The prominence was scanned with 
line from the left to the
right of the filtergrams of Fig. 1 (click here), and the slit was parallel to
the left edge of each filtergram. The scanning time was 0927UT to 0930UT.
Combining the spectra and 
filtergrams, we can
infer
that the materials in the left leg of the eruptive prominence spiraled to rise.
This is consistent with the results obtained in section two.
The spectra obtained were measured with the PDS microdensitometer
at the Purple Mountain Observatory. Comparing the spectra with the
filtergrams, we can locate the position of each
line in the filtergrams of the prominence. After calculating the
center-of-gravity position (
) of each chosen line profile and
then the Doppler shift 
(
, 
), we
obtained the velocity along the line of sight 
(
).
Figure 4 (click here) gives
the distribution of the line-of-sight velocity, which indicates that the
velocity increased from the bottom of the left leg to the top. After the
materials in the left leg reached the climax, those dropped out in succession
and did not spiral.
Figure 4 (click here)b summarises the configuration of the velocity
distribution. The materials are usually frozen in the magnetic field of a
prominence, and they move mostly along the magnetic lines. So we think of
the magnetic field of
the prominence helical, stretched primarily along the axis, and
tied to the photosphere only at the end of the prominence. There is
growing
evidence that filaments (or prominences) emerge from the photosphere already
twisted (Rust & Kumar 1994; Yang et al. 1988;
Liggett & Zirin 1984; Vrsnak et al. 1991). The
prominence motion shown in Fig. 4 (click here)b may be called "right-hand
screw" when it is viewed by a hypothetical observer in the chromosphere.
Rust & Kumar (1994) infer that the magnetic fields of
prominences are twisted in a systematic way: right-hand helices predominate
in the south and left-hand helices predominate in the north. The result
obtained by spectral analyses confirms their inference.
The velocity increased perhaps because (1) the angular velocity increased from
the bottom to the top due to the release of the torsion of the prominence;
(2) the angular velocity may be constant, but the radius of the leg increased
from the bottom to the top.
Figure 4: a) Distribution of the velocity (in km/s) along the line
of sight. b) Sketch for explanation of the velocity distribution.
The arrow shows the direction of the material motion