Table 2 (click here) gives the differences in the FK5(J2000) system with the VLBI radio
positions published by Johnston, in the sense radio minus this work (CL).
With the exception of the source 1236-684 (see previous section), the
comparison shows a good overall agreement, consistent with the declared
precision of the results being compared. Two other sources (0454-463 and
2314+038) for which we give an optical position in Table 1 (click here), have been declared
unsuccessfully observed by Johnston. As might be expected, a comparison of
their a-priori radio positions given in the NRL/USNO radio/optical survey
working list (Fey 1994) with our optical data gives very high residuals, and
therefore were not included in Table 2 (click here). For both objects optical positions
have been obtained by Jauncey, in the FK4 system. In the case of 0454-463 the
optical position given by Jauncey (transformed to the FK5, J2000 system) is in
good agreement with ours. As pointed out by Johnston, the radio counterpart
was apparently correlated at the wrong position. As for 2314+038, our optical
position is not consistent with Jauncey's.
Table 2: Comparison in the FK5 system with the VLBI radio positions given by
Johnston et al. (1996)
Combining the results presented in Table 2 (click here) with analogous results
published
in Papers II and III of our series, a comparison subset of 45 objects with the
VLBI positions given by Johnston is obtained. After the omission of the
sources
1116+128 and 1236-684, which have (radio - CL) residuals more than 3 times the
standard deviation of the average residuals of the subset, the following
statistics are obtained:
n = 43
.
The 
and 
(radio - CL) residuals
for the 43 sources are plotted as a function of the coordinates in Figs. 3 (click here)a-d. Quick inspection of Figs. 3a and 3b, which illustrate the dependence of
on both coordinates, shows the existence of a
general deviation in RA between the radio reference frame, represented by the
VLBI radio system established by Johnston, and the optical reference frame
represented by our optical positions of CERS referred to the FK5 system
through the intermediary of the IRS catalogue. This result confirms what was
reported in Paper III on the basis of a smaller subset of comparison objects.
Our result is consistent with recent results by Argyle et al. (1996),
hereafter Argyle who, based on optical observations of CERS and radio stars
measured in the system defined by the annual series of Carlsberg Meridian
Catalogues, also report the existence of a similar deviation in RA (
south of the equator) between the VLBI frame and the optical frame of the
FK5 implicit in the Carlsberg system.
Figure 3: a-d) Residuals in the sense (radio - CL), plotted as a
function of the
coordinates, for the 43 sources in common between the VLBI survey of
Johnston
et al. (1995) and the optical survey of Costa & Loyola (Papers II, III and
IV).
See Sect. 5 for details
As shown by Argyle, roughly half of this deviation is attributable to an offset in the origin of RA between the radio and optical reference frames, the remnant probably being intrinsec to the Carlsberg system itself. Although with the precision of our data, and the small subset of comparison sources available, it is not possible in our case to disentangle the contributions to the general deviation in RA found by us, it is reasonable to suppose that it is also due in part to the existence of an offset in the origin of RA between the radio and optical frames as defined at present.
On the contrary to what is reported by Argyle, we do not find evidence for
a general deviation in declination between the radio and optical frames. As
shown by Figs. 3 (click here)c and 3 (click here)d, which ilustrate the dependence of 
on both
coordinates, the only obvious effect is a local deviation in the 
to
zone seen in the relation 
vs. 
.
Acknowledgements
We are indebted to Cerro Tololo Interamerican Observatory for donating the plate material that made possible the continuation of our survey. We are also indebted to Prof. C. Anguita for his interesting comments, and to M. Wishnjewski for carefully measuring many sets of plates. This work was partially financed by the Fondo Nacional de Investigación Científica y Tecnológica (proyecto 1930784, Fondecyt).