NASA/CR -.-- j 207773 Planet. Space Sci., Vol. 45, No. 10,pp. 1271-1277, 1997 _) Pergamon _, 1997Elsevier Science Ltd All rights reserved. Printed in Great Britain 0032-0633/97 $17.00+0.00 PIh S0032-0633(97)00075-5 / 7- Comet Shoemaker-Levy 9: an active comet Terrence W. Rettig _and Joseph M. Hahn 2 1Department of Physics, University of Notre Dame, Notre Dame, IN 46556, U.S.A. :Lunar and Planetary Institute, 3600 Bay Area Boulevard, Houston, TX 77058-I 113, U.S.A. Received 30 September 1996;accepted 21March 1997 sion that was triggered by the comet's tidal breakup in Abstract. The important elements of the debate over July 1992. Arguments for and against the comet's activity the activfiy versus dormancy of comet Shoemaker- are described herein, and comparisons of observations of Levy 9 (g:E_are reviewed_ It isargued that the the S L 9 fragment K to dust comae models will show "circularity of the isophotes in the inner comae of S L that it was indeed an active, dust-producing comet. Esti- _ell as the _a]-_d-epe?i- en_s of the comae mates of its dust size distribution, mass loss rates, and Qarightness profiles are indicators of sustained dust pro- dust outflow velocities are presented. duction by S L 9. It is also shown that the westward LLail orientations, which were formerly interpreted as a _--Ngff"of the comet's dormancy, are not a good indicator _1 either activity or dormancy. Rather, the tail orien, The activity of comet Shoemaker-Levy 9 "-Igtions simply place constraints on the dust production /rj!__tefor grains smaller than _ 5ym. All the availabcle Determining whether comet S-L 9was active or dormant _idencc points to S L9as having been an active, dust-- is necessary in order to successfully interpret the dust ;ipr°ducing comet. Synthetic images of an active comet- observations as well as to correctly infer the dusty-gas ±_r¢ fitted to Hubble Space Telescope images of t_l_-? dynamical processes that may have occurred on its com- S-L 9 fragment K, and its grain size and outflow velocity etary surfaces. In fact, the distinction between activity -izIisfi'qb-titions are extractediTl_seTq_diggs show that versus dormancy alters the inferred mass of the S L 9dust the appearance of the dust coma was dominated by grains by several orders of magnitude. According to the Iarge grains having radii between _ 30i_mand _ 3mm_ dormant-comet hypothesis, if most of the observed S-L 9 produced at a rate of M._22kgs -_, and ejected at-; dust had been emitted by the comet fragments during the outflow velocities of _0.5m s-_. Only upper limits o!1 months just following the 1992 tidal breakup event and the production rates of smaller grains are obtained.. that S L 9 had been relatively inactive ever since (e.g. The nu_le-u_ _f fragment K was not o_er-ve-d-d]-rectiy Sekanina et aI., 1994; Sekanina, 1996a), then one must but _ts s_ze _srestricted to lie within a rather narrow conclude that the surviving comae grains observed one to interval 0.4 < R_< 1.2kin. _ 1997 Elsevier Science Ltd, two years later were large, having sizes exceeding about Icm. Initially, much smaller grains may once have been present but they had since been swept from the fragments' comae by solar radiation pressure. However if the S L 9 Ever since the discovery of comet Shoemaker Levy 9 fragments were instead continuously emitting dust, then (S L 9) in the spring of 1993 there has been an ongoing much smaller grains could have steadily replenished the debate regarding its activity (Sekanina et al., 1994; Cher- comae as they continually drifted down the dust tails due nova et al., 1996; Hahn et al., 1996; Rettig et al., 1996b; to radiation pressure (e.g. Hahn et al., 1996). Sekanina, 1996a). This point is of no minor consequence, Regardless of whether the comet fragments were active for in order to correctly understand the 1993-1994 obser- or dormant, there could not have been a large contribution vations of the S L9 dust comae and tails it is necessary to to the observed comae optical depth by grains much smal- know