Melissa A. McGrath
Space Telescope Science Institute, Baltimore, MD 21218 USA
abundance in the Venus atmosphere since 1991; significantly changed our
understanding of the Martian climate; made possible measurement of the
zonal wind fields and their change with height on Jupiter and Saturn;
determined the location of the UV auroral oval on Jupiter; resulted in
the detection of an oxygen atmosphere on Europa, and ozone on Ganymede;
witnessed the formation and destruction of satellites of Saturn; and
directly imaged for the first time the surface of Titan and surface
albedo patterns on Pluto.
Keywords: planets
Since its launch in 1990, the Hubble Space Telescope (HST) has been used
to observe every planet in the solar system save one, numerous planetary
satellites, ring systems, asteroids and comets. These observations have
resulted in a multitude of remarkable discoveries, some highlights of
which are reviewed in this article. A review of cometary observations
with HST is presented separately, in an accompanying article by P.
Feldman (p. 
). The subject matter of this paper is organized
as follows: section 1 discusses the inner planets, Mercury and Venus;
section 2 discusses Mars; section 3 presents results on the gas giant
planets organized by the general research topics of atmospheric dynamics,
aurorae, satellites and the comet Shoemaker-Levy 9 (SL9) impact with
Jupiter; and section 4 presents results on Pluto/Charon.
Observations within 50^o of the sun are prohibited by the
guidelines set forth in the HST Constraints and Requirements Document
(CARD). Reaching a maximum elongation of only 
25^o, Mercury
is 
25^o inside the current CARD limit, and so close to the
sun that operation of the telescope, which was not designed for solar
exposure, could be dangerous. The technique used to observe Venus, that
is, observing the target while the sun was blocked by the Earth, is not
feasible for Mercury because the time required to take the exposures
and then slew the telescope beyond the 50^o limit exceeds the time
available between target and sun rise. Illumination of the HST primary
mirror would occur at a sun angle of about 20^o, while pointing within
25^o from the sun would result in illumination of the interior of the
telescope tube. Such illumination would quickly heat up the tube to high
temperatures, releasing contaminants and potentially causing warping
and misalignments significant enough to affect optical performance.
A failure during such a dangerous pointing could well result in severe
degradation or loss of the telescope. Consequently, Mercury remains the
only planet in the solar system that has not been observed with HST.
By contrast with Mercury, the maximum elongation of 46--47^o for
Venus occurs much nearer the CARD limit. After extensive engineering
analysis considering possible risks to the telescope, a waiver of the CARD
limit by STScI and Project management permitted the first observations
of Venus to take place in January 1994. The special precautions taken
for these observations allowed only about 5 minutes of exposure time per
orbit during the four orbit duration of the program. The goal of the Venus
program, under the direction of Principal Investigator Larry Esposito
(University of Colorado), was to measure the SO
and SO abundances
and vertical distributions near the cloud tops in the middle atmosphere.
SO
is of particular interest because the chemical process creating the
Venusian clouds is thought to be oxidation of upwelling SO
. Knowledge
of SO
vertical distribution and abundance is, therefore, essential for
photochemical models of the Venus atmosphere. The photochemical reduction
of SO
to polymorphic S or S
O may create the characteristic
dark UV markings that distinguish Venus. Freshly created H
SO
aerosols that are chemically derived from SO
are thought to account
for the patches of bright haze observed above the Venusian poles.
Thus, the horizontal motions of the atmosphere, the vertical uplift,
and the major chemical and dynamical processes at the cloud tops, are
all observable by monitoring SO
.
The SO
abundance at the time of its discovery in 1978 was orders of
magnitude larger than the previously established upper limits (Stewart
et al. 1979, Barker 1979, Conway et al. 1979, Jenkins et al. 1979,
Owen & Sagan 1972). Continued observations by Pioneer Venus and the
International Ultraviolet Explorer (IUE) from 1978 through 1990 showed
a steady decline from 
100ppb at 40mbar in 1978 to 
40ppb at
40mbar in 1990 (Esposito 1984, Esposito et al. 1988, Na et al. 1990).
There have also been indications that an unusual event occurred in 1958,
when a brightening of polar haze and small aerosols was observed from
Earth (Dollfus et al. 1979). This evidence led Esposito (1984) to
conclude that a possible natural physical explanation for the episodic
nature of SO
is active volcanism on Venus. Volcanic domes and surface
wind streaks emanating from vents observed by the Magellan spacecraft
and widespread and recurring lightening also support the case for active
volcanism. However, the connection between SO
abundance above the
clouds and volcanic activity has not been unambiguously established,
since the abundance of SO
below the clouds has not changed since 1978.
The broad question being addressed by the HST observations is whether the
change in SO
abundance at the Venus cloud tops since 1978 is consistent
with the decrease expected after a single large scale injection event,
or whether it can be explained by variations in atmospheric circulation.
A total of three GHRS spectra and two WFPC2 images were obtained.
GHRS observations were made with the G200M grating centered at 207 nm
(near the peak of the SO
absorption cross section) and at 216 nm,
and with the Echelle-B grating centered at 207 nm. The WFPC2 images
were obtained with the F218W and F255W filters. An SO
abundance of
20
10ppb with a scale height of 2km at the altitude of the cloud
tops has been derived from the G200M spectra using a model atmosphere with
inhomogeneous distribution of SO
, SO and S
O (Na & Esposito 1995).
