Annu. Rev. Astron. Astrophys. 1977. 15: 69-95 Copyright © 1977 by Annual Reviews. All rights reserved

7. RELATION TO QSOs

The similarities between the nuclei of Seyfert galaxies and QSOs have been pointed out many times, and numerous efforts have been made to demonstrate a continuity between these objects (e.g. Sandage 1971, Lynden-Bell 1971, Notni & Richter 1972, Arakelian 1971, Weedman 1976a). It is the Sy 1 nuclei whose properties most closely resemble the QSOs. It is argued that as long as compact sources of high luminosity are known to exist in the nuclei of Seyfert galaxies, there should be no objection to interpreting QSOs as distant, more luminous examples of such sources. The Seyfert nuclei are therefore used to demonstrate empirically that QSOs really have cosmological redshifts. The most recent presentation of this continuity argument is by Khachikian & Weedman (1977). They consider the magnitude-redshift and angular diameter-redshift diagrams for Seyfert galaxies. There is substantial scatter in the former (the Hubble diagram), which is attributed to the differences in absolute luminosity among the various Seyfert nuclei. Assuming a linear redshift-distance relation gives a scatter (m) = 1.04 mag for Seyferts compared to Sandage's 0.28 mag for first ranked ellipticals. It is argued that angular diameters should be a more reasonable measure of relative distance for Seyferts since such diameters are determined by the galactic disks rather than the nuclei. The angular diameters measured from the Sky Survey decrease with redshift and have a scatter about the mean line of ( log ) = 0.21 for Sy 1. This compares to 0.11 for first-ranked ellipticals. It is therefore argued that, while Seyferts are not as homogeneous as first-ranked ellipticals, they still show good evidence for a linear redshift-distance relation.

There is not yet agreement, however, that the extended nebulosities associated with "Seyfert galaxy" nuclei are really the disks of galaxies (Burbidge 1973). The morphological studies discussed previously (Section 3) have demonstrated that a lot of Seyferts are spiral galaxies, but such demonstrations are difficult for the high redshift Seyferts that are critical in providing a transition to QSOs. There are a few Sy 1 that look like spiral galaxies whose luminosities would approach those of cosmologically redshifted QSOs. These include Mkn 79 and 618, NGC 7469, IC 4329A, I Zw1, and II Zw 136.

If cosmological redshifts are accepted for Seyferts and assumed for QSOs, there is an overlap in luminosity between the nuclei of Sy 1 and the QSOs. This is summarized in Khachikian & Weedman (1977). The luminosity indicators in their redshift-luminosity diagram are the broad hydrogen lines from Sy 1 and QSOs. Such lines arise only in the nuclei and are roughly proportional to the luminosity of the nonthermal continuum that carries most of the energy. The H lines of lower redshift objects are related to the L lines in high redshift QSOs by assuming an intrinsic L: H ratio of 40 (Davidson 1972). Luminosities are given for 37 Sy 1 and 62 QSOs using fluxes collected in Weedman (1976b) and Osmer & Smith (1976). The optically brightest QSOs are presumably included because nine of the high redshift QSOs were discovered from their strong L emission on objective prism survey plates (Smith 1976). The Sy 1 nuclei cover a luminosity range of 10³, and most QSOs are less than 10 times brighter than the brightest Seyfert. The entire phenomenon, from faintest Sy 1 to brightest QSO, covers a factor of 10⁵ in luminosity. As the bolometric luminosities are approximately 10³ times the H luminosity (Weedman 1976b), this corresponds to a luminosity range of 10⁴³ ergs sec^-1 to 10⁴⁸ ergs sec^-1 (for q₀ = 0 and H₀ = 50 km sec^-1 Mpc^-1). Improvements in these numbers can be expected as more far-infrared data are accumulated (e.g. Rieke & Low 1975a) and as better composite spectra of Sy 1 and QSOs are assembled (e.g. Chan & Burbidge 1975, Baldwin 1975).