whether one was viewing recent and possibly ongo- ler than a few microns (e.g. Sekanina et at., 1994; Hahn ing dust emission, or perhaps more "ancient" dust emis- et at., 1996; Sekanina, 1996a). This fact may be inferred from the observed dust tails' orientations. As cometary dust grains recede anti-sunward due to radiation pressure, keplerian shear causes the grains to drift in the direction Correspondence to" T. W. Rettig opposite of the comet's velocity vector. In heliocentric 1272 T. W. Rettig and J. M. Hahn :Comet Shoemaker Levy 9:an active comet Eastwa_rdWestward the fragments when viewed either before or after solar opposition (see Fig. 1). Comet S -L9was observed through two successive solar oppositions and no eastward dust features were detected (Sekanina et al., 1994 ;Chernova et al., 1996). Prior studies ofS-L 9 often imply or conclude that the fragments were relatively dormant due to the absence of any eastward dust features (e.g. Sekanina et al., 1994; Chernova et al., 1996; Sekanina, 1996a). However this conclusion is Jupitei[r premature since the observed dust tail orientations were consistent with both the active and the dormant comet hypotheses. The westward tail orientations observed after opposition instead provide an upper limit on the rate at which an S L 9fragment could have emitted grains smal- ler than a few microns, which is about 0.5kgs -_ as esti- mated by Sekanina (I996a) based on HST detection limits. Frequently, cometary dust emission occurs from a few \ discrete spots on the surface a comet nucleus which, if rotating, can produce dust streams and spirals sometimes "May seen propagating through a dust coma. It should be noted that comet S L 9 exhibited rather featureless dust comae which have also been interpreted as an indicator of the fragment's inacth, ity (e.g. Sekanina, 1996a). However, a July March 31 14 featureless dust coma is not a strict indicator of inactivity, Jan25 for it is also consistent with dust emission that was more evenly distributed across a fragment's sunlit surface. This Fig. I. Lines of sight from Earth @ to S L 9 are shown for is a reasonable possibility considering the S L 9fragments selected HST observations acquired in 1994. A continuous had effectively been stripped of any ancient surface mantle stream of dust grains larger than about 5j_mcontribute curved during the tidal disruption event which might otherwise tails that always appear west of the fragment (dark grey curve), have localized dust production to discrete spots. whereas smaller grains contribute straighter tails that appear There are, however, two lines of evidence that suggest east of the fragment (light grey curve) when observed after solar opposition. Dust tails are not drawn to scale S L 9 had been actively replenishing its dust comae. The first is that contour maps of the light distribution in the fragments' innermost _ 1"comae regions remained quite circular throughout most of the comet's orbit 2(Weaver et space, comet S L 9 was moving roughly eastward at aI., 1994, 1995 ;Hahn et al., 1996 ;Rettig et al., 1996b). It approximately Jupiter's orbital velocity, _ so keplerian has been noted that if S L 9 had been dormant and its shear was responsible for bending the dust tails west of comae had consisted of large grains simply co-orbiting the anti-solar direction. However the orientation of the with the fragments, then the inner comae contours should tails projected onto an Earth-observer's sky plane have become progressively elongated along the fragment depended upon the degree of the taws curvature, and this train axis as the fragment train itself lengthen with time was governed by the dust grains' size. Smaller grains that (Weaver et al., 1995 ;Hahn et al., 1996 ;Weissman, 1996). were rapidly accelerating anti-sunward gave rise to In contrast, a comet that continuously replenishes its dust straighter tails, whereas large and slowly drifting grains comae will maintain circular isophotes in its inner coma, contributed shorter tails having greater