An SO mixing ratio of 3ppb at the same altitude has also been estimated.
Comparison with earlier SO
measurements shows the SO
abundance
above the clouds tops is lower by about a factor of 2 to 5 than the
amount derived from the most recent sounding rocket observation in 1991.
HST observations, therefore, show the SO
abundance continuing to decline,
implying no recent volcanic outburst. A more detailed analysis of the
data will be required to establish whether this continued decline can
be explained by photochemical and dynamical processes in the atmosphere.
The first General Observer (GO) observations performed with HST were of Mars in December 1990. The solar system moving target software was not operational at the time, and all of the early observations had to be programmed as fixed targets. These observations initiated a major GO program led by Phil James, University of Toledo, that has continued since 1990. The major goals of the program are to measure interannual (seasonal) variability of surface and atmospheric phenomena, to compare the observations with previous spacecraft and telescopic observations during the same season, and to constrain the state and composition of the atmosphere using UV observations.
Conventional ground-based observations of Mars are generally limited
to a 3--4 month period around opposition when the angular size of Mars
is larger than 
10 arcsec. HST significantly extends the useful
observing period to about 13 months of the 26-month Martian synodic year.
These windows are separated by about 1 year, and they move relatively
slowly with respect to the Martian seasons. Increasing the portion
of the Martian year during which synoptic changes can be resolved
is critical to disentangling the various time scales for large-scale
change on the planet. Large scale albedo changes, for example, were
previously thought to be the result of seasonal effects largely because
the restricted ground-based observing window, which moves relative to
the Martian seasons in a 15-year cycle, made it difficult to disentangle
changes on a seasonal time scale from those on an interannual time scale.
Furthermore, the lowest resolution HST imaging (near quadrature) is
equivalent to the best ground-based imaging near opposition.
The program includes both PC imaging with filters ranging from 230--890nm, and FOS spectroscopy to measure atmospheric ozone abundance. To date, three sets of observations have been performed from Dec 1990--May 1991, May 1992--April 1993 and Aug 1994--Aug 1995. A cycle 6 program has also been approved. A summary of the 1990--1991 observations has been published by James et al. (1994), and more recent observations have been discussed by Clancy et al. (1996a,1996b) and Martin et al. (1995). This observing program has resulted in a significant change in our understanding of the Martian climate, showing that the Mars atmosphere is cooler, clearer and drier than during the time of the Viking missions.
During the first Martian year of the Viking missions (1977), there were two major global dust storms on Mars, which resulted in dramatic changes in the Syrtis Major region following the second storm. A general regional brightening occurred after the dust storms, followed by regional darkening. The regional albedo increase following global dust storms was thought to be caused by enhanced deposition of bright dust from the atmospheric dust load. Net removal of dust, which is readily ejected by the effective regional winds, results in a decreased regional albedo as the year progresses. Once most of the dust has been transported from the region, significant variability will not occur until a dust storm again initiates the cycle. This process would presumably be repeated in any global dust storm cycle.
To first order, the Syrtis Major region has not shown any significant
change over the time period of the HST observations, as would have been
expected following major global dust storms during the preceding season.
The dust opacity of the Martian atmosphere has been constrained by
comparing ultraviolet images with atmospheric scattering models,
and by examination of contrast changes of surface features at visible
wavelengths. An upper limit of 0.1 for equatorial dust opacity near
vernal equinox has been derived, compared to values of 0.4 to 0.5
measured in Chryse by Viking Lander 1. An upper limit of 0.2 was found
for dust in the southern hemisphere. Except for suggestions of local dust
activity near the north polar cap in late spring, there has been no visual
evidence of global dust storms on the planet from 1990--1994. However,
two dust storms, in April and August 1995, have recently been reported,
and a major dark albedo feature (Cerberus, 
1500
500km) has
disappeared from the surface since Viking (Lee, personal communication).
Colder atmospheric temperatures, attributed to the absence of major
dust activity, have also been recently measured (Clancy et al. 1996a).
Dust absorbs sunlight, which in turn heats the atmosphere, so a relatively
dust-free atmosphere should be cooler. At the colder temperatures, water
vapor at low altitudes freezes out to form ice crystal clouds, now seen
in abundance by HST (Martin et al. 1995). There is also an inverse
photochemical relationship between Mars atmospheric water and ozone.
The broad Hartley band of ozone (centered at 255nm) is within the
F230W filter bandpass, and the 230W/336W ratio is, therefore, sensitive
to ozone, and provides a uniquely global view of Mars atmospheric ozone
at good spatial resolution. High latitude ozone has been observed at a
concentration similar to that observed by Mariner 9: a seasonal dependence
in ozone concentration was observed with concentration decreasing at
northern arctic latitudes and increasing at southern mid-latitudes as L
increases from 350^o to 60^o. More recent FOS observations (Clancy
et al. 1996b) have found that ozone now extends from the north pole to mid-
and lower latitudes, and has increased by a factor of 3 at the equator.
Despite the difference in dust loading, some aspects of the surface are unaffected by the different climate: the north pole cap is like that seen during the Viking and Mariner years. The behavior of condensate clouds and the recession of the north polar cap are consistent with those observed during Viking years. Finally, the albedo of the seasonal ice cap was found to decrease with wavelength, implying that the surface is not continuous clean ice.