A key link between the nuclei of Seyfert galaxies and the QSOs is the fact that both can be variable in luminosity. As a consequence of the small radiating volumes implied by this variability, there are severe constraints on theoretical models of the radiation sources. Extensive considerations of such compact, nonthermal sources have been presented by Jones et al. (1974a, b) and Burbidge et al. (1974). Results relevant to this problem are also reviewed by Stein et al. (1976) in the context of the BL Lacertae objects. The energy-density paradox that arises in a luminous source that varies rapidly was first pointed out by Hoyle et al. (1966). Simply put, the problem is that electrons producing photons via synchrotron radiation will lose their energy in subsequent Compton scatterings with the very photons they created. The radiation source quickly quenches itself unless there is continuous injection of high-energy electrons and/or rapid source expansion (see references in Stein et al. 1976). Recent VLBI measures of the nuclei of the Sy 1 3C 120 and other radio sources do indicate relativistic expansions, or at least separations of individual radio-source components at relativistic speeds (Schilizzi et al. 1975, Wittels et al. 1976). The references cited include numerous references to earlier work in which such expansion was suspected (e.g. Shapiro et al. 1973, Kellermann et al. 1973 for 3C 120). The important progress in the recent work has been the demonstration that source contraction is never observed. This means that "Christmas-tree" models, in which the changes in source structure are attributed to flaring at random places in the source, are ruled out.

The long-term monitoring of Seyfert galaxies with different techniques has shown that variability is a common property. At one time or another, observers of all parts of the spectrum have reported a variable Seyfert galaxy. As such variability is the only measure of source size for most Seyfert nuclei, it is certainly important to know about. Data on the optical variability of Seyferts has been assembled by Cannon et al. (1971), Lyutyi (1973), Penston et al. (1974), Lyutyi & Pronik (1975), and Scott et al. (1976). Other studies of individual objects are included in the UBV references in Table 1. (Some useful comparison star sequences are in Penston et al. 1971.) One Seyfert (X Comae) was even discovered because of its variability (Bond 1973). The problem in interpreting the results is the tedious accumulation of data required before correlations can be searched for. Are there, for example, any correlations between variability and source luminosity (e.g. Elliot & Shapiro 1974)? It appears at the moment that the optical continua of all Sy 1, like the QSOs, are variable if observed closely enough. It is of great importance to demonstrate whether the continua of Sy 2 are variable as it is not yet known whether or not these sources are nonthermal.

Reports have existed for some time that the emission lines can also vary in Seyfert galaxies. Such variations have been reported for NGC 1068, 1275, 1566, 3227, 3516, 4151, 5548, 7469, 7603, 3C 390.3 and Markarian 6 (references to Table 1 and Lyutyi & Pronik 1975). Such variations are important for understanding the volume of the ionized gas and the location of the gas relative to the variable continuum source. Unavoidably, most of the observations from which line variability has been deduced are very inhomogeneous. Results from different observers using different spectrographs have been compared and changes claimed. Sometimes, the emission lines have been monitored with narrow-band interference filters without obtaining spectra. NGC 1068, in which the emission arises from a resolved nucleus has been observed with entrance apertures smaller than the nucleus. Consequently, much of the data regarding emission-line variability is hard to evaluate with certainty. It is particularly perplexing when such results are reported for NGC 1068 in which low-density gas is thought to encompass a volume several hundred parsecs in diameter. It is important to continue attempts to confirm such results, however, especially because there are convincing cases in which line variability has been carefully monitored under the same observing conditions. The best examples seem to be the emission-line variations in NGC 7603 (Tohline & Osterbrock 1976) and the variable interstellar absorption lines in NGC 4151 (Anderson 1974a). As systematic monitoring of any sort requires diligence and patience, it is likely to be quite some time before sufficient data is accumulated to understand the nature of variability in Seyfert nuclei.