Next Contents Previous


The full complement of active galactic nuclei constitutes a zoo of different names, detection criteria, and spectral, polarization, and variability characteristics. As in biology, however, taxonomy derived from empirical observations can impose some order on the chaos. Table 1 shows the principal classes of AGN (adapted from Lawrence 1987, 1993), organized according to their radio-loudness and their optical spectra, i.e., whether they have broad emission lines (Type 1), only narrow lines (Type 2), or weak or unusual line emission. Within each of the groupings, different types of AGN are listed by increasing luminosity. We now explain Table 1 in more detail.

Table 1. AGN Taxonomy

Roughly 15-20% of AGN are radio-loud, meaning they have ratios of radio (5 GHz) to optical (B-band) flux F5 / FB gtapprox 10 (Kellermann et al. 1989), although this fraction increases with optical (Padovani 1993; La Franca et al. 1994) and X-ray (Della Ceca et al. 1994) luminosities, reaching for example ~ 50% at MB ltapprox -24.5. With few exceptions, the optical and ultraviolet emission-line spectra and the infrared to soft X-ray continuum of most radio-loud and radio-quiet AGN are very similar (Sanders et al. 1989) and so must be produced in more or less the same way. The characteristic of radio-loudness itself may be related in some way to host galaxy type (Smith et al. 1986) or to black hole spin (Blandford 1990; Wilson and Colbert 1995), which might enable the formation of powerful relativistic jets.

Based on the characteristics of their optical and ultraviolet spectra, AGN can be separated into the three broad types shown in Table 1.

(1) Those with bright continua and broad emission lines from hot, high-velocity gas, presumably located deep in the gravitational well of the central black hole, are known as Type 1 AGN. In the radio-quiet group, these include the Seyfert 1 galaxies, which have relatively low-luminosities and therefore are seen only nearby, where the host galaxy can be resolved, and the higher-luminosity radio-quiet quasars (QSO), which are typically seen at greater distances because of their relative rarity locally and thus rarely show an obvious galaxy surrounding the bright central source. The radio-loud Type 1 AGN are called Broad-Line Radio Galaxies (BLRG) at low luminosities and radio-loud quasars at high luminosities, either Steep Spectrum Radio Quasars (SSRQ) or Flat Spectrum Radio Quasars (FSRQ) depending on radio continuum shape, with the dividing line set at alphar = 0.5 (where the radio spectrum is measured at a few GHz). Other than luminosity, little distinguishes Seyfert 1s from radio-quiet quasars, or BLRG from radio quasars.

(2) Type 2 AGN have weak continua and only narrow emission lines, meaning either that they have no high velocity gas or, as we now believe, the line of sight to such gas is obscured by a thick wall of absorbing material. Radio-quiet Type 2 AGN include Seyfert 2 galaxies at low luminosities, as well as the narrow-emission-line X-ray galaxies (NELG; Mushotzky 1982). The high-luminosity counterparts are not clearly identified at this point but likely candidates are the infrared-luminous IRAS AGN (Sanders et al. 1988; Hough et al. 1991; Wills et al. 1992b), which may show a predominance of Type 2 optical spectra (Lawrence et al. 1995, in preparation).

Radio-loud Type 2 AGN, often called Narrow-Line Radio Galaxies (NLRG), include two distinct morphological types: the low-luminosity Fanaroff-Riley type I radio galaxies (Fanaroff and Riley 1974), which have often-symmetric radio jets whose intensity falls away from the nucleus, and the high-luminosity Fanaroff-Riley type II radio galaxies, which have more highly collimated jets leading to well-defined lobes with prominent hot spots (see Sec. 5.2). Examples of FR I and FR II radio morphologies are shown in Fig. 2.

Figure 2. Radio images of the two types of radio galaxies: (a) at low luminosity, an FR I radio galaxy, 1231+674, with diffuse, approximately symmetric jets whose surface brightness falls off away from the center, and (b) at high luminosity, an FR II radio galaxy, 1232+414, with sharp-edged lobes and bright hot spots; the jets in this case are often too faint to see. (Courtesy of Frazer Owen and Mike Ledlow.

(3) A small number of AGN have very unusual spectral characteristics. Inventing a term, we call these Type 0 AGN and speculate that they are related by a small angle to the line of sight (``near 0 degrees''). These include the BL Lacertae (BL Lac) objects, which are radio-loud AGN that lack strong emission or absorption features (typical equivalent width limits are set at Wlambda < 5 Å). In addition, roughly 10% of radio-quiet