curvature. A sche- as was exhibited by S- L 9 throughout most of its orbit. matic diagram of the Sun-comet Earth viewing geometry A second line of evidence that favors activity is the for HST observations acquired in 1994 is sketched Fig. 1. comae surface brightness profiles. An idealized comet This diagram indicates that the dust tails should always experiencing steady and isotropic dust emission will have appeared west of the fragments when viewed prior develop a dust coma having a column density varying as to solar opposition. But after solar opposition, if there p-_ with projected distance p from the coma photocenter. were detectable quantities of grains smaller than a critical When the effects of radiation pressure acting upon a dis- size, the dust tails would have rotated 180_from west to tribution of grains are considered, the brightness profile east on the sky plane. The critical grain size is about 5lLm along the tail still varies as p-_ (generalizing the results for S L 9 (Hahn et al., 1996). It is important to realize of Wallace and Miller (1958)) whereas the azimuthally that grains larger than this threshold would have con- averaged brightness profile will develop a p-3,,2 power law tributed tails that would always have appeared west of for sufficiently large p (Jewitt and Meech, 1987). Bright- ness profiles for fragment K, given in Fig. 2, clearly evi- dence such phenomena. All of the bright, on-axis frag- _Except just prior to impact, the velocity of S-L 9 relative to Jupiter was small compared to the orbital velocity of the Jup- iter/S L 9system about the Sun. Jupiter's gravity did not play a significant role in determining the appearance of the S L 9 :The exception, of course, occurred during the month just prior comae and tails until about amonth before impact (Hahn et al., to impact as the comae and tails become progressively elongated 1996). along the comet-Jupiter direction. T.W.RettigandJ.M.Hahn:Comet Shoemaker Levy 9:an active comet 1273 to the sequence of S L 9 observations obtained with the HST throughout 1994, and minimizing the fit's X2, the 1.00 comet's grain size and outflow velocity distributions are extracted. Only a brief description of the modeling efforts is described here ;a detailed account of shall be provided in a future communication as well as results obtained from observations of several other S L 9 fragments. The motion of a model S L 9 fragment is numerically 0.10 integrated forward in time following the moment of tidal breakup. As it orbits Jupiter, the simulated comet frag- ment ejects dust grains of various radii R at velocity V(R) Zo tailward in random directions from its sunlit hemisphere. The I model nucleus and its dust grains are subject tojovian and '1 solar gravities with the grains also experiencing radiation 0.01 pressure appropriate for their size. The model's dust size II distribution is divided into nine discrete size bins ranging sunward [[I from R = !ttm on up to 1cm. The allowed ejection vel- ocities V(R) are similarly discrete on 0.25m s-t velocity intervals (this quantization of the problem makes it 1.0 10.0 computationally tractable). If N(R,t) is defined as the Projected distance 9 (arcsec) cumulative number of all grains having radii smaller than Fig. 2.Sunward, tailward, and azimuthally averaged (360°) sur- R emitted by a given S L 9 fragment up until some face brightness profiles of fragment K on March 30, 1994.A p-I time t, then the task at hand is to solve for curve is drawn over the tailward profile and a p 3ncurve is dR(R,t) -_ ARdN(R,t)/dR, which is the differential dust plotted over the azimuthally averaged profile beyond p> 1" production rate of all grains in the size interval R + AR/2 at time t. The velocity distribution V(R) must also be ments imaged with HST throughout 1994 have comae solved for each size bin R. An important assumption made light distributions similar 3to Fig. 2 (Hahn et al., 1996). It here is that the differential dust production rate dR(R) is should be noted that brightness