Additional HST observations of Mars have been performed to search for
deuterium Ly-
emission (Bertaux et al. 1992), and to determine
the CO abundance in the Mars' atmosphere (Barnet et al., this volume,
p. ).
The majority of HST observations of the gas giant planets Jupiter, Saturn, Uranus and Neptune can for the most part be grouped into four broad scientific areas, atmospheric dynamics, aurorae, satellites, and the comet Shoemaker-Levy 9 impact with Jupiter, in lieu of a planet by planet summary.
The outer planets have strong prevailing zonal winds of amplitude
larger than their rotational velocities. The approximate magnitude of
these winds are 100m/s, 400m/s, 200m/s and 400m/s for Jupiter, Saturn,
Uranus and Neptune, respectively, relative to their rotational velocities.
The winds are strongest at the equator, alternate between prograde and
retrograde relative to rotation direction with increasing latitude, and
diminish in magnitude with increasing latitude. This pattern, seen on
all the gas giant planets, is commonly referred to as the zonal winds (or
jets), and the largest amplitude zone at the equator as the equatorial
jet. The zonal winds are very stable, as evidenced by ground-based,
Voyager and HST observations spanning more than a decade which show
nearly identical wind fields, and their energy source is not understood.
The magnitude of the winds on the different planets is not proportional
to any readily identifiable forcing function, such as solar insolation,
which is 
30 times weaker on Neptune than on Jupiter, or the internal
heat sources of the various planets. The radiative time constant,
which is a measure of the inertia of an atmosphere in its response to
an external radiative forcing, is very long on these planets (e.g.,
4
10
or greater in Jupiter's troposphere and stratosphere,
which is so long that diurnal variations attributable to purely radiative
effects are precluded). It is difficult to understand how the very
stable, large-scale order implied by the zonal winds exists given the
small-scale chaos that is readily apparent in the visible clouds.
Most of the observational studies are limited to determining the wind field at the altitude of the visible clouds in the troposphere. Although very difficult to determine observationally, a key piece of information relevant to distinguishing among the various proposed models is determination of how the strength of the wind field changes with height, both below and above the visible cloud tops. Two major events studied extensively with HST, the Saturn storm of 1990 and the comet Shoemaker-Levy impact with Jupiter in 1994, offered unique opportunities to map the wind fields much higher in the atmosphere due to the unusual ejection of relatively long-lived (t > few rotation periods) material into the stratosphere during these events.
The Saturn storm of 1990 was observed in detail by HST (Beebe et al, 1992, Westphal et al. 1992) using WFPC. Although dependent on a number of assumptions, the difference in average zonal winds found in the green F547M filter and the methane F889N filter (which measures winds about one scale height above the green filter) was interpreted as evidence for vertical shear in the zonal winds in the equatorial region. This was the first direct detection of the decay of the zonal winds with height on Saturn (Barnet et al. 1992). HST observations also resulted in interpretation of the storm as a single eruptive event.
Observations of the evolution of the SL9 impact sites provided the first-ever opportunity to measure wind fields in the stratosphere, and even higher in the thermosphere, on Jupiter. The cores of the impact sites were observed to break up and be dispersed by the ambient tropospheric winds, indicating that some impact debris had settled in the troposphere. Gradients in the wind shear as a function of latitude and meridional transport by vortex motion accelerated the spread of material. The zonal wind field, measured by tracking features in the images, was found to be essentially identical to the wind profile determined from Voyager data. Even material that was known, because of it similar appearance in the F410M and methane filters, to reside in the stratosphere seemed to ``sense'' the ambient tropospheric wind field (Hammel et al. 1995a). Finally, optically thin material was seen to drift northward with an average velocity of 30m/sec, which was consistent with the measurements of the drift at 50m/sec of much higher-altitude material seen in UV images (Clarke et al. 1995).
In addition, observations of visible clouds on Uranus (Seidelman et al. 1995) and Neptune (Hammel et al. 1995b) with HST have provided important observational data concerning the stability of the zonal winds on these planets.
The Jovian aurora has been one of the most extensively studied solar
system phenomena with HST. It dominates the energetics of the upper
atmosphere, depositing energy of 10
--10
W, which is 100--1000
times more powerful than the Earth's auroral energy input and a few to
50 times greater than the global solar UV flux absorbed at the same
altitudes. This energy is believed to produce extreme conditions in
Jupiter's auroral ionosphere and thermosphere and to dominate the global
thermospheric circulation. The auroral emission occurs primarily in
the UV and is produced by secondary electron excitation of atmospheric
H
and H gas, which produces strong H
Lyman and Werner band and
H Ly
emission.
Until recent HST images, Voyager spectroscopic observations had provided
the only quantitative measurement of the location of the UV auroral oval
on Jupiter, and these provided only a constraint, not an actual location.
Following the Voyager missions, IUE spectroscopy was the workhorse for
studying the Jovian aurora. IUE observations indicated a morphology of a
bright spot at 180^o longitude in the north, and 50^o in the south.
The location of the auroral oval provides reliable information about the
magnetic field at the ``surface'' (1R
) of the planet, as well as
information in particular about the higher order moments of the field.
The Voyager missions provided accurate measurements of the field only
in to about 5R
for Jupiter, which accurately samples only the far
(primarily dipolar) field. Models of the field extrapolated to the
surface are expected to be inaccurate.
