The nature of the optical-radio correlations for powerful radio galaxies

The nature of the optical-radio correlations for powerful radio galaxies is investigated using spectroscopic observations of a complete sample of southern 2-Jy radio sources. In line with previous work, we find that significant correlations exist between the luminosities of the [O III]lambda 5007, [O II]lambda 3727 and H beta emission lines and the radio luminosity. However, our observations are not easily reconciled with the idea that these correlations are caused by the increase in the power of the photoionizing quasar as the jet power increases, with average ISM properties not changing appreciably with redshift or radio power: not only do we find that the scatter in the L-[O III] versus L-radio correlation is significantly larger than in L-[O II] versus L-radio and L-H beta versus L-radio correlations, but the ionization state deduced from the emission lines does not increase with radio power as predicted by the simple, constant ISM, photoionization model. We conclude that (a) there exists a considerable range in the quasar ionizing luminosity at a given redshift, and (b) the mean density of the emission-line clouds is larger in the high-redshift/high-power radio sources. The latter density enhancement may be either a consequence of the increased importance of jet-cloud interactions or, alternatively, the result of a higher pressure in the confining hot ISM, in the high-redshift objects. Apart from the general scatter in the correlations, we identify a distinct group of objects with [O III]lambda 5007 luminosities which are more than an order of magnitude lower than in the general population radio galaxies at similar redshift. These weak-line radio galaxies (WLRGs) are likely to be sources in which the central ionizing quasars are particularly feeble. Deep spectra show that many of the sources in our sample are broad-line radio galaxies (BLRGs). The fact that the BLRGs are observed out to the redshift limit of the survey, overlapping in redshift with the quasars, argues against the idea that BLRGs are simply the low-radio-power counterparts of high-power, high-redshift quasars. Either there exists a considerable range in the intrinsic luminosities of the broad-line AGN for a given redshift or radio power, or the BLRGs represent partially obscured quasars. The degree of scatter present in the L-[O III] versus L-radio correlation supports the former possibility.