profiles extracted from constant with time: The comet dust is also assumed to the 1993 observations of S-L 9differ distinctly from Fig. obey the usual phase law log _b(_) = -_/3/2.5, where :t is 2 and obey a ,_p-O,7 dependence (from Fig. 6 of Jewitt the Sun comet Earth phase angle and the free parameter (1995), see also Weaver et al. (1994)). This suggests that /3is the phase coefficient. The additional light contributed the S-L 9dust production rate may have been decreasing by an unresolved spherical fragment of radius Rr is also with time prior to the 1994 observations considered here. included, and it is assumed to have a light distribution Spectroscopic searches for sublimating cometary gases given by the HST point-spread-function. Thus 20 par- have yielded null results for S-L 9 (Cochran et al., 1994; ameters specify a simulated set of S L 9 observations --a Weaver et al., 1994, 1995; St_we et aI., 1995), which is dN(R), V(R) pair for each grain size bin plus /3and Rr. not regarded as unusual for small icy bodies 5.4 AU away Once a set of parameters are chosen, brightness maps of from the Sun (Weissman, 1996). An unobserved surface the coma and tail are computed for various observation gas flow is the most likely source of the observed dust dates. emission, although alternate theories exist (see Olson and The downhill simplex method is used to search the Mumma, 1994; Rettig et al., 1996a). While only cursory available parameter space and minimize the fit's X2(Nelder comparisons of the appearance of dormant comet models and Mead, 1965; Press et al., 1994). A time-sequence of to the observations have been made (Hahn et al., 1996), contour maps of fragment K observed with HST is given their ability to fully explain the observed S L 9 phenom- in Fig. 3 as well as the resulting isophotes of the fitted ena are not promising. Below we describe Monte Carlo synthetic image. Although discrepancies exist between the simulations of an active comet in S-L 9's orbit about observed and modal isophotes at faint light levels, there Jupiter. Synthetic images of model comae are constructed, is overall good agreement between the observations and and a search of parameter space provides excellent fits to the fit. Figure 4 shows the fragment's dust production the HST observations of fragment K. Fragment K's grain distribution dN(R) and its mass loss distribution size and outflow velocity distributions are presented as dM(R) = (4rc/3)pgR _dN(R) which assumes a dust geo- well as its dust phase law coefficient and an upper limit metric albedo of 0.04 and a grain mass density on the fragment's radius. pg = 1gcm -_. The arrows represent upper limits on the production rates for the indicated size bins. These findings are typical of many comets in that production of the Simulations of an S-L 9coma and tail smallest grains are numerically favored but total mass loss rates are governed by the largest grains ejected. The mass Synthetic images of cometary dust comae and tails have been computed for comet S-L 9. By fitting model images 4The fragment's _p-_ and _p-3:2 brightness profiles, as well 3Again the exception isjust before impact when Jupiter's gravity as its nearly constant distance from Sun, suggest (but do not altered the coma/tail structures. Also, several of the dim, off- guarantee) that itsdust production rate did not vary significantly axis fragments did not exhibit profiles like those seen in Fig. 2. during the span of observations considered here. 1274 T. W. Rettig and J. M. Hahn : Comet Shoemaker-Levy 9 :an active comet January 25 March 31 1o .... i , J 101_' .... "' .... ' '__'_' [ff o u,l -5 , i .... i .... / . , , -I0 -5 0 5 10 N -I0 -5 0 5 10 aresee EJ arcsec May 18 June 26 .... i .... i .... i , . , 10 10 W 5 o 0 -5 -5 i _ i i i -10 -5 0 5 10 -10 -5 0 5 10 al'csec arcsee Fig. 3. Contour maps of fragments K (gray curves) and the fitted synthetic images (black curves). The coma observed on May 18 was clipped by the detector edge, and the "fins", along the faint isophotes are an artifact of the model which ejects dust having discrete, rather than continuous sizes and velocities. A star trail lies north of the fragment on March 31, and the arrow in the June 27 figure indicates the projected direction to Jupiter loss rate of grains detected in fragment - K's coma The uncertainties quoted in Fig. 4 are 68% confidence (30 pm< R < 3ram) is _/- 22 +_5 kg s- _. Evidently, coma intervals in the model-parameters _where pg = Igcm =_ ;: grains Smaller than 30 _m did not contribute a detectable and a = 0.04 has been assume& Howeve? 