\font\bfS=cmbx10 scaled\magstep1
\font\bfSS=cmbx10 scaled\magstep2
\def\degfrac{$^\circ\!\!$}
\def\deg{$^\circ$}
\def\arcsecfrac{$^{\rm\prime\prime}\!\!$}
\def\arcsec{$^{\rm\prime\prime}$}

\def\lse{\mathrel{\mathchoice {\vcenter{\offinterlineskip\halign{\hfil
$\displaystyle##$\hfil\cr<\cr\sim\cr}}}
{\vcenter{\offinterlineskip\halign{\hfil$\textstyle##$\hfil\cr<\cr\sim\cr}}}
{\vcenter{\offinterlineskip\halign{\hfil$\scriptstyle##$\hfil\cr<\cr\sim\cr}}}
{\vcenter{\offinterlineskip\halign{\hfil$\scriptscriptstyle##$\hfil\cr<\cr\sim\cr}}}}}
\def\gse{\mathrel{\mathchoice {\vcenter{\offinterlineskip\halign{\hfil
$\displaystyle##$\hfil\cr>\cr\sim\cr}}}
{\vcenter{\offinterlineskip\halign{\hfil$\textstyle##$\hfil\cr>\cr\sim\cr}}}
{\vcenter{\offinterlineskip\halign{\hfil$\scriptstyle##$\hfil\cr>\cr\sim\cr}}}
{\vcenter{\offinterlineskip\halign{\hfil$\scriptscriptstyle##$\hfil\cr>\cr\sim\cr}}}}}


\nopagenumbers

\rm

\centerline{\bfSS A search for comets encountering Jupiter --- the connection}
\centerline{\bfSS to the dust stream detected by Ulysses\footnote{\rm*}{\rm
Based on observations collected at the European Southern Observatory,
La Silla, Chile.}}

\bigskip
\centerline{M. Lindgren and G. Tancredi}
\medskip
\centerline{Astronomical Observatory, Box 515, 751 20 Uppsala, Sweden}

\bigskip
%\bigskip
\bigskip\noindent
{\it
From a project where the vicinity of Jupiter is examined for
new short--period comets, we present observations made at the same time as
the space probe Ulysses passed by Jupiter. The results does not confirm the
existence of an active comet close to Jupiter as suggested by the detection
of a dust stream during the encounter.
Dynamical calculations are also presented, where we show the
possibility that the dust could have a cometary origin.
}

%\bigskip
\bigskip\bigskip
\noindent{\bfS Introduction}

\medskip\noindent
Consider a ``snapshot'' of the inner solar system as in figure 1. The figure
shows the positions, on February 29, 1992, of the planets Earth, Mars and
Jupiter together with 4295 asteroids and 122 Jupiter family comets.
When comparing the positions of the asteroids and the comets
one can make an interesting observation: in the general direction of Jupiter
the only objects between the outer main--belt of asteroids and the orbit of
the planet are comets. That is, if one would like to make a search for new
comets without having the trouble of mixing with the Trojans or the Hildas,
the vicinity of Jupiter is a good place to look. Based on this, maybe a little
na\"\i ve, observation Tancredi and Lindgren initiated a project, using the
ESO Schmidt telescope, where the vicinity of Jupiter is examined for new
short--period comets. In Tancredi and Lindgren (1992) it is shown that
in a rotating--pulsating frame of reference with the instantaneous 
distance Sun--Jupiter as unit distance, and the instantaneous angular
velocity of Jupiter as the rotation velocity, the number density of comets
at any time is clearly at a peak close to Jupiter (figure 2),
and how at the moment of Jupiter opposition it is possible to distinguish
between foreground objects and objects close to Jupiter.

%\bigskip
%[FIGURE 1 \& 2]
%\bigskip
%\vskip 90mm
\vfill\eject

\noindent
The project is now in progress: a first set of 36 plates has been taken, and
the reductions are almost finished. Results will be reported elsewhere
(Tancredi and Lindgren, 1992b).
However, by pure chance, the dates of exposure of the plates corresponds,
within a week, to the moment when the NASA/ESA space probe Ulysses made its
gravity--assist fly--by at Jupiter. This would of course be of no significance
if nothing out of the ordinary happened. But, as reported by Gr\"un et al. 
(1992), something did happen when Ulysses passed by Jupiter. The
spacecraft appeared to pass through a previously unknown stream of dust,
possibly due to a passage through the dust tail of a comet. This coincidence
that the plates in the comet
search project were exposed almost at the same time as an in--situ detection
of a dust stream obviously raises the question wether one on the plates can
actually see a comet at the predicted position.