For Jupiter, in particular, there have been several outstanding issues:
are the aurorae triggered by precipitating particles that are intrinsic
to the Jovian magnetosphere, or are they external (i.e., solar wind)
particles?; what is the identity of these primary particles, heavy
ions of S and O from the Io atmosphere and torus, or protons?; where do
the particles precipitate from?; what is the relationship between the
UV auroral emission and the IR auroral emission, which are thought to
originate from very different levels in the Jupiter atmosphere; and what
is the global effect of the aurora on the atmosphere? Direct imaging
was expected to provide concrete answers to many of these questions,
particularly on the location from which particles precipitate (i.e., the
precise location of the auroral oval on the planet). It was, therefore,
a major breakthrough when HST obtained the first direct images of the
Jupiter UV auroral emission on 8 Feb 1992 (Caldwell et al. 1992) using
the Faint Object Camera (FOC) in its f/96 zoom mode. To suppress red
leak in the FOC UV filters (which would otherwise produce a visible image
despite the filter bandpass), double filter combinations had to be used,
in this case F140W+F152M. This technique has been used successfully by
numerous observing teams with various permutations of the double filter
combinations to obtain either the H
emission or the H Ly
emission, both for Jupiter (Dols et al. 1992, Gerard et al. 1993,
Gerard et al. 1994, Prange et al. 1995) and Saturn (Gerard et al. 1995).
The disadvantage of this technique is that it results in very low signal
to noise ratio (S/N) data.
An equally significant breakthrough in Jovian auroral studies came with the installation of the new WFPC2 camera (Trauger et al. 1994) during the first servicing mission. The WFPC2 team had installed an innovative new filter (F160BW, the ``Woods'' filter), particularly well-suited for direct imaging of the Jovian UV aurora. This filter has insignificant red leak, and significantly higher throughput (Clarke et al. 1995b) than the double filter combination required for FOC imaging, resulting in significantly higher S/N images with the new camera. The first images using this technique were obtained in May 1994, and a summary of the early WFPC2 imaging results on the Jovian aurora were presented at this meeting by J. T. Clarke. WFPC2 has also obtained significantly improved auroral imaging of Saturn (Trauger et al. 1995).
The GHRS has also been used to observe the Jupiter aurora
spectroscopically. Observations by several teams (Trafton et al. 1994,
Kim et al. 1994, Clarke et al. 1994) in 1992 and 1993 using
the identical GHRS setup (G160M and the large science aperture)
have covered various wavelength regions centered on the strongest
vibrational-rotational lines of the H
Lyman and Werner band systems,
and on H Ly
. These spectra have been used to derive H
rotational temperatures at the altitude where most of the emission
originates. IR auroral observations give hydrocarbon temperatures of
180--500K over 1--0.001 mbar, and H
values of 900--1500 K over
0.1 and 0.001 
bar, implying a rapid increase in temperature
with altitude. The UV auroral emission is thought to originate at
intermediate altitudes near the hydrocarbon homopause from modeled
hydrocarbon absorption of the UV auroral emission. Although there
are relatively large uncertainties of as much as 100--200K in derived
temperatures, the results of the three groups are in good agreement,
with best-fit temperatures ranging from 
250--750 K. Part of
the spread in derived temperatures is also likely to be real, since
spatially distinct auroral regions of both bright and faint emissions,
which would be expected to show different temperatures, were observed,
and because large temporal variability, also thought to be connected
with temperature changes, was observed.
Galilean Satellites
The Galilean satellites of Jupiter have been a primary target for HST. Unresolvable from the ground except at infrared wavelengths, HST has performed the first spatially resolved imaging of these satellites at visible wavelengths since the Voyager flybys in the late 1970s, and the first spatially resolved imaging at UV wavelengths. Loss of some of the high-resolution imaging sequences from the Galileo mission has emphasized the need for the Hubble capability.
Io, because of its very unusual nature as the most volcanically active
body in the solar system, and its direct links to the Jovian magnetosphere
(through its associated plasma torus) and the Jovian atmosphere
(via the strong current system that links the two) has been the most
extensively observed planetary satellite with HST. Io is known to have
visibly bright regions of SO
frost, and a tenuous SO
atmosphere.
It was an early target of the FOC GTO program at UV wavelengths, and
the first spatially resolved UV image of this satellite was acquired
on 5 March 1991 (Paresce et al. 1992). Combined with subsequent imaging
at UV and visible wavelengths acquired with both the FOC and WFPC, the
early imaging observations emphasized that despite extensive volcanism,
the majority of the satellite's surface albedo patterns remained unchanged
since 1979 (implying that recurring volcanism tends to occur in the same
places), and that regions that are bright in the visible tend to be dark
in the UV, as expected for SO
-rich frost patches. One large area
(Colchis Regio) exhibited a significant change in albedo since 1979.
In addition, the presence of an SO
gas
patch (presumably a volcanic plume) over the volcano Pele was a possible
interpretation of the data (Sartoretti et al. 1994,1995). Complete
global imaging of the satellite with the WFPC2 from UV to near-IR
wavelengths was also performed in 1994 (Spencer et al. 1994) and again
in 1995. The 1995 imaging showed a dramatic change in the surface albedo
of the satellite surrounding Ra Patera, probably composed of frozen SO
\
gas ejected by a large volcanic explosion or fresh lava flow between March
1994 and July 1995 (Spencer et al. 1995), and also provided tentative
evidence for an active plume over the Pele volcano, imaged at the limb.