6fil- Systematic : _ amount of light scattering cross section, so only upper uncertainties are affected by the unknown grain density limits on their production rates are obtained. It is noted and albedo. Radiation pressure sorts dust grains accord- here that the mass loss rate for grains smaller than 3ltm ing to the product p_R, so if the true S L 9 grain density is at most 0.1 kgs -_ and well below Sekanina's earlier p_ differs from the value assumed here then the R axis in upper limit. Figs 4 and 5 should be divided by a factor pg expressed in It is interesting to compare the grain size distribution cgs units. The observed flux reflected by each grain size for fragment K with that for comet Halley. As is evident bin determines aR 2dN(R), so if an alternate albedo a is in Fig. 4, a single power law cannot accurately represent also preferred, the grain production rates in Fig. 4 should the fragment's grain size distribution. Nonetheless, com- be multiplied by O.04p2ja. If p_ is independent of grain ...... puting the logarithm slope of d:v-(R) over the 30pm< size (which might not be true if smaller cometary grains R<3mm size interval indicates dN(R),_R _ with are fluffy instead of compact) then it can be shown that a = 2.2+0.2, which is considerably flatter than the R 3.7 the mass loss rates dM(R)_pgR 3dN(R) are independent power law measured for comet Halley (Tokunaga et al., of the assumed grain density but still uncertain by a factor 1986; Waniak, 1992). If comet S L 9 had produced small of 0.04/a. grains in the same proportions as comet Halley, then they The dust outflow velocity distribution V(R) for frag- • would have been well above detection limits, as indicated ment K is given in Fig. 5for the 30t_m<R<3 mm grains, by Fig. 4. which, if described by a power law, would obey V_ R h T. W. Rettig and J. M. Hahn :Comet Shoemaker-Levy 9:an active comet 1275 1014 102 the fitting algorithm could not uniquely disentangle any \ x_ ctR-3.7 light reflected by an embedded comet fragment from that contributed by the surrounding dust coma. Only an upper ',, /,j 1012 ._ .....; " 10! limit of Rf< 1.2 km is obtained for fragment K's radius assuming an a = 0.04 albedo. This limit isabout two times •.,. , v/ dM(R) lolo .... ,, tighter than that reported by Weaver et al. (1995) and is _'. _ 100 slightIy smaller than the size estimate given by Sekanina (1996b). Note that even smaller size limits/estimates have lOs been obtained from models of the tidal disruption of the S-L9 progenitor [Rf_C(O.4)km (Asphaug and Benz, 1996)], models of the Jupiter impact events [Rt-<0.5km 106 :: x,,xx d (Mac-Low, 1996)], as welI as the minimum impactor size f':: _, 10"2 that may be inferred from the amount of CO observed in Jupiter's atmosphere at the K impact site [R_> ((0.4)km 104 x (Lellouch, 1996)], the source of which is thought to be of _x 10"3 cometary origin. 