\bigskip
\noindent{\bfS Observations}

\medskip\noindent
A total of 36 Schmidt plates (hypersensitized IIIaJ emulsion and using
a blue cut--off filter GG385) were obtained with the ESO Schmidt telescope in
March and April 1992. The area covered consists of 3 by 3 regions 
covering a total area of about 16 by 16 degrees centered on
Jupiter as shown in figure 3 and table 1. Four exposures of each region
were made; 3 in March and 1 in April. The March plates were exposed for 90
minutes and the April plates 45 minutes. As suggested by Tancredi and Lindgren
(1992a) the telescope was, during the March observations, tracked with a
speed corresponding to the apparent motion of Jupiter, while the April
observations were done using sidereal tracking speed. 
The seeing was
generally around 1\arcsecfrac .5, and based on exposures of stars of
known magnitude, the limiting apparent magnitude for point sources
was found to be 20.5-21.0 leading to an approximate magnitude limit of 19
for the trails of the slower moving objects (relative velocity to Jupiter
$\sim$ 5\arcsec /h).

%\bigskip
%[FIGURE 3 \& TABLE 1]
%\bigskip
\vskip 75mm

\noindent
The plates were visually scanned by the authors, and among the $\sim$10
objects per sq. deg. that was seen, potential cometary candidates
were selected
according to the criterion that if a trail was shorter than the trails of
the stars (i.~e. objects moving slower than the 18\arcsec /h
relative velocity of Jupiter)
 it was marked and included for further astrometric measurements.
However, {\sl all} trails in the region associated with the Ulysses
encounter were marked as discussed below.
The positions of the trails were then measured using the ESO Optronics
measuring system in Garching and the measuring machine at the Uppsala
observatory.

\bigskip
\noindent{\bfS Ulysses' dust stream encounter}

\medskip\noindent
As reported by Gr\"un et al. (1992) and  Mann et al. (1992, this volume) the dust
detectors on the space probe
Ulysses signalled an enhancement in the sub--micron dust density during
its encounter with Jupiter. The enhancement began on
March 10.4 and ended on March 11.4, with a peak on March 10.92 when the
probe was at a distance of 5.4 AU from the sun and 0.26 AU from Jupiter.
The duration of the encounter and the speed of Ulysses lead to an estimated
extension of the stream along the path of Ulysses of about 800,000 km, and the
estimated relative velocity of the stream is 40--50 km/s.

By knowing the position and velocity of Ulysses and the spin angle
distribution during the dust impacts on the, to the spin axis almost
perpendicularly mounted, detector, the relative velocity vector of the dust
can be estimated. But due to calibration problems the relative speed of the
dust is very uncertain (a factor of 2 according to Gr\"un et al., 1992).
Using a set of different values for the dust velocity, Gr\"un
(1992, private communication) calculated the corresponding heliocentric
orbits that can be seen in table 2.

%\bigskip
%[FIGURE 4 \& TABLE 2]
%\bigskip
%\vfill\eject
%~
\vskip 85mm

\noindent
Using these orbits we then integrated ``hypothetical comets'' back to the
dates of the exposures of the Schmidt plates, and plotted the corresponding
ephemerides. In figure 4 can be seen the projected positions for a comet
encountering Ulysses on March 10.92 and its supposed positions on the
dates of the exposures of the relevant regions
(2/3, 4/3 and 7/3) before the encounter. The labels 0, 10, 20 km/s etc.
correspond to the different values of the absolute velocity of the
dust relative to Ulysses used for the determination of the orbits.

Since what we are looking for is a cometary coma and bearing in mind that
the ejected dust can be present several million km. from the coma, we have
to examine a larger field than the field defined above by the plotted ephemeris
lines.
The area
around the lines out to a distance of 1.5 degrees from the 20 km/s line
%(correspondig to a distance at Jupiter of 17 million km.)
was then closely examined for possible cometary candidates.


\bigskip
\noindent{\bfS Results}

\medskip\noindent
A total of 45 objects were found in the above defined area, none of which
showed any sign of activity. Nor did we find any sign of a dust stream. One
previously known comet (P/Forbes) was during this time in the field (although
at a distance of 3.4 AU), but none of the objects matched the predicted
position of this comet.

Of the 45 objects found, 26 could be identified on three plates and linked
together by the appearance (length, position angle and brightness) of the
trails. Of these we measured the positions of the
13 that were within 0.4 degrees %(5 million km at the distance of Jupiter)
from the area defined in figure 4. This preference for considering only
objects close to the projected position of the dust is due to the simple fact
that the dust tail of a comet at this position (solar elongation $\sim$180\deg)
would point away along the line of sight, and hence we concentrated on a
smaller region where we did not have to consider the projection of an extended
tail.