HST imaging of Europa (see Figure 1), Ganymede, and Callisto from UV to
near-IR wavelengths was also obtained for the first time in late June
1995 using WFPC2 (Noll et al. 1995b). Major surface features such as
the Asgard impact crater on Callisto and Galileo Regio on Ganymede are
readily identifiable in the images. Numerous discrete albedo features
appear on all three satellites. The UV images (
260
and 280nm) do not differ significantly from the visible images with the
exception of a UV-bright region at high southern latitudes on Callisto's
leading hemisphere. Spectrophotometry of selected regions shows variation
suggesting compositional and/or structural differences between bright
and dark albedo features.
Extensive spectroscopic observations have also been performed of the
Galilean satellites and the Io plasma torus. Observations using the GHRS
in 1992 resulted in the first detection of OII 
2471 from the Io
plasma torus (McGrath et al. 1993), which is important because O
is
the most abundant ion in the torus but is difficult to measure due to the
relative weakness of the collisionally excited forbidden transitions at
3726 and 3729Å, and because it has no strong transition in the far-UV,
a region well-studied by the IUE satellite over the past decade.
HST GHRS observations of Io in March 1992 resulted in the first detection
of gaseous SO
band absorption in the ultraviolet spectral region
(Ballester et al. 1994, Trafton et al. 1996), where the SO
\
absorption cross section peaks. This observation comprised only the
third positive means of detection of the SO
atmosphere since its
discovery in 1979, and resulted in quantitative estimates of the column
density of atmospheric SO
gas. Detailed modeling of the spectra
resulted in best fit models that incorporated a ``patchy'' atmosphere,
a result consistent with earlier observations of Lellouch et al. (1992).
Combined with the inference of SO
gas over the Pele volcano discussed
above, there is now growing evidence from several sets of observations
that the Io atmosphere contains regions of pronounced, highly localized
density enhancements, as might be expected from localized source regions
such as volcanoes or surface frost patches.
Because of its ice-covered surface, there have long been suggestions
that an oxygen atmosphere should accumulate around Europa. The source
is thought to be surface plasma bombardment triggering reactions which
break up the water molecules and form molecular hydrogen and oxygen.
The lighter H
molecules escape from Europa relatively easily, leaving
behind an atmosphere rich in oxygen. Observational confirmation of this
theory was obtained in May 1994 with the detection of atomic oxygen
emission from Europa (see Figure 1), which was interpreted as being
produced by the simultaneous dissociation and excitation of atmospheric
O
by electrons from Jupiter's magnetosphere (Hall et al. 1995a).
Europa's molecular oxygen atmosphere is very tenuous, with a surface
pressure of only 
10
that of Earth's atmosphere.
Figure: Left: HST image of Europa obtained with the WFPC2 in June 1995
(courtesy K. S. Noll). Right: Detection of OI on Europa in May 1994
implying the presence of a tenuous oxygen atmosphere (from Hall et
al. 1995a)
Ultraviolet FOS spectra (
200--330 nm) of Europa and Ganymede
obtained in June 1995 (Noll et al. 1995b) showed albedos lower than
expected based on existing IUE broad-band albedos (Nelson et al. 1987) and
phase coefficients. Discrete spectral features are present in the spectra,
most notably a broad feature at 280nm on Europa's trailing hemisphere
attributed to SO
. The ratio of Ganymede's trailing to leading
hemisphere spectrum results in a very broad absorption feature centered
at 
260nm. The position and width of the band is closely matched by
the gas-phase Hartley band of ozone. The possibility of O-band absorption
was suggested by Nelson & Lane (1987) based on IUE spectra. Sack et
al. (1992) have identified a UV spectral feature in laboratory samples
of Ar-irradiated H
O ice as possibly due to O
, strengthening the
idea that particle irradiation of Ganymede's trailing hemisphere might
form O
in radiation damaged ice. The recent detection of spectral
features attributed to O
in Ganymede's visible spectrum (Spencer,
Calvin, & Person 1995) is also consistent with this hypothesis.
This evidence for oxygen chemistry in the surface ice implies that
Ganymede, like Europa, is likely to have a tenuous oxygen atmosphere.
truein Titan
Titan possesses a very dense atmosphere of molecular nitrogen, about
four times as dense as the earth atmosphere, and it is shrouded in
clouds and haze. Titan is of great interest because its atmosphere
is thought to be similar to the pre-biotic atmosphere of the Earth.
The atmospheric aerosols are more transparent at near-IR wavelengths
than at shorter wavelengths, and IR observations should penetrate to
the surface. Ground-based observations have slowly revealed that
Titan's brightness varies with orbital longitude, being 
23%
brighter at 1.6
m at eastern elongation than at western elongation
(Lemmon et al. 1995). These observations determined unambiguously that
Titan is in synchronous rotation with a period of 15.949 
0.006
days, and has a bright surface feature located near 100^o longitude.
The first direct imaging of the Titan surface was performed by HST using
the Planetary Camera in October 1994, in a program led by Peter Smith
(University of Arizona). Titan was imaged with a suite of filters ranging
from F336W to F1042M at 14 intervals throughout one rotation. A bright
region on the leading hemisphere near the equator of about 2500km (about
the size of Australia) extending from 
70--130^o longitude was
found, in remarkable agreement with the earlier ground-based results.