1 I0 100 1000 10000 The lower limit on the fragment's radius is Grain radius R (_tm) R,,,_,,= (3]VIAt/4npO I:__ 0.4 km in order for a fragment of Fig. 4. Fragment K's differential dust production rate d_r(R) density pf_0.6gs -_ to sustain a dust production rate of and mass production rate dM(R) as a function of grain radius _! _-22 kg s- 1during the At 2 2yr time interval between R. Arrows indicate upper limits. The dashed curve has the same breakup and impact. Consequently if it had an initial size logarithmic slope as an R- _7Halley-type grain size distribution smaller than 0.4 km, it would have completely evaporated before striking Jupiter. However, fragment K was indeed observed to strike Jupiter. Supposing this fragment main- 2.0 tained a constant dust production rate during its final orbit, then with an initial fragment radius of at least 0.5km, the final impactor radius would have had been 0.4 km or larger and would have ejected less than half its 1.5 mass. Similarly, the smaller fragments F, J, P_, P2, T, and U did not exhibit any impact signatures (Hammel et aI., > 1995; Chodas and Yeomans, 1996), so perhaps they sim- ply exhausted most of their mass before they could strike 1.0 " '"--, . .t_ the planet. ...t....................... O 5 _ 0.5 Conclusions Since molecular fluorescence from any cometary gases were below detection limits, it is only from pre-impact 10 1O0 1000 10000 observations of the dust that one might glean further Grain radius R (p_m) insight into the structural properties of the tidally dis- Fig. 5. The dust outflow velocity for fragment K and an R o._ curve. The arrow indicates an upper limit rupted comet S-L9 fragments as well as the physics of their cometary atmospheres. For this reason it iscritically important to correctly infer the history of dust emission with b = 0.1 _+0.2. The observed power-law dependence by S-L9 and its grain sizes and outflow velocities. The is significantly weaker than b = 0.5 predicted by the the- photometry of S- L 9 favors an active-comet hypothesis ory of dusty-gas emission from cometary surfaces (Gom- over dormancy, as evidenced by the circular isophotes in bosi et al., 1986) but this finding is typical of studies of its inner comae and its p _and p-3;2 brightness profiles. other comets (Fulle, 1990, I992, 1996; Waniak, 1992). Detailed comparisons of observations to an active-comet The observed dust velocities, _0.5 to lm s-1, are in good model further strengthens this contention. agreement with earlier estimates (Hahn et al., 1996 ;Rettig The appearance of fragment K's coma and tail at visible et al., 1996b), but they are significantly slower than the wavelengths was governed by relatively large 30pm< outflow velocities measured for large grains emitted by R<3mm grains ejected at velocities V-_0.5 1.0ms -_. other distant comets: V2:15ms -1 for Schwassmann Grains 10pm and smaller were not detected above a one- Wachmann 1(S-WI)at 6AU (Fulle, 1992), V>32ms -_ a confidence level, but it isreasonable to assume they were for Hale Bopp at 7AU (Kidger et al., 1996), V_5ms -1 produced at a rate below the _0.1 kgs -_ detection limit. for Chiron at 9AU (Fulle, 1994), and V_25ms -I for The non-detection of small grains that are othe_qse Halley at 14AU (Sekanina et al., 1992). The low velocities characteristic of most other comets isdue to S L 9's rather observed in S-L9 may indicate that its relatively large flat size distribution, dA/-(R),_:R--' 2+°-'. One may speculate grains were poorly coupled to an unseen gas flow. that its unusual size distribution was a consequence of the The parameter search algorithm yielded a phase law comet's tidally disrupted nature. Dust grains from most coefficient fl = 0.010_+_0.006 magnitudes deg- _.However comets emanate from, and are perhaps filtered by, an 1276 T.W. Rettig and J. M. Hahn :Comet Shoemaker-Levy 9:an active comet ancienotverlyinsgurfacemantleH.owevearsurfaceman- Noll, H. A. Weaver and P. D. Feldman, pp. 1 30.Cambridge tleisabsenotnatidallydisruptedcometa,ndthismay University Press, Cambridge. havepermittedthelargeSL9dustgrainstoescapweith Cochran, A. L., Whipple, A. L., MacQueen, P. J., Shelus, P. J., Whited, R. W. and Claver, C. F. (1994) Preimpact charac- greateeraseandresulitnadustsizedistributionthatwas terization of P/'Comet Shoemaker-Levy 9. Icarus 112, 528- flatterthanseeninmostcometsT.helargegrainsseenin 532. theSL9comameayalsobeanindicatoorftheparticular Fulle, M. (1990) Meteoroids from short period comets. Astron. icespecietshatwasresponsibfloerthecomet'dsustemis- Astrophys. 