It is well established from radio properties (see Antonucci 1993) that core-dominated AGNs are simply lobe-dominated AGNs viewed from near the jet axis (i.e., near face-on). Gaskell et al. (2004) showed from a comparison of continuum shapes and line ratios that core-dominated and lobe-dominated AGNs have the same underlying optical-to-UV continuum shape and that the SED differences are just due to increased reddening in the lobe-dominated AGNs. We thus have every reason to expect the BLRs of core-dominated and lobe-dominated AGNs to be the same on average. Lobe-dominated radio-loud AGNs should therefore be an excellent laboratory for studying how orientation affects the appearance of the BLR. Miley & Miller (1979) found that lobe-dominated AGNs preferentially had broader and more irregular line profiles. Wills & Browne (1986) discovered that the FWHM of H increases as we see AGNs more edge-on. This provided strong support for a flattened BLR.
AGNs with the peaks of their broad Balmer lines blueshifted or redshifted from the systemic velocity have long been known (Lynds 1968). It was proposed (Gaskell 1983) that these peaks might represent separate BLRs each associated with a member of a supermassive black hole binary, but line profile variability observations on long and short timescales have delivered two fatal blows to this hypothesis. Firstly, although for a while it looked like the expected binary orbital motion was being seen in long-term profile variations in 3C 390.3 (Gaskell 1996), further observations showed that the radial velocity changes were completely inconsistent with a binary black hole (Eracleous et al. 1997) but were instead consistent with orbital motion of concentrations of BLR gas orbiting in a disk. The second fatal blow was that velocity-resolved reverberation mapping of 3C 390.3 strongly ruled out the binary BLR hypothesis because the redshifted and blueshifted peaks varied simultaneously on a light-crossing timescale (O'Brien et al. 1998, Dietrich et al. 1998). This demonstrated conclusively that the double-peaked profiles arose from an inclined disk, as had been widely proposed (see references and discussion in Gaskell & Snedden 1999). Despite these double fatal blows to the idea that displaced broad-line peaks might be due to supermassive binaries, the topic of what signs there might be of sub-parsec supermassive binaries nonetheless remains one of considerable current interest (see review by Tamara Bogdanovic in these proceedings).
A subsequent comprehensive survey of radio-galaxies by Eracleous & Halpern (1994, 2003) revealed many disk-like Balmer line profiles. They found the FWHMs of the Balmer lines to be approximately double those of AGNs with single-peaked Balmer lines. As is shown in Fig. 14, a factor of two reduction in line width is sufficient to make displaced peaks disappear. Gaskell & Snedden (1999), Popovic et al. (2004), and Bon et al. (2006) have argued that a disk-like emission line contribution is probably present in all BLRs but simply hard to recognize because, as illustrated in Fig. 13, the classic double peaks become hard to see when the disk is near to face-on.
Figure 13. The effect of broadening lines on the appearance of structure in line profiles. The left frame shows a Lorentzian and two Gaussians chosen to approximate the appearance of H or H in 3C 390.3 in 1981 or 1988. The right frame has the same line widths and peak intensities as in the left frame, but half the velocity displacements. Figure from Gaskell & Snedden (1997).
It is straight forward to estimate the inclinations of the BLRs from broad disk-like line profiles. These can often be estimated to within a few degrees. Eracleous & Halpern (1994, 2003) get inclinations which predominantly have i > 25deg. Their fits to the disk profiles also provide important confirmation that a significant turbulent velocity is needed and give the turbulent velocity for each object. Without the turbulent velocity component the peaks of the line profiles would be much too sharp. The turbulent velocities are fairly well determined from the line profile fits (to ± 250 km s-1). The average BLR turbulent velocity needed is 1300 km s-1. This is roughly what would be expected from the height of the BLR/torus. The 1- scatter in the derived turbulent velocities is only ± 400 km s-1, which is only slightly greater than the average formal uncertainty in the estimates.
Bon (2008) has estimated inclinations for single-peaked AGNs. For these we mostly see disks with inclinations of i < 25deg (see also Bon et al. in these proceedings). The difference in sini between the displaced-BLR-peak AGNs and single-peaked AGNs is thus about a factor of two. This agrees with the ratios of FWHMs for the two samples.