The objects were then dynamically linked together by using orbit determination
procedures according to Neutsch (1981) after making preliminary identifications
by using so called ``V\"ais\"al\"a--type''
orbits (V\"ais\"al\"a, 1939 and Marsden, 1985). Figure 5 shows a plot of the
positions of these 13 objects, and the orbits can be seen in table 3. Clearly,
none of them are anywhere near Jupiter. The object closest to Jupiter is
object E, which is at a heliocentric distance of only 3.9 AU.

Thus we conclude, based on our observations, that the dust stream detected
by Ulysses did not originate
from a comet brighter than the limiting apparent magnitude of 19 in the
case of a slow moving object (relative velocity to Jupiter
$\sim$ 5\arcsec /h), and in the case of a faster moving object
in a parabolic orbit (relative velocity to Jupiter $\sim$ 25\arcsec /h), not
brighter than apparent magnitude 17.5.

%\bigskip
%[FIGURE 5 \& TABLE 3]
%\bigskip
%\vfill\eject
~
\vskip 80mm

\bigskip\bigskip
\noindent{\bfS Dynamical speculations}

\medskip\noindent
Obviously, the above conclusion does not role out the possibility that the dust
stream is the tail of a comet: it could simply be fainter than the limiting
magnitude of the search. But there is an  argument against a cometary
origin as pointed out by Gr\"un et al. (1992): the high speed of the dust
implies a hyperbolic orbit with perihelion distance of about 5 AU or more,
and, so far, no comet with such an orbit has been detected (Kres\'ak, 1992).
Nevertheless, we would like to point out some facts that may support the theory
for a cometary origin on the following grounds:
the mere fact that the detection of high velocity dust took place very close
to Jupiter and not on the way out to the planet makes one wonder if it
was only by chance or if Jupiter is responsible for accelerating the dust.
It is known that during close encounters between minor bodies and
planets, the osculating orbits can undergo extreme variations on very short
time scales (e.~g. Tancredi et al., 1990), and not represent the orbit of
the object when it is not encountering the planet.

To explore this we have made numerical integrations of test objects with
initial positions at the point where Ulysses encountered the dust stream
and initial velocities divided into two groups: one group where only the
magnitude of the velocity vector is varied according to different values
of the relative velocity to Ulysses (i.~e. the table 2 values). The other group
of velocity vectors were created in such a way that we obtained, for 3
different values of the relative velocity (14, 20 and 45 km/s) sets of 49
vectors (7 by 7 degrees) with different directions centered around the
direction of the velocity vector determined from measurements by Ulysses.

The test objects were then numerically integrated one year backward and
forward in time, recording the osculating orbital elements every day.
The dynamical model
used in the integrations consisted of the Sun, Jupiter and the (massless)
objects. The integrator used was the variable timestep RADAU integrator as
described by Everhart (1985).

The results from the forward integration yielded nothing spectacular for
either of the two groups of initial velocity vectors: all objects continue
on hyperbolic orbits, which was to be expected since the general direction
of the vectors was pointing away from Jupiter and hence no close encounter
occurs. For the backward integrations, however, we found for all of the 
objects that they have made close encounters with Jupiter 1--2 weeks before the
Ulysses encounter, with corresponding perturbations of the orbits. Table 4
summarizes the results for the backward integration for the case where the
direction of the initial velocity vector is the same as the measured one. The
table shows that for relative velocities higher than about 15 km/s the
objects have passed by Jupiter in relatively unchanged orbits. But for 
relative velocities below 15 km/s we see that the orbits before Jupiter
encounter are all elliptic with periods down to values of typical short--period
comets. Their orbital directions remain retrograde, with the exception of the
object corresponding to zero relative velocity, but since this velocity
corresponds to the orbit of Ulysses it should not be considered as realistic
by the simple fact thet it would mean that the dust stream had been around
Ulysses before the detection.

The results from the backward integration of objects starting with different
directions show the same general behaviour as the results in table 4: the
objects with 45 km/s relative velocity  to Ulysses just pass by Jupiter without
much change in their orbits, even though the closest approach is at a distance
of only one Jupiter radius from the surface of the planet. For the objects
starting with relative velocities of 20 km/s we also obtain only hyperbolic
pre--Jupiter encounter orbits; one interesting case, though, may be one orbit
which switched from being a direct orbit before encountering Jupiter to be a
retrograde orbit after the encounter. However, that particular case is one where
the distance to Jupiter at closest approach was less than the radius of the
planet. Finally, for the test objects started with velocity vectors corresponding
to 14 km/s relative velocity to Ulysses we find that some of the objects were in
retrograde elliptic orbits before the Jupiter encounter. It was not
possible to find a case with a direct elliptic pre--encounter orbit. Figure 6
shows examples of the three main outcomes of our test: the possibility to
reproduce orbits that are either elliptic/retrograde, hyperbolic/retrograde
or hyperbolic/direct before encountering Jupiter.

