The B3 radiosource catalogue (Ficarra et al. 1985),
lists 13354 sources brighter than 100 mJy at 408 MHz and covers 0.78
steradians in a sky strip from to
, all hour angles.
From the B3, down to the catalogue limit of 0.1 Jy,
Vigotti et al. (1989) selected five
complete subsamples separated by equal increments in
logarithmic flux density and containing approximately the same number of
sources, through the choice of different declination limits for each subsample.
Furthermore, the right ascention range was restricted to exclude sky areas
at low galactic latitude (R.A.
and
).
From VLA maps in C configuration at 1465 MHz for the 1103 sources in the sample, Vigotti et al. (1989) obtained radio positions (accurate to 0.5 arcsec rms) which allowed the optical identification on POSS-I prints based on positional coincidence. The analysis of these maps has also led to the exclusion of 53 sources from the sample because they were either lobes or because some sources joined to form a single source.
These sources were reobserved at the same frequency with the VLA in A configuration allowing a better resolution, i.e. HPBW 1.4 arcsec to be compared with 14 arcsecs of C configuration. A-configuration VLA maps will be presented elsewhere (Vigotti et al. 1996, in preparation).
The final radio sample contains 1050 radiosources. Table 1 (click here) shows the sample definitions, the number of sources and the solid angle covered in each subsample. All the sky area limits are at 1950.0 epoch, except the declination limits for sample 4, which correspond to the whole B3 catalogue, and are referred to 1978.0 epoch. The solid angles are corrected for the incompleteness regions of the B3 survey.
Table 1: Definition of the B3-VLA sample
Out of the original 352 identified objects, described in Vigotti et al. (1989), 183 were quasars candidates, defining a quasar candidate on the basis of its star-like appearance, regardless of the color, to avoid color biases.
These were examined
using the positional coincidence between the optical and radio
positions and the structural informations coming both from the A-configuration
and C-configuration VLA maps.
In order to build a complete sample of quasar candidates, we have accepted also
identifications which are formally at low probability.
For the unresolved sources the radio-optical rms
displacement is and the
radius is
,
but we decided to extend the search up to 6''. Infact, examining
the radio-optical
displacement histogram (see Fig. 1 (click here))
for the 57 unresolved (in C-configuration)
spectroscopically confirmed quasars, we see that the
choice
would have
implied a loss of
of the unresolved sources (4/57) with a
displacement greater than
.
A posteriori looking the A-configuration
maps we understand why the distribution is not gaussian. In all the four
cases the quasar was coincident with the faint component of a double
radiosource of 4-5 arcseconds diameter,
and with a very high flux ratio between
the two radio components.
Furthermore, we obtained CCD images with the 1.5 meters Loiano Telescope of Bologna Observatory or with 3.5 meters telescope at Calar Alto of all the identified objects with uncertain optical classification from plates in the original sample, in order to be sure not to have missed any star-like object.
From the 183 quasar candidates in Vigotti et al. (1989) we
excluded:
(a)five candidates with (0019+391, 0800+399,
0800+472, 0805+406 and
1357+392).
(b)three candidates which from the A-configuration
VLA maps turned into empty fields
(0209+390, 1012+389 and 1317+389).
(c)three candidates which revealed to be plate-flaws from CCD images
(1033+408, 1258+395 and 1301+393).
Figure 1: Distribution of the radio-optical displacements for 57 unresolved quasars
The candidates magnitudes in Vigotti et al. (1989) were estimates on
POSS-I enlargements obtained by comparison with reference star images in
the Selected Area 57. The accuracy of those
estimated R magnitudes is 0.5 m. As machine measured R magnitudes are
available from the APM Catalogue (Irwin 1992) which have an rms
error of , we list these in Table 4 (click here).
There is a good agreement between the two magnitude scales
(see Fig. 2 (click here))
with a 0.024m zero point.
In two cases (0143+446b, 1342+389a)
we adopted the estimated
magnitudes because APM was clearly wrong (they were giving the magnitude of
the quasar plus a star few seconds apart).
The sample is complete down to POSS-I limit in red, nominally 20.0;
however a plate to plate variation of 0.25 magnitudes could affect the
sample.
Figure 2: B3 magnitudes versus APM magnitudes