On the basis of the different radio properties (especially different radio morphologies) between GPS galaxies and GPS quasars, I suggest that GPS galaxies are a more homogeneous class mainly consisting of compact symmetric objects, while the majority of GPS quasars are an intrinsically different type of radio source, similar to the common flat spectrum radio sources. For CSS radio sources the difference between galaxies and quasars is less evident. This is interpreted as a bias correlated with size: samples of smaller objects are more likely to be contaminated by unrelated objects. This is consistent also with the high percentage of blazar/quasar objects present in a sample of the even smaller HFP radio sources.

Radio observations of linear polarisation provide a useful probe of the physical conditions in the gaseous environments of compact steep spectrum (CSS) sources. Here, we present polarimetric VLBA observations at 8.4 GHz of the CSS quasar 3C 43. The jet in this source curves gently up to about 220 mas from the core, at which point it bends sharply, probably due to a jet–cloud encounter. Regions of polarised emission are detected along the jet, while the core component shows no significant polarisation.

III Zw 2 shows dramatic radio outbursts roughly every five years. Here we present the light curves and VLBA observations of the latest flare with an excellent time sampling. We have detected superluminal motion with a lower limit for the apparent expansion speed of 1.25±0.09 c at 43 GHz. Spectral and spatial evolution are closely linked and the evolution of III Zw 2 in a turnover frequency vs. linear size diagram is similar to the correlation for GPS and CSS sources. Before and after this rapid expansion we observe a period of virtually no expansion. However, at 15 GHz III Zw 2 shows a constant slow expansion (∼0.6 c). The difference is qualitatively explained by optical depth effects in an ‘inflating-balloon model’, describing the evolution of radio lobes on an ultra-compact scale. The stop-and-go behaviour could be explained by a jet interacting with a molecular cloud or the molecular torus.

The study of both neutral and ionised gas in young radio sources is providing key information on the effect the radio plasma has on the ISM of these objects. We present results obtained for the compact radio sources PKS 1549–79, 4C 12.50 and PKS 1814–63 and for the intermediate-size radio galaxy 3C 459. At least in the first two, low ionisation optical emission lines and HI absorption appear to be associated with the extended, but relatively quiescent, dusty cocoon surrounding the nucleus. The [OIII] lines are, on the other hand, mostly associated with the region of interaction between the radio plasma and the ISM, indicating a fast outflow from the centre. A case of fast outflow (up to ∼1000 km s-1) is also observed in HI in the radio source 4C 12.50. As the radio source evolves, any obscuring material along the radio axis is swept aside until, eventually, cavities (of the same kind as observed e.g. in Cygnus A) are hollowed out on either side of the nucleus. We may witness this phase in the evolution of a radio source in the radio galaxy 3C 459.

We report results from nearly simultaneous pentachromatic VLBI observations towards a nearby GPS galaxy NGC 1052. The observations at 1.6 and 4.8 GHz with the VSOP, and at 2.3, 8.4, and 15.4 GHz with VLBA, provide linear resolutions of ∼0.1 pc. Convex spectra of a double-sided jet imply that synchrotron emission is obscured through foreground cold dense plasma, in terms of free–free absorption (FFA). We found a central condensation of the plasma which covers about 0.1 and 1 pc of the approaching and receding jets, respectively. A simple model with a geometrically thick plasma torus perpendicular to the jets is established to explain the asymmetric distribution of FFA opacities.

Data from new Australia Telescope Compact Array observations of a nearby gigahertz peaked spectrum radio source, PKS 1718–649, show significant variability both below and above the spectral turnover frequency during a nine month period. Simple models of synchrotron self-absorption (SSA) and free–free absorption (FFA) cannot explain the individual spectra or the spectral variability. Novel SSA or FFA models will need to be investigated in order to explain the complex variability in this source. Some potential explanations for the spectral variability are explored here but we can make no strong conclusion about which of SSA or FFA (or alternative mechanisms) causes the peaked spectrum in PKS 1718–649.