230, 220-226. sion.AtSL9'sdistancferomtheSunw, aterproduction Fulle, M. (1992) Dust from short period comet P/Schwassmann- wouldhavebeentoofeebletolaunchanygrainslarger Wachmann 1and replenishment of the interplanetary dust than_6(1)#mfromthesurfacoefan cloud. Nature 359, 4244. Rf_C(1) km comet Fulle, M. (1994) Spin axis orientation of 2060Chiron from dust fragment (see Hahn et at., 1996). However models indicate coma modeling. Astron. Astrophys. 282, 980 988. that the sublimation of more volatile species such as CO Fulle, M. (1996) Dust environment and nucleus spin axis of or CO2 may have been sufficiently vigorous to loft grains comet P/Temple 2 from models of the infrared dust tail as large as a few millimeters. observed by IRAS. Astron. Astrophys. 311,333--339. During the 1994 observations, fragment K was ejecting Gombosi, T. I., Nagy, A. F. and Cravens, T. E. (1986) Dust and the 30_m<R<3mm grains at a rate of M= neutral gas modeling of the inner atmospheres of comets. 22+ 5kg s- _.In order to sustain this rather vigorous mass Rev. Geophys. 24, 667 700. Hahn, J. M., Rettig, T. W. and Mumma, M. J. (1996) Comet loss rate the radius of fragment K must have been Shoemaker Levy 9dust. Icarus 12, 291-304. Rf> 0.4 km at the time of breakup, while the dust coma Hammel, H. B., Beebe, R. F., Ingersoll, A. P., Orton, G. S., modeling indicates that Rr< !.2 km during the 1994 obser- Mills, J. R., Simon, A. A., Chodas, P., Clarke, J. T., De Jong, vations. Thus without ever observing fragment K directly, E., DoMing, T. E., Harrington, J.,Huber, L.F., Karkoschka, its radius is constrained to lie within a fairly narrow size E., Santori, C. M., Toigo, A., Yeomans, D. and West, R. A. interval. (1995) HST imaging of atmospheric phenomena created by A comparison of dust production by the S L 9fragment the impact of Comet Shoemaker Levy 9. Science 267, 1288 K to comets SW 1and Chiton is in order; these bodies 1296. have estimated mass loss rates of _600 and _2kgs -T, Jewitt, D. (1995) Pre-impact observations of Comet Shoemaker Levy 9. In Proc. European SL-9/Jupiter Workshop, eds R. respectively (Fulle, 1992, 1994). However these comets West and H. B6hnhardt, ESO Conference and Workshop are much larger than S L 9, having radii Rs w 1_ 15km Proceedings No. 52, pp. lq. (Meech et al., 1993) and Rch_ro°_84km (Altenhoff and Jewitt, D. C.and Meech, K. J. (1987) Surface brightness profiles Stumpff, 1995). A comparison of mass loss rates per of 10comets. Astrophys. J. 317, 992 1001. nuch, us surface area reveals that the surface of fragment Kidger, M. R., Serra-Ricart, M., Bellot-Rubio, L. R. and Casas, K was at least _ 6 times more active than S W 1, and at R. (1996) Evolution ofa spiral jet inthe inner coma of Comet least _5 x 103 times greater than Chiron. Thus this S-L Hale Bopp (1995 O1). Astrophys. J. 461, L119 L122. Lellouch, L. (1996) Chemistry induced by impacts: obser- 9 fragment, and perhaps all the others, were extremely vations. In The Collision of Comet Shoemaker-Lecy 9and active in comparison to other comets orbiting at com- Jupiter, eds K. S. Noll, H. A. Weaver and P. D. Feldman, parable distances from the Sun. This fact may also be a pp. 213-242. Cambridge University Press, Cambridge. consequence of S L 9's tidal disruption which stripped Mac-Low, M.-M. (1996) Entry, and fireball models vs. obser- any surface mantle from the fragments that might other- vations: what have we learned? In The Collision of Comet wise have constricted their dust production. Shoemake_Lery 9andJupiter, eds K. S.Noll, H. A. Weaver and P. D. Feldman, pp. 157-182. Cambridge University Acknowledgements. This research was performed while J.M.H. Press, Cambridge. was in residence at the University of Notre Dame. Support for Meech, K. J., Belton, M. J. S., Mueller, B. E. A., Dicksion, this work was also provided by NASA through grant 5624.21- M. W. and Li, H. R. (1993) Nucleus properties of P/ Schwassmann Wachmann 1.Astronom. J. 106, 1222-1236. 