The survey consists of almost 8000 spectra which were taken
between February and November 1984. The area observed covered
and 
.
Spectra were obtained, both in longitude and latitude, every 
(0.85 beam-widths) for 
, and every 
(1.7
beam-widths)
for 
.
Each spectrum was integrated for about 10 minutes, to yield an antenna
temperature noise level of 
K (rms) for a 1.3 km s-1
velocity
resolution (Table 2 (click here)). In order to subtract instrumental effects and the
sky background the antenna was switched, every 15 seconds,
between the target position (on) and an emission-free position
(off) at elevations closer than 
. The off positions were verified
to have
emission not greater than 0.04 K (Bitran 1987).
Table 3 (click here) lists the emission-free
positions.
Observations were carried out only under good weather conditions, i.e. when the zenith opacity of water vapor at 115 GHz was between 0.05 and 0.20. During the observations the spectra were examined visually and fitted with linear baselines only. If a spectrum showed evidence of baseline distortion it was discarded and the observation repeated.
Figure 2: Spatial maps of CO line intensity 
integrated over the velocity range of -320 to 320 km
s-1 displayed at three different contour levels. The contour intervals
and the lowest contours have been set at 13, 32.5 and 65 K km s-1,
corresponding to 
and 
,
in the top, middle and
bottom maps, respectively
Figure 5: Two longitude-velocity diagrams integrated below and
above the galactic plane. The contour interval is 2 K, with the lowest
contour at 2 K. The arrows point to two intense sources with
unusually large velocity widths
Within 
of the galactic center the
velocity span of the CO emission exceeded the full range of the
spectrometer. This problem was circumvented by taking, for all positions
within
, two spectra centered at velocities 306 km
s-1 apart (Table 2 (click here)).
These two spectra covered a combined
range of 640 km s-1 with an overlap of 26 km s-1 (20
channels) allowing ample emission-free sections to set baselines,
and enough overlap to check that the spectra matched properly (Fig.
1 (click here)). The
emission level in the overlapping channels was checked to coincide in both
spectra within the antenna temperature rms noise. Poor matchings were
rejected and the corresponding positions re-observed. A sample of combined
spectra along two adjacent longitude strips are shown in Fig. 1 (click here); the high
signal-to-noise ratio and the flat baseline of these spectra are
characteristics of the whole survey.
An important goal of this survey was to cover the full latitude
extent of the galactic CO emission. We found most of the emission
to be confined in , therefore our latitude
coverage (
) seems adequate. However, because HI
emission is observed at higher latitudes in opposite galactic
quadrants (at 
for 
, and at 
for 
), we conducted a search between 
and
for , and between 
and 
for , detecting only local emission.