation of Murray, Leighton, and Mars, Jet Propul. Lab. Tech. Rep. 32-884, 1967. Sharp has been underwritten by the California Leighton, R. B., N.H. Horowitz, B.C. Murray, 330 Mvaa^¾

R. P. Sharp, A. H. Herriman, A. T. Young, B. A. 4, The south polar cap, J. Geophys.Res., 76, this Smith, M. E. Davies, and C. B. Leovy, Mariner 6 issue, 197 lb. television pictures: First report, Science, 165, Shirk, J. S., W. A. Hazeltine, and G. C. Pimentel, 684-690, 1969a. Sinton bands: Evidence of deuterated water on Leighton, R. B., N.H. Horowitz, B.C. Murray, Mars, Science,1J7, 48-49, 1965. R. P. Sharp, A. H. Herriman, A. T. Young, B. A. Shklovskii, I. S., and C. Sagan, Intelligent Life Smith, M. E. Davies, and C. B. Leovy, Mariner 7 in theUniverse, p. 270, Holden-Da y, San Francisco, television pictures: First report, Science, 165, Calif., 1966. 787-795, 1969b. Shoemaker, E. M., Preliminary analysisof the fine Leighton, R. B., N.H. Horowitz, B.C. Murray, structure of the lunar surface in Mare Cognitum, R. P. Sharp, A. H. Herriman, A. T. Young, B. A. in The Nature of the Lunar Surface, pp. 23-78. Smith, M. E. Davies, and C. B. Leovy, Mariner 6 Johns Hopkins Press, Baltimore, Md., 1965. and 7 television pictures: Preliminary analysis, Silver, L. T., Uranium-thorium-lead isotopes in Science, 166, 49-67, 1969c. some Tranquility Base samples and their impli- Leighton, R. B., and B.C. Murray, One year's cations for lunar history, in Proceedingsof the processing and interpretation--An overview, Apollo 11 Lunar Science Conference,2, Chemical J. Geophys.Res., 76, this issue, 1971. and Isotope Analysis, Geochim.Cosmochim. Acta, Lincoln Laboratory, Radar Studies of Mars, Final Suppl. l, 1533, 1970. Rep., Contract NAS 9-7830, 15 January 1970. Soderblom, L. A., A model for small-impact erosion Moore, H. J., Density of small craters on the lunar applied to the lunar surface, J. Geophys.Res., 75, surface, in AstrogeologicStudies Annual Progress 2655-2661, 1970. Report, August 1962-July 1963, part D, pp. Trask, N.J., Size and spatial distributions of 34-51, U.S. Geological Survey, Washington, craters estimated from Ranger photographs, in 1). C., 1964. Ranger 8 and 9, 2, Experimenters' Analyses and National Academy 'of Sciences, Biology and the Interpretations, Jet Propul. Lab. Tech. Rep. Exploration of Mars, pp. 8 and 21, Nat. Acad. 32-800, 249-338, 1966. Sci., Nat. Res. Council Publ. 1296, 1966a. Wetherill, G. W., Collisions in the asteroid belt, National Academy of Sciences, Space Research: J. Geophys.Res., 72, 2429-2444, 1967. Directions for the Future, pp. 18 and 96, Nat. Wetherill, G. W., Dynamical studies of asteroidal Acad. Sci. Nat. Res. Council Publ. 1J05, 1966b. and cometary orbits and their relation to the •pik, E. J., Collisionprobabilities with the planets origin of meteorites, in Origin and Distribution and the distribution of interplanetary matter, of the Elements, edited by L. H. Ahrens, pp. Proc. Irish Acad., 5J, Sec. A, 1951. 423-443, Pergamon, New York, 1968. Rindfieisch, T. C., J. A. Dunne, H. J. Frieden, Wetherill, G. W., Origin of prairie network fireballs W. D. Stromberg, and R. Ruiz, Digital processing and meteorites (abstract), Trans. AGU, 50, of the Mariner 6 and 7 pictures, J. Geophys.Res., 224, 1969. 76, this issue, 1971. Witting, J. F., F. Narin, and C. A. Stone, Mars: Ryan, J. A., The case against thermal fracturing Age of its craters, Science,149, 1496-1498, 1965. at the lunar surface, J. Geophys.Res., 67, 2549- Young, A. T., High-resolution photomerry of a 2558, 1962. thin planetary atmosphere, Icarus, 11, 1-23, Sharp, R. P., L. A. Soderblom, B.C. Murray, and 1969. J. A. Cutts, The surface of Mars, 2, Uncratered terrains, J. Geophys.Res., 76, this issue, 1971a. Sharp, R. P., B.C. Murray, R. B. Leighton, L. A. (Received July 31, 1970; Soderblom, and J. A. Cutts, The surface of Mars, revised September 17, 1970.)