Table 3 (click here) and Table 4 (click here) show the parameters derived from the present data. When possible the modelfitting was done independently of the image plane by means of the AIPS task JMFIT and also using the visibilities using the task modelfit of the Caltech-Difmap package. Generally the solutions found were in good agreement, and we used the differences in the results provided by the two methods as an estimate of the errors on the parameters reported in the tables. The component labelling is reported on the images for those sources were multiple components have been fitted.
| Name | Comp. |  |  | Name | Comp. | | |
| 10-5erg/cm3 | mGauss | 10-5erg/cm3 | mGauss | ||||
| (1) | (2) | (3) | (4) | (1) | (2) | (3) | (4) |
| 0201+113 | > 41 | > 67 | 1629+120 | C1 | > 7.2 | > 28 | |
| 0237-027 | > 10 | > 33 | C2 | 0.15 | 4.1 | ||
| 0237-233 | C1 | 1629+680 | C1 | 10 | 33 | ||
| C2 | C2 | 0.34 | 6.0 | ||||
| 0500+019 | C1 | > 5.9 | > 25 | 1801+010 | C1 | > 32 | 59 |
| C2 | 2.9 | 18 | C2 | 3.0 | 18 | ||
| 0511-220 | C1 | > 8.9 | > 31 | 1848+283 | > 47 | > 71 | |
| C2 | 0.42 | 6.7 | 2044-027 | > 0.95 | > 10 | ||
| 0743-006 | > 14 | > 39 | C2 | 0.03 | 1.6 | ||
| 0922+005 | > 12 | > 36 | 2126-158 | > 25 | > 51 | ||
| 0941-080 | C1 | 0.02 | 1.6 | 2128+048 | C1 | > 9.9 | > 33 |
| C2 | 0.03 | 1.7 | C2 | > 2.4 | > 16 | ||
| 1143-245 | > 17 | > 43 | C3 | > 2.7 | > 17 | ||
| 1237-101 | > 6.2 | > 26 | C4 | 0.08 | 3.0 | ||
| 1317-005 | > 2.6 | > 17 | 2137+209 | C1 | > 8.7 | > 31 | |
| 1402-012 | > 33 | > 60 | C2 | 0.06 | 2.5 | ||
| 1502+036 | > 1.9 | > 14 | 2210+016 | C1 | > 1.7 | > 14 | |
| 1518+047 | C1 | 3.2 | 19 | C2 | 0.83 | 9.5 | |
| C2 | 0.94 | 10 | C3 | 0.89 | 9.8 | ||
| 1602+576 | C1 | > 50 | > 73 | C4 | 0.12 | 3.7 | |
| C2 | 19 | 45 | 2223+210 | > 20 | > 46 | ||
| C3 | 0.63 | 8.3 | 2351-006 | C1 | > 1.9 | > 14 | |
| C4 | 0.37 | 6.3 | C2 | 0.36 | 6.2 | ||
| 1607+268 | C1 | > 4.7 | > 23 | ||||
| C2 | 0.39 | 6.5 | |||||
In Table 5 (click here) we summarize the physical parameters of the sources as
derived from the assumption of equipartition conditions within the
radio emitting region. We further assumed 
as index of the
energy distribution of the relativistic electrons (corresponding to an
optically thin spectral index of 0.8, in a region not affected by
synchrotron ageing), a filling factor 
and particle energy
equally shared between electrons and protons (k=1), integrating the
radio spectrum between 10 MHz and 100 GHz, using the relationship by
Miley (1980).
When unknown, we arbitrarily used z=1. We did not apply the
"k-correction'' to take into account the values in the rest frame,
since the instrinsic values are not discussed.
Finally the parameters are derived at the most convenient frequency
(either 2.3 or 8.3 GHz) where the radio emitting component is better
described (the former for resolved components, the latter for the
unresolved regions). Generally the equipartition magnetic
field is of the order of 
, while the minimum
energy density is about 
. These figures
are higher than those found in CSS sources, and are similar to those
derived for the GPS sources with the highest turnover frequencies.
This is partly due to the compactness of the majority of the
components detected here; in fact the present data are more favorable
to the detection of high brightness compact regions.
All the six galaxies in the sample show some structure, which can always be interpreted in terms of a double (often asymmetric) or a triple source. Quasars instead are mostly found to be represented by a single component and the incidence of significant VLBI structure (core-jet or double) is limited to 8 objects out of 22. This might also be due to the low dynamic range of our images which prevents the detection of low surface brightness regions in objects with a very bright component or very asymmetric radio emission. In general the radio structures we observe are consistent with those seen by Stanghellini et al. (1997) for their sample of powerful GPS sources.
Most of the sources presented here can be considered good GPS candidates, since the properties we were able to study are those typical of the CSS/GPS class. Flux density variability has been inferred from literature data for a few candidates (all quasars), but the increase/decrease was never enough to exclude the candidate. We note that the optically thin spectral index is flatter than 0.5 in 0922+005, while a flattening at mm wavelengths is likely to occur in 0201+113 and 0237-027. On the other hand, a number of sources have ill defined radio spectra due to the lack of data; in particular, further multifrequency flux density measurements would be required for 0237-027, 0941+261, 1402-012, 1502+036, 1602+576, 1801+010 and 2351-006.
Acknowledgements
We thank the European VLBI Network and the personnel at the geodetic stations of Matera and Wettzell. This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. DD and MB acknowledges the Commission of the European Union for the award of a Fellowship under contract ERBCHBGCT920212. DD acknowledges the Joint Institute for VLBI in Europe (JIVE) and the National Foundation for Research in Astronomy (Dwingeloo, NL) and MB acknowledges the Nuffield Radio Astronomy Laboratories (Jodrell Bank, UK) for hospitality during the fellowship. DD and MB acknowledge the Italian Space Agency (ASI) for support under contract CNR/ASI ARS 96-13. We wish to thank Richard Schilizzi for the encouragement and the referee, Chris O'Dea, for insightful comments on the manuscript.