Next Contents Previous

5.3.2 Narrow Emission Lines

Assuming the narrow emission lines are emitted isotropically, unification schemes predict similar narrow-line luminosities in quasars and FR II radio galaxies. Due to the intrinsic correlation between narrow-line luminosity and extended radio emission (Baum and Heckman 1989; Rawlings and Saunders 1991; Zirbel and Baum 1995), the comparison has to be done matching objects of similar extended radio luminosity. For flux-limited samples, this is roughly equivalent to matching in redshift (although the scatter in the redshift-Pext correlation can be large for core-dominated objects since it is an induced correlation). Also, matching the objects in redshift will exclude the effects of cosmological evolution, as well as observational selection effects due to the bandpass used.

Quasars do have systematically higher [O III] lambda 5007 luminosities than radio galaxies (Jackson and Browne 1990), but this line appears to be emitted anisotropically when compared to [O II] lambda 3727 (McCarthy 1989; Hes et al. 1993), possibly due to contribution to [O III] from the broad-line region (although there is no sign of broad wings on the [O III] line) and/or partial obscuration by the thick torus. (Excluding the low-excitation FR II galaxies also reduces the discrepancy in [O III] luminosities: Laing et al. 1994.) As shown in Fig. 7, the [O II] lambda 3727 line luminosities of 3CR steep-spectrum quasars and FR II radio galaxies matched in radio power are completely overlapping (Hes et al. 1993), as expected from the unification scheme. (9)

Figure 7. The [O II] emission line luminosity for quasars (filled circles) and powerful (FR II) radio galaxies (open squares) in the 3CR sample, plotted versus redshift to allow comparisons at similar radio luminosity (which is well correlated with redshift in a radio-flux-limited sample) and/or observed wavelength of [O II]. The narrow emission line luminosities of radio galaxies and quasars, which should be unaffected by obscuration or relativistic beaming, span the same values over the full redshift range. (Data from Hes et al. 1993.)

9 Any trend with redshift is induced through the correlation of luminosity and redshift in flux-limited samples, and should therefore be ignored in this and similar plots. Back.

Next Contents Previous