The bright region is about 10--15% brighter than surrounding areas at
1.04
m. It is unclear whether the bright region is higher or lower
than the surrounding darker regions. The HST images provide invaluable
information for targeting the Huygens probe of Titan which will be
delivered by the Cassini mission to Saturn, to be launched in 1997.
Images of Titan were also obtained with the PC on 26 August 1990, and comparison of these images with Voyager 1 and 2 images obtained 10 and nine years earlier showed that the seasonal hemispheric brightness asymmetry had reversed near 440 and 550 nm wavelengths, with the northern hemisphere being brighter in the 1990 data (Caldwell et al. 1992).
truein Saturn ring plane crossing
Four times from May 1995 to Feb 1996 a unique opportunity to search for faint satellites and observe the tenuous outer rings of Saturn occurred when the earth and sun crossed the Saturn ring plane. These events are rare, occurring in groups of two or four every 14.5 years. Although Saturn was too close to the sun for HST to observe the last event on 11 Feb 1996, the 22 May, 10 Aug and 19 Nov 1995 events were each observed.
During the first event, in which the sun was 3^o above the ring
plane and the earth crossed from above to below, Bosh & Rivkin (1995)
obtained 27 400s images over 10 hours using the WFPC2 camera with the
890-nm methane absorption band filter (to minimize scattered light).
The time of minimum ring light occurred at 5:38 UT 
3 min, 23 min
later than the predicted time, which they interpreted by a combination
of change in pole position at the reference epoch and change in pole
precession rate. Accounting for the entire change by the precession rate
alone would require it to be 65% of its predicted value. They reported
the detection of four probable satellites (S/1995 S1, S2, S3 and S4)
outside the A ring in the vicinity of the F ring, each seen in at least
11 images. They suggested that S1 could be Atlas, and S2 Prometheus,
although they would vary from their predicted positions by 26^o and
21^o respectively, if these identifications are correct. All other known
satellites were accounted for except Saturn SVIII (Pan). Further analysis
of the data showed that S3 has an orbit identical to the F ring, and it
is now thought to be an arc or clump of F ring material.
During the second event, Nicholson et al. (1995) obtained 48 WFPC2 images, also using the 890-nm methane filter. They confirmed detection of S/1995 S2, and found a lag of 19^o behind the predicted position of Prometheus, comparable to that of Bosh and Rivkin, confirming its likely identification as Prometheus. It was 0.5 mag fainter than expected. They also reported detection of three new objects: S/1995 S5, S6, and S7. S5 and S6, whose orbits are similar to the F ring, were interpreted as arcs or clumps within that ring rather than previously undetected satellites, based on their brightness and the completeness limit of Voyager imaging searches. S5 also showed appreciable brightness variations with orbital phase, including a decrease in brightness at the edge of the ring, further suggesting an arc-like structure. The orbit of S/1995 S7 was found to be indistinguishable from that of Prometheus (= S/1995 S2), with the new object trailing the latter by 15^o.
The three new objects (S4, S5, and S6) seem to have disappeared by 21 Nov, when HST again viewed Saturn as the sun passed through the ring plane. Voyager found evidence of similar activity, but the new images mark the first time that observations from earth's vicinity have documented the actual formation (in the 14 years since Voyager) and subsequent destruction of ring material in a matter of weeks. The discovery of new objects in this transitional phase is not totally unexpected since they orbit Saturn near the narrow F ring, which is a dynamic transition zone between the main rings and the larger satellites. Moonlets in this region can be easily disrupted by Saturn's tidal pull if they are fractured by an impact, forming a cloud of debris. The apparent slip of Prometheus in its orbit by 20 degrees from the predicted position may also be a consequence of a collision with F ring material, which is believed to have occurred in 1993 as the moon passed close enough to one of the denser, lumpy regions of the F ring to have its orbit changed.
Figure: Left: Ratio of two WFPC2 F255W images taken on 4 March 1995
and 15 July 1994 showing the continued presence of SL9 impact-related
material near latitude -40^o nearly eight months after the impacts.
Note the latitudinal spreading of the material obvious near the limbs of
the planet. Figure courtesy R. A. West. Right: Ratio of pre-impact
(14 July 1994) and post-impact (3 March 1995) FOS spectra of the
impact latitude showing the continued presence of CS
and enhanced
NH
nearly 8 months after the impacts. From McGrath et al. (1995b)
During the second ring plane crossing in August, Hall et al. (1995b)
obtained UV spectra near the rings using the FOS that showed the
presence of fluorescent OH A
- X
(0,0) band emission.
These observations showed that Saturn's rings are embedded in a
tenuous gas of hydroxyl (OH) molecules with a scale height of 0.4R
.
Fluorescent emission from OH was also detected much further out in the
Saturn magnetosphere, near the orbit of Tethys (d
4.5R
)
in 1992 (Shemansky et al. 1993). The OH is thought to come from H
O
extracted from the icy satellites, and previous models had assumed charged
particle sputtering as the primary source. The large source rate implied
by the observed OH brightness implies that a micrometeorite source is
required in addition to charged particle sputtering. The inferred flux
of micrometeorites is much larger than previously predicted.
