8.4. Absolute Magnitudes and the Hubble Diagram
Work in progress by Gelderman and collaborators will eventually allow accurate magnitudes to be obtained for CSS sources. This section concentrates on recent work on the GPS host galaxies by O'Dea et al. (1996a).
The absolute magnitudes in the Cousins RC band are between -22 and -24.0 (which corresponds to the range from M* to 2 mag brighter). This range of absolute magnitudes is consistent with the host galaxies of other powerful radio galaxies (see, e.g., Owen & Laing 1989; Owen & White 1991; Smith & Heckman 1989b) and first-ranked cluster galaxies (Postman & Lauer 1995). Thus, the distribution of optical absolute magnitudes is at least consistent with these sources being the progenitors of the large-scale radio sources. Note, however, that examination of the Rc Hubble diagram below does reveal differences between the GPS and 3CR host galaxies.
Figure 18 shows the Rc band Hubble diagram for both the GPS sources and powerful extended radio galaxies (mostly from 3CR). The GPS Hubble diagram has been discussed by Stanghellini (1992), de Vries (1995), Snellen et al. (1996b), and O'Dea et al. (1996a). The good agreement between the GPS and extended radio galaxies (especially at z 0.5) suggests that to first order the host galaxies of the GPS sources and extended sources are essentially the same. However, there is a suggestion that at redshifts of order unity the 3CR galaxies are brighter by about 1 mag than the GPS galaxies. In addition, O'Dea et al. (1996a) show that the GPS galaxies have a steeper slope than the sample as a whole. Since the 3CR galaxies are now completely identified, the offset in magnitude is due not to incompleteness in the 3CR.
Figure 18. (a) The Cousins RC band Hubble diagram for powerful radio galaxies. Magnitudes are not K-corrected. GPS galaxies are represented by asterisks, extended radio galaxies from Owen & Laing (1989) by open squares, 3CR galaxies from Eales (1985) by solid squares, and 3CR galaxies from Dickinson (1995, private communication) by solid triangles. Includes models from Bruzual & Charlot (1993). The key for the models is as follows: Long-dashed curve, nonevolving E galaxy; dotted curve, E galaxy with zf = 10; solid curve, E galaxy with zf = 5; short-dashed curve, E galaxy with zf = 2; dotted-long-dashed line, Sa galaxy with zf = 5. All models assume H0 = 75 and q0 = 0.1. (b) r - i color vs. redshift for GPS host galaxies. The key to the plotted models is as follows: Long-dashed curve, nonevolving E galaxy; dotted curve, E galaxy with zf = 10; solid curve, E galaxy with zf = 5; short-dashed curve, E galaxy with zf = 2; dotted-long-dashed line, Sa galaxy with zf = 5. All models assume H0 = 75 and q0 = 0.1. Adapted from O'Dea et al. (1996a).
One possibility is that the extra light in the 3CR galaxies could be related to radio power - i.e., if the 3CR galaxies are stronger radio sources, then they may have more of the extra component. O'Dea et al. (1996a) showed that there are no systematic differences between the GPS and 3CR galaxies regarding the relationship between optical magnitude and radio power. Thus, the result that, at high redshift, the 3CR galaxies are brighter than the GPS galaxies is not due to the 3CR galaxies having more powerful radio sources.
O'Dea et al. compared the data with models computed with the latest version of the Bruzual-Charlot population synthesis code (described by Charlot, Worthey, & Bressan 1996) assuming a single 0.5 Gyr duration burst of star formation at some formation redshift zf, followed by "passive" evolution. The models are normalized in magnitude to the data at low redshift. The "no evolution" model is that for an average elliptical at age 13.5 Gyr (see Fig. 5 of Bruzual & Charlot 1993). Figure 18 shows that the GPS galaxies are consistent with both the nonevolving elliptical model and passive evolution of an old stellar population (zf 5). Models in which the stellar populations are younger (the elliptical model with zf = 2 and the Sa model with zf = 5) tend to be brighter than the GPS sources at redshifts z 1. The 3CR galaxies at z 1 are brighter than predicted by these passive evolution models with an old stellar population (cf. Spinrad 1986; Spinrad & Djorgovski 1987; Djorgovski 1987). The fact that the GPS galaxies agree with the models for passive evolution of old stellar populations suggests that difference between the GPS and 3CR galaxies is not due to the GPS galaxies being redder than normal elliptical galaxies, but to the 3CR galaxies being bluer than normal elliptical galaxies. This is also consistent with the observed agreement of the GPS sources with the 3CR on the K-band Hubble diagram (de Vries et al. 1998b).
As discussed above, the GPS sources are missing the blue aligned component that is seen in larger scale radio galaxies. This lack of a blue component may be responsible for their different behavior on the Hubble diagram (see also Snellen et al. 1996b). One consequence of this is that the stellar evolution of GPS galaxies may be simpler than the 3CR galaxies with large-scale radio sources. Thus, the "optically boring" GPS galaxies may provide examples of "normal" giant elliptical galaxies at high redshift (Snellen et al. 1996b; O'Dea et al. 1996a; Bremer & Snellen 1996).