93A from the Space Telescope Science Institute which is oper- ated by the Association of Universities for Research in Astron- Nelder, J. A. and Mead, R. (1965) A simplex method for func- omy, Incorporated, under NASA contract NAS5-26555. This tion minimization. Comput. J. 7,308 313. paper represents contribution 924from the Lunar and Planetary Olson, K. M. and Mumma, M. J. (1994) Simulations of the Institute, operated by the Universities Space Research Associ- breakup and dynamical evolution ofComet Shoemaker/Levy ation NASA contract NA6W 4575. 9employing aswarm model. Bull. Am. Astron. Soc. 26, 1574- 1575. Press, W. H., Flannery, B. B., Teukolsky, S. A. and Vettarling, W. T. (1994) Numerical Recipes b_C. Cambridge University References Press, Cambridge. Rettig, T. W., Mumma, M. J., Sobczak, G. J., Hahn, J. M. and Altenhoff, W. J. and Stumpff, P. (1995) Size estimate of"aster- DiSanti, M. (1996a) The nature of Comet Shoemaker Levy oid" 2060 chiron from 250GHz measurements. Astron. 9 sub-nuclei from analysis of pre-impact HST images. J. Astrophys. 293, L41 L42. Geophys. Res. Planets 101,9271-9282. Asphaug, E and Benz, W. (1996) Size, density, and structure of Rettig, T. W., Sobczak, G. W. and Hahn, J. M. (1996b) Dust Comet Shoemaker Levy 9inferred from the physics of tidal outflow velocity for Comet Shoemaker-Levy 9. h'arus 121, breakup, h'arus 121,225- 248. 281-290. Chernova, G. P., Jockers, K. and Kiselev, N. N. (1996) Imaging Sekanina, Z. (1996a) Tidal breakup of the nucleus of Comet photometry and color of Comet Shoemaker-Levy 9. Icarus Shoemaker Levy 9. In The Collision of Comet Shoemaker 121,38_45. Let'), 9 and Jupiter, eds K. S.Nolt, H. A. Weaver and P. D. Chodas, P. W. and Yeomans, D. K. (1996) The orbital motion Feldman, pp. 55--80. Cambridge University Press, and impact circumstances of Comet Shoemaker-Levy 9.In Cambridge. The Collision of Comet Shoemaker Let,3'9and Jupiter, eds K Sekanina, Z. (1996b) Fragmentation ofComet Shoemaker Levy T. W. Rettig and J. M. Hahn : Comet Shoemaker-Levy 9: an active comet 1277 9's nuclei during flight through the jovian atmosphere. In Waniak, W. (1992) A Monte Carlo approach to the analysis of Physics, Chemistry, and Dynamics of lnterplanetary Dust, eds the dust tail of comet P/Halley. Icarus 100, 154_161. B. A. S. Gustafson and M. S. Hanner, pp. 383-386. Weaver, H. A., Feldman, P. D., A'Hearn, M. F., Arpigny, C., Sekanina, Z., Larson, S. M., Hainaut, O., Smette, A. and West, Brown, R. A., Helin, E. F., Levy, D. H., Marsden, B. G., R. M. (1992) Major outburst of periodic comet Halley at a Meech, K. J., Larson, S. M., Noll, K. S., Scotti, J. V., Sekan- heliocentric distance of 14AU. Astron. Astrophys. 263, 367- ina, Z., Shoemaker, C. S., Shoemaker, E. M., Smith, T. E., 386. Storrs, A. D., Yeomans, D. K. and Zellner, B. (1994) Hubble Sekanina, Z., Chodas, P. W. and Yeomans, D. K. (1994) Tidal Space Telescope observations of Comet P/Shoemake_Levy disruption and the appearance of periodic comet Shoemaker 9(1993e). Science 263, 787 791. Levy 9. Astron. Astrophys. 289, 607-636. Weaver, H. A., A'Hearn, M. F. A., Arpigny, C., Boice, D. C., Stt_we, J. A., Schulz, R. and A'Hearn, M. F. (1995) NTT obser- Feldman, P. D., Larson, S. M., Larny, P., Levy, D. H., vations of Shoemaker Levy 9. In Proc. Europoean SL-9/ Marsden, B. G., Meech, K. J., Noll, K. S., Scotti, J. V., Jupiter Workshop, eds R. West and H. B6hnhardt, ESO Sekanina, Z., Shoemaker, C. S., Shoemaker, E. M., Smith, Conference and Workshop Proceedings No. 52, pp. 17-22. T. E., Stern, S. A., Storrs, A. D, Trauger, J. T., Yeomans, Tokunaga, A. T., Golisch, W. F., Griep, D. M., Kaminski, C. D. K. and Zellner, B. (1995) Hubble Space Telescope (HST) D. and Hanner, M. S. (1986) The NASA Infrared Telescope observing campaign on Comet Shoemaker Levy 9. Science Facility Comet Halley monitoring program--l. Preperihelion 267, 1282-1287. results. Astronom. J. 92, 1183 1190. Weissman, P. R. (1996) If it quacks like acomet .... Icarus 121, Wallace, L. V. and Miller, F. D. (1958) Isophote configurations 275-280. for model comets. Astronom. J. 63, 213-219. r- ....-z _ ....... -_ _ ;_-!7 , ! L± ..... :;L...._± J _7.... ; ' ........ 7' l