%\vfill\eject
%~
% FIGURE 6 & TABLE 4
\vskip 115mm

%\bigskip
\noindent{\bfS Discussion}

\medskip\noindent
Obviously, these integrations are just a small and
preliminary exploration of a small part of the much larger phase space defined
by the uncertainties in the position and velocity vectors involved. A study of
this kind was made by Tancredi (1990) for the case of P/Helin--Roman--Crockett,
showing the complexity of the dynamics involved.  And to
fully investigate this case one would like to make a thorough analytical study
of the complicated dynamics inherent to this situation where we are dealing
with extreme close approaches, but that is outside the scope of this paper.

To make some kind of conclusion about our numerical investigation we would
like to point out that although a relative velocity of $\lse$14
km/s is a little on the low side considering the estimated 40--50 km/s (but
almost within the errors), the results show that the possibility exists that the
dust originates from a short--period comet which has just been perturbed
into a hyperbolic orbit bound for interstellar space, as opposed to the
tentative explanation given by Gr\"un et al.
(1992) that the hyperbolic orbit indicates an interstellar origin for the dust.

Considering that during the Voyager I \& II encounters with Jupiter there was
a comet (P/Helin--Roman--Crockett) very close to the planet, actually
undergoing a temporary satellite capture (Tancredi et al., 1990), and the fact
that Ulysses has detected a dust stream possibly connected to a tail of a comet,
we would finally like to point out the possibility for future missions to Jupiter
(e.~g. Galileo) to detect new comets during a very interesting phase of their
dynamical lifetime.

\bigskip
\noindent{\sl Acknowledgements.}\quad We would very much like to thank Prof.
E. Gr\"un for giving us very early information about the Ulysses dust stream
encounter, and much appreciated is also the travel funds made available
by the Liljewalch scholarship to participate in this meeting.

\bigskip
\noindent{\bfS References}

\medskip\noindent
Everhart, E., 1985, An efficient integrator that uses Gauss--Radau
spacings. In {\sl Dynamics of comets: Their origin and evolution}
(A. Carusi and G. B. Valsecchi, eds.), p. 185, D. Reidel publishing
company.

\medskip\noindent
Gr\"un, E., Zook, H. A., Baguhl, M., Fechtig, H., Hanner, M. S., Kissel, J.,
Lindblad, B.--A., Linkert, D., Linkert, G., Mann, I., McDonnell, J. A. M.,
Morfill, G. E., Polanskey, C., Riemann, R., Schwehm, G. and Siddique, N.,
1992, Ulysses dust measurements near Jupiter, {\sl Science,} submitted.

\medskip\noindent
Kresak, \v L, 1992, Are there any comets coming from interstellar space?,
{\sl Astron. Astrophys,} {\bf259}, 682.

\medskip\noindent
Mann, I., Gr\"un, E., Baguhl, M., Fechtig, H., Hanner, M. S., Kissel, J.,
Lindblad, B.--A., Linkert, D., Linkert, G., McDonnell, J. A. M., Morfill, G. E.,
Polanskey, C., Riemann, R., Schwehm, G., Siddique, N. and Zook, H. A.,
1992, Measurements with the Ulysses and Galileo dust detectors,
this volume.

\medskip\noindent
Marsden, B. G., 1985, Initial orbit determination: The pragmatist's point
of view, {\sl Astron. J.,} {\bf90}, 1541.

\medskip\noindent
Neutsch, W., 1981, A simple method of orbit determination, {\sl Astron.
Astrophys.,} {\bf102}, 59.

\medskip\noindent
Tancredi, G., 1990, The capture of P/Helin--Roman--Crockett by Jupiter, in
{\sl Nouveaux developpements en planetologie dynamique,} (D. Benest,
C. Froeschl\'e and J. M. Petit, eds.), Nice, p. 273.

\medskip\noindent
Tancredi, G., Lindgren, M. and Rickman, H., 1990, Temporary satellite capture
and orbital evolution of comet P/Helin-Roman-Crockett, {\sl Astron. Astrophys.,}
{\bf239}, 375.

\medskip\noindent
Tancredi, G. and Lindgren, M., 1992a, The vicinity of Jupiter: a region
to look for comets, in {\sl Asteroids, Comets, Meteors 1991}
(A. Harris and E. Bowell, eds.), Lunar and Planetary Institute,
Houston, in press.

\medskip\noindent
Tancredi, G. and Lindgren, M., 1992b, A search for comets encountering
Jupiter, in prep.

\medskip\noindent
V\"ais\"al\"a, Y., 1939, Eine einfache Methode der Bahnbestimmung,
{\sl Mitteilung der Sternwarte der Universit\"at Turku Nr. 1}, Annales
Academiae Scientiarum Fennicae. Ser. A. Tom. LII, No. 2, Helsinki.


\bye