3.1. Extended Emission
An important question is whether the CSS and GPS sources are "complete" and compact sources in their own right, or whether they are just the inner parts of much more extended sources as are the "steep-spectrum cores" in 3C 236 and 3C 293 (Schilizzi et al. 1988; Bridle, Fomalont, & Cornwell 1981). Several of the sources in the Fanti et al. 3C + PW sample are now known to have larger scale structure (3C 216, Barthel, Pearson, & Readhead 1988, van Breugel et al. 1992; Taylor, Ge, & O'Dea 1995; 3C 299, van Breugel et al. 1992; 3C 346, van Breugel et al. 1992, Dey & van Breugel 1994, ; Cotton et al. 1995; 3C 380, van Breugel et al. 1992, Wilkinson et al. 1991a). 3C 299 is a very asymmetric double, and 3C 346, 3C 380, and 3C 216 are probably large doubles that are projected with their radio axes at a small angle to the line of sight. 7 However, the majority of the 3C+PW CSS sample do not show any evidence for very extended emission up to limits of 3% of the compact emission (van Breugel et al. 1992).
Baum et al. (1990) reported the discovery of extended emission around the CSO 0108+388. Stanghellini et al. (1990a, 1990b) reported extended emission associated with the candidate GPS sources 0201+113 and 1045+019. The spectrum of 1045+019 is not well determined and its identification as a GPS source requires confirmation. However, 0108+388 and 0201+113 both have GPS-type spectra (Baum et al. 1990; O'Dea et al. 1990b). Thus, these first results suggested that 3/14 or 21% of GPS sources had extended emission. Stanghellini et al. (1998b) have searched for extended emission associated with the Stanghellini complete sample to limits of about 0.1% of the compact emission and found extended emission around two additional sources, 0738+313 (see also Murphy, Browne, & Perley 1993) and 2134+004. In addition, two other sources show secondary emission whose association with the source is still uncertain, 0248+430 and 0941-080. Thus, the Stanghellini complete sample contains 3-5 sources out of 33 with extended emission for a fraction of 9%-15%; if we combine the Stanghellini et al. (1990c) and Stanghellini et al. (1998b) searches, we have 4/42-7/42 or 9%-17% with extended emission. In all cases, the extended emission is diffuse and very faint. Thus, extended structure associated with GPS sources is not common, and the vast majority of GPS sources appear to be truly compact and isolated.
Baum et al. (1990) suggested two scenarios to explain the fraction of GPS sources with extended emission (cf. section 11). (1) "Smothered" sources: If the parsec-scale jet in a large-scale radio source is smothered and the jet propagation is halted on the scales of tens of parsecs, then the source would have both the still existing large-scale structure and the confined parsec-scale source that would have a GPS. The smothering could be produced by a large influx of dense gas acquired in galactic cannibalism (Baum et al. 1990) or by a distortion of the central torus causing the jet to ram into the dense torus gas (Gopal-Krishna 1995). Since the extended emission is no longer being resupplied by the jet, the extended structure would become faint and diffuse on the shorter of the radiative and adiabatic loss timescales. (2) Recurrent sources: If radio sources are recurrent and the new phase of activity (which is initially a GPS source) begins while the extended relic of the previous cycle of activity is still present, this would also give the appearance of a GPS core surrounded by faint extended emission.
7 These four sources are removed from the statistical analysis in this paper. Back.