Based on the free–free absorption (FFA) model of gigahertz peaked spectrum (GPS) sources, we explain both Faraday rotation asymmetry and opacity asymmetry consistently between two components of a GPS quasar 2134+004. The FFA model assumes dense plasma around the central core to produce FFA, and the difference of path lengths in the plasma toward each component could cause these asymmetries. The component that is closer to the observer has a shorter path length, and consequently, smaller opacity of FFA and lower Faraday rotation. In a simple case, the ratio of Faraday rotation between two components is a function of the ratio of path length, and is the same as the ratio of opacity. Then these two ratios are shown to be essentially the same by our observations. We could thus distinguish between near-side and far-side components by the asymmetries.

We present initial results from a program that we have begun, aimed at increasing the number of GPS and CSS sources for which good spectroscopic data are available for a variety of emission lines. We combine new observational data with previously published results to investigate whether there are any relationships between the excitation state of the gas and the radio properties of the sources, retaining distinctions between GPS and GSS galaxies and quasars. In scenarios where the youngest, most compact sources interact most strongly with the gas, it is perhaps to be expected that this can have an effect on the overall excitation state of the gas, which could therefore correlate with radio properties such as turnover frequency that plausibly trace the source size and age. However, we find a degree of scatter between the turnover frequency and gas excitation state that is several times larger than the characteristic degree of uncertainty in our line ratio measurements. This would suggest that either there is no underlying relationship between turnover frequency and excitation state, or alternatively that any potential relationship is completely diluted by variations in the intrinsic properties of the gas from one source to the next, which would thus be an important factor in determining the source properties.

Flux density monitoring data at 2.3 and 8.4 GHz is presented for a sample of 33 southern hemisphere GPS sources, drawn from the 2.7 GHz Parkes survey. This monitoring data, together with VLBI monitoring data, shows that a small fraction of these sources, ∼10%, vary. Their variability falls into several categories: sources whose spectral classification is, at best, ephemeral on a timescale of years; sources with a stable GPS classification that vary, but retain their classification; and a small number of sources that exhibit interstellar scintillation, but that maintain a mean GPS spectrum. Existing data on GPS sources with higher frequency peaks, ≥3 GHz, reveals that many such sources vary. However, the majority of these sources possess a GPS spectrum only during outbursts, and hence should perhaps be classified as ephemeral GPS sources. In addition, significant levels of circular polarisation have been found in a number of GPS sources, both amongst the variables and non-variables. Remarkable amongst these is PKS 1519–273, which possesses strong and variable circular polarisation, and which exhibits IDV in all Stokes parameters.

X-ray observations of PKS 1934–63 and S5 1946+708 have been made with Beppo SAX lasting 100 and 40 ksec, respectively. Both sources were detected, and in both there is evidence (at 90% confidence) of a strong iron Kα line which would indicate that the nuclei are surrounded by Compton thick material (˜1 g cm−2).

For 1946+708 VLBI H I absorption data are available from the literature. With the condition that free–free absorption should be modest for the source to be observable a minimum radius of around 20 pc is derived for the absorbing torus. The torus is predominantly molecular with a density of 108 cm−3. The corresponding pressure is rather high. Observations with XMM have been requested to verify the large equivalent width of Kα with a better confidence level.

In 1934–63 no important H I absorption was previously detected and no nucleus has been seen in radio data at 8 GHz. If this is due to free–free absorption a molecular torus with a radius less than 5 pc could be responsible. The density would be 108 cm−3 or more.

The relatively low S/N ratio of our observations does not allow a fully convincing conclusion to be reached. But our discussion shows that the combination of X-ray and high resolution 21-cm absorption data can provide important information on the physical parameters in the absorbing ‘torus’.

Paper to appear in full in Astronomy and Astrophysics.

We report very long baseline array (VLBA) observations at 2.3, 8.4, and 15.4 GHz towards nine gigahertz peaked spectrum (GPS) sources. One Seyfert 1 galaxy, one Seyfert 2 galaxy, three radio galaxies, and four quasars were included in our survey. We obtained spatial distributions of the free–free absorption (FFA) opacity with milliarcsecond resolution for all sources. It is found that type 1 (Seyfert 1 and quasars) and type 2 (Seyfert 2 and radio galaxies) sources showed different distributions of the FFA opacities. The type 1 sources tend to show more asymmetric opacity distributions towards a double lobe, while those of the type 2 sources are rather symmetric. Our results imply that the different viewing angle of the jet causes the difference of FFA opacity along the external absorber. This idea supports the unified scheme between quasars and radio galaxies, proposed by Barthel (1989).