The 10-kpc "ring'' of M 31 has an angular diameter along the major
axis of 
while the beamsize of a single VLA telescope
(the primary beam) at 
20 cm is only 
. To cover the
whole area, observations of 7 fields ("pointings'') with the VLA-D
array were collected during 9 years (Table 1). The angular resolution is
45
. As the upgrade of the L-band systems took place between
1990 and 1993, only the last 3 pointings could profit from the
improved sensitivity. Unfortunately, at the same time the
interference situation became worse so that the second IF at 1515 MHz
could not be used anymore. The rms noise in the final maps varies
between the different pointings due to different observation times,
different system sensitivities and a different amount of data
affected by interference. In linear polarization, the rms noise is
almost the same for all pointings.
All data were reduced at the MPIfR Bonn using the standard AIPS
package. 3C 48 and 3C 138 were used for absolute flux density and
polarization angle calibration, 0038+328 for gain and phase calibration.
Self-calibration was performed on each field individually using
the AIPS task ASCAL. Comparison of the total flux densities of
unresolved background sources in overlapping regions of two (or more)
fields yielded good consistency within a few percent, except for field A
(the first one observed). Here all flux densities were found to be
higher by 
, for which the maps in Stokes parameters
I(total intensity), Q and U were corrected.
For each Stokes parameter the maps of all fields were combined by
applying the AIPS task LPTES which is similar to the standard LTESS,
but cuts the individual maps at a chosen radius from the field centre.
We chose a radius where the sensitivity is down to 
in order
to avoid a large increase of the noise level at the edges of the
combined map. The combined maps in Stokes Q and U were transformed
into maps of the polarization angle and of the polarized intensity
by taking 
where
(Wardle & Kronberg 1974).
The resulting maps in total and linearly polarized intensities are
shown in Figs. 1 (click here) and 2 (click here), respectively. The rms noise within the main
regions of M 31 is 
in total power and 
in polarized intensity, increasing to about twice these
values at the singly-covered map edges. Hence these maps represent the
most sensitive radio continuum survey of M 31 so far and the first one in
linear polarization with the VLA.
Figure 1: Mosaic of 7 fields on M 31 showing the total power emission
at 20.5 cm observed with the VLA-D array. The angular
resolution is 45 . The rms noise generally is
increasing to 
at the borders of the map
Figure 2: Mosaic of 7 fields on M 31 showing the linearly polarized
emission at 20.5 cm observed with the VLA-D array. The
angular resolution is 45 . The rms noise generally is increasing to 
at the borders of the map. Point
sources brighter than 1 mJy in polarized intensity are subtracted. The
position of the bright "ring'' is shown by contours of the total
emission (see Fig. 4). The contour levels are
0.6, 1.2, 1.8, 2.4, 3.0 and 6.0 mJy/beam
The largest structure visible to the VLA-D array at
20.5 cm is 
due to the lack of short
interferometer spacings. To fill these short spacings, observations
at 1400 MHz (21.4 cm) were performed with the 100-m
Effelsberg telescope of the MPIfR between 1993 and 1995. The
bandwidth used was 20 MHz. The system temperature was 26 K and
the angular resolution 9
35. A field of 
in M 31 coordinates (
position angle of the major
axis) was scanned 2 times in - and 3 times in
-direction, yielding a noise level of 
. The
scanning effects were removed by a weighted addition of the maps
with different scanning angles in the Fourier plane (see
Emerson
& Gräve (1988) for explanation of the algorithm).
M 31 is located on a spur of Galactic radio continuum emission (Gräve et al. 1981). Subtraction of this foreground emission in total intensity was achieved by a careful application of the BGF algorithm (Sofue & Reich 1979) after subtraction of known background point sources.
The final map in total intensity is shown in Fig. 3 (click here). Unfortunately, at this low angular resolution the polarized emission from M 31 cannot be separated from that from the Galactic foreground.
Figure 3: Total emission (contours) and E-vectors of the polarized
emission at 
cm of a field
containing M 31 observed with the 100-m Effelsberg telescope. The
angular resolution is 9 35. The rms noise is about
5 mJy/beam. Contour levels are 20, 40, ..., 100, 160, 200, 300, ...,
600 mJy/beam. The maximum polarized intensity
of 121 mJy/beam corresponds to a vector plotted with 11
\
length
