The data consists of a large number of constant declination scans
(antenna temperature profiles) 24-hour long, and separated approximately
by
. Many of the effects that mar the data can
be readily recognized, and often corrected, because of their short
time scale (e.g. solar bursts, radio stations). Meteorological effects
like rain and fog can be eliminated simply by not observing under
those conditions. Some ionospheric effects produce smooth fluctuations
of the order of ten minutes and can be present for hours. These
effecs cannot be corrected so the affected data are rejected. Other
ionospheric effects result in slow variations of the incoming signal
with a time scale of many hours, and we attribute them to changes in
the opacity of the ionosphere. As a result of the combined effects
just discussed the temperature profiles observed for a given
declination are not quite repetitive. In order to minimize the
ionospheric absorption, for each pointing we selected the best data
and drew the envelope to them; this envelope was then accepted as the
true profile for the corresponding declination. Similar procedure have
been followed by
Yates & Wielebinski (1966) and
Briddle & Purton
(1968).
As it is explained
in detail elsewhere (Alvarez et al. 1994), contrary to what we had
assumed at first from a study of the antenna pattern near the zenith,
we found by precise measurements that the separation between
beams is not constant at 
, but that it increases with
zenith distance up to 
at 
(
).
In matching the data
by overlapping end-beams of adjacent positions we adopted the criterion
of averaging the temperatures measured by the two beams, and of
asigning that value to the average of their declinations. In this way
we tried to minimise the scanning effects in the final map. In all we
used 58 different declinations.
We also found that the azimuth of the N-S axis of the array (measured
from north towards the east) is 
, and that the plane of the
array is inclined 
toward the east (Alvarez et al. 1994). These
parameters were used to correct the observed right ascensions. The
data were also corrected by side lobe effects; this was accomplished
by computing theoretically the array radiation pattern, and by
complementing
this survey with
observations made by Maeda (1994) at 46 MHz and north of 
.
Many scanning effects were eliminated by
selecting the best data and by reobserving several times the conflicting
positions.
Residual scanning effects were corrected as much as possible by
the method of unsharp masking (Sofue & Reich 1979).
The extragalactic radio sources were not
removed.
We estimate the error in the
temperature scale at 10% or better.