A comprehensive HST campaign was executed in July 1994 to observe the Shoemaker-Levy 9 impact with Jupiter. The HST campaign was led by six principal investigators: J. Clarke (University of Michigan), H. Hammel (MIT), M. McGrath (STScI), K. Noll (STScI), H. Weaver (ARC), and R. West (JPL), and included an additional 60 co-investigators. Initial results were published in March, 1995, in a special issue of Science with articles by each of the six PIs.
Highlights of the HST campaign included resolved imaging of the plumes
and impact sites, which showed clearly the morphology of the events.
Sequences of images of the impact sites revealed circular features that
increased in radius with time, and were interpreted as resulting from
tropospheric gravity waves, the speed of which implied that the Jupiter
O/H ratio is 10 times the solar value (Ingersoll & Kanamori 1995).
HST spectroscopic measurements following the impacts revealed numerous
new gaseous species present in the atmosphere, including S
, CS
,
CS, H
S, Mg, Si, and Fe; conspicuously absent were SO
, SO, and
CO
(Noll et al. 1995, Atreya et al. 1995, Yelle & McGrath 1996).
Enhanced amounts of ammonia (NH
), and H
emission were also
detected. HST observations also provided the best constraints on the
composition of the dark brown particulate material so readily apparent
in the impact sites (West et al. 1995a). Spectroscopic and imaging
observations well after the event, as late as March and September 1995,
still show clear evidence for the continued presence of impact-related
material in the Jovian stratosphere (West et al. 1995b, McGrath et al.
1995b see Figure 2).
Until the discovery of spherical aberration in the HST primary mirror,
the FOC GTO science team had hoped to utilize their instrument in the
high resolution f/288 mode, in combination with image restoration to
discern surface albedo distributions predicted by models developed
from observations of the Pluto-Charon mutual eclipses that occurred
from 1985--1990 (Buie, Tholen & Horne 1992). The first images, taken
in September 1990 several weeks after the initial deployment of the
telescope, suffered severely from the spherical aberration and were not
useful for the original goal. However, they were useful for refining
the Pluto-Charon orbit, which is critical in developing accurate surface
albedo models from the mutual eclipse data (Albrecht et al. 1991).
Additional early observations with the WFPC made in August 1991 also
made possible the first determination of the Charon/Pluto mass ratio,
0.0147, and significant improvement by almost a factor of
two for Charon's semimajor axis, 
86km, and the Pluto system
mass, M
=(1.401
inverse solar masses (Null et
al. 1993). The implied Pluto and Charon masses were M
=(13.10
0.24)
10
g and M
=(1.10
0.18)
g,
respectively.
Figure: Left: Two FOC F550M images of Pluto co-added, deconvolved, and
resampled into a factor of 2 finer grid. Right: Model reconstruction
of Pluto oriented as the data on the left and based on Johnson B band
albedo distributions determined by Buie et al. (1992). Both figures are
courtesy of R. Albrecht, from Albrecht et al. (1994).
After the HST First Servicing Mission, the FOC observations were repeated.
The f/288 mode was no longer available, but the original f/96 imaging
mode was changed to f/151 by the addition of the corrective optics.
The first post-COSTAR images were obtained during two orbits on 21
Feb 1994 using filters F342W and F550M when Charon was near maximum
southern elongation. The separation of the two bodies was 0.92
.
These images resulted in radius determinations for Pluto and Charon of
1160km and 635--650km, respectively. The value of 1160km for Pluto
compares with a value 1151
6km from Pluto-Charon mutual events
(Tholen & Buie 1990). The radius range for Charon disagrees with
determinations based on the Pluto-Charon mutual events (593
13km),
but is consistent with a lower limit based on stellar occultations (Elliot
& Young 1992). Assuming a value of 19,500km for the orbital radius,
a larger radius implies a somewhat lower mean density of between 1.87
and 1.89 g/cm
for the system.
Comparison of the restored average of two frames at F550M with an albedo map in the Johnson B band from Buie et al. (1992) projected onto a globe is shown in Figure 3. The FOC data have not been corrected for limb darkening, and so do not reproduce the bright polar region. Albrecht et al. (1994) note a general agreement of the major features, however, they also caution that it is very difficult to separate Pluto surface albedo distributions from detector effects. More observations, allowing better discrimination of possible effects, and better knowledge of the FOC calibration will be required to substantiate their results.
My very grateful thanks go to the following colleagues who have graciously shared and patiently explained their latest results for this review: Rudy Albrecht, Amanda Bosh, Todd Clancy, John Caldwell, John Clarke, Larry Esposito, Doyle Hall, Heidi Hammel, Steve Lee, Chan Na, Keith Noll, Renee Prange, Bruno Sicardy, John Spencer, Alan Stern, Peter Thomas, Larry Trafton, and Bob West. I would also like to thank the Goddard Space Flight Center Laboratory for Astronomy and Solar Physics, in particular Ken Carpenter and Susan Neff, for providing generous visitor support during the writing of this paper.
