One possibility is that galaxies merge on bound, eccentric orbits rather than the parabolic orbits assumed in the fiducial models. A bound orbit would slow the encounter velocity at perigalacticon and may result in the formation of more prominent tails. While it is unlikely that any galaxy pairs which might form on tightly bound orbits would survive to the present time, it seems at least plausible that much wider encounters than those studied here may decay into a bound orbit for the second, much closer passage. Dynamical friction after this second passage would then result in a rapid merger, like those modeled in .
To explicitly test the evolution of tails in a bound encounter, we have
set up a merger of Model D galaxies on a bound orbit. We choose a
prograde orbit with 
and eccentricity 
. At perigalacticon,
the relative velocity of the galaxies is 
, considerably slower
than the 
expected for zero energy parabolic orbits.
For comparison, the relative velocities for zero energy orbits in Models A
and B at 
are 
and 
(Figure 2).
The results of this simulation are shown in Figure 9.
The tidal tails which form initially, quickly fall
back into the merging pair; whatever gain was achieved by slowing the
interaction down was offset by the deep potential wells, and again the
tails are quite anemic and very short-lived. Even slower, more tightly
bound orbits, therefore, are not likely to improve the situation and,
furthermore, seem astrophysically unreasonable.
Figure 9: Merger of Model D galaxies on a bound orbit.