Plots received from Paul Chodas, 4 May 1994:


  trainangle_vel.ps  Plot of the angle between the nuclear train and the
                     velocity vector of its midpoint, vs. time
  trainangle_rad.ps  Plot of the angle between the nuclear train and the
                     radial vector from its midpoint to Jupiter, vs. time

  Although both plots show only the angle in the orbit plane, the out-of-
plane angle is very small.  The angle in the first plot is positive when
the train is between the velocity vector and the "cross-track" vector,
which is perpendicular to the velocity vector, in the orbit plane, and
pointing outwards (eg. pointing away from Jupiter at apojove).  The train
is perpendicular to the velocity vector on July 11, 1993, but aligns itself
to within 1 degree of the velocity vector from mid-December 1993 through 
impact.

  The angle shown in the second plot is the angle, measured at the center 
of the train, from the radial direction away from Jupiter (R) towards
the negative T direction, where T is the "transverse" vector in the RTN
frame (in the orbit plane, perpendicular to R, and such that R x T = N,
the positive orbit normal).  If one is sitting on the midpoint of the
train, facing away from Jupiter with the orbit plane horizontal, then
fragment W is always off to the right by the angle shown in the plot.
The plot shows that the train has been less than 7 degrees away from
the radial direction since discovery.

============================
posted 5 May 94, acr.