Albrecht et al. 1991, ApJ, 374, L65
Albrecht et al. 1994, ApJ, 435, L75
Atreya, S. K. et al. 1995, GRL, 22,1625
Ballester, G. E. et al. 1994, Icarus, 88, 1
Barker, E. S. 1979, GRL, 6, 117
Barnet, C. D., Westphal, J. A., Beebe, R. F., & Huber, L. F. 1992, Icarus, 100, 499
Beebe, R. F., Barnet, C., Sada, P. V., & Murrell, A. S. 1992, Icarus, 95, 163
Bertaux, J.-L., Clarke, J. T., Mumma, M., Owen, T., & Quemerais, E. 1992, in Science with the Hubble Space Telescope, P. Benvenuti & E. Schreier, ESO, 459
Bosh, A. & Rivkin, X. 1995, BAAS, 27, 77; also IAU Circ #6192; and submitted to Science, 1996
Buie, M. W., Tholen, D. J., & Horne, K. 1992, Icarus, 79, 38
Caldwell, J., Turgeon, B., & Hua, X.-M. 1992, Science, 257, 1512
Caldwell, J. et al. 1992, Icarus, 97, 1
Clancy, T. et al. 1996a, Icarus, in press
Clancy, T. et al. 1996b, submitted to JGR
Clarke, J. T. et al. 1995a, Science, 267, 1302
Clarke, J. T. et al. 1995b, BAAS, 27, 1149
Clarke, J. T., Ben-Jaffel, L., Vidal-Madjar, A., Gladstone, G. R., Waite, Jr, J. H., Prange, R., Gerard, J.-C., Ajello, J., & James, G. 1994, ApJ, 430, L73
Conway, R. R., McCoy, R. P., Barth, C. A., & Lane, A. L. 1979, GRL, 6, 629
Dollfus, A., Auriere, M., & Santer, R. 1979, AJ, 84, 1419
Dols, V., Gerard, J.-C., Paresce, F., Prange, R., & Vidal-Madjar, A. 1992, GRL, 19, 1803
Elliot, J. J., & Young, L. A. 1992, AJ, 103, 991
Esposito, L. W. 1984, Science, 223, 1072
Esposito, L. W. et al. 1988, JGR, 93 5267
Gerard, J.-C., Dols, V., Grodent, D., Waite, J. H., Gladstone, G. R., & Prange, R. 1995, GRL, 22, 2685
Gerard, J.-C., Dols, V., Paresce, F., & Prange, R. 1993, JGR 98, 18793
Gerard, J.-C., Grodent, D., Prange, R., Waite, J. H., Gladstone, G. R., Dlos, V., Paresce, F., Storrs, A., Ben Jaffel, L., & Franke, K. A. 1994, Science, 266, 1675
Hall, D. T. et al. 1995a, Nature, 373, 677
Hall, D. T. et al. 1995b, BAAS, 27, 1131
Hammel, H. B. et al. 1995a, Science, 267, 1288
Hammel, H. B. et al. 1995b, Science, 268, 1740
Ingersoll, A. P. & Kanamori, H. 1995, BAAS, 27, 1128
James, P. et al. 1994, Icarus, 109, 79
Kim, Y. H., Caldwell, J. J., & Fox, J. L. 1994, ApJ, 447, 906
Lellouch, E., Belton, M., De Pater, I., Paubert, G., Gulkis, S., & Encrenaz, T. 1992, Icarus, 98, 271
Lemmon, M. T., Karkoschka, E., & Tomasko, M. 1995, Icarus, 113, 27
Martin, L. J. et al. 1995, BAAS, 27, 1061
McGrath, M. A. et al. 1993, ApJ, 415, L55
McGrath, M. A. et al. 1995a, Science, 267, 1313
McGrath, M. A. et al. 1995b, BAAS, 27, 1118
Na, C. & Esposito, L. W. 1995, BAAS, 27, 1071
Na, C. Y., Esposito, L. W., & Skinner, T. E. 1990, JGR, 95, 7485
Nelson, R. N. et al. 1987, Icarus, 72, 358
Nicholson, P. et al. 1995, IAU Circ 6243
Noll, K. S. et al. 1995a, Science, 267, 1307
Noll, K. S. et al. 1995b, BAAS, 27, 111
Null, G. W. et al. 1993, AJ, 105, 2319
Owen, T. & Sagan, C. 1972, Icarus, 16, 557
Paresce, F. et al. 1992, A&A, 262, 617
Sartoretti, P. et al. 1994, Icarus, 108, 272
Sartoretti, P. et al. 1994, JGR, 100, 7523
Spencer, J. S. et al. 1994, BAAS, 26, 1135
Spencer, J. S. et al. 1995, BAAS, 27, 1160
Shemansky, D. E. et al. 1993, Nature, 363, 329
Stewart, A. I. F., Anderson, D. E., Esposito, L. W., & Barth, C. A. 1979, Science, 203, 777
Tholin, D. J. & Buie, M. W. 1990, BAAS, 22, 1129
Trafton, L. et al. 1996, ApJ, 456, 384
Trafton, L., Gerard, J.C., Munhoven, G., & Waite, Jr., J. H. 1994, ApJ, 421, 816
Trauger, J. et al. 1994, ApJ, 435, L3
Trauger, J. et al. 1995, BAAS, 27, 1148
Weaver, H. A. et al. 1995, Science, 267, 1282
West, R. A. et al. 1995a, Science, 267, 1296
West, R. A. et al. 1995b, BAAS 27, 1119
Westphal, J. A. et al. 1992, Icarus, 100, 485
Yelle, R. V. & McGrath, M. A. 1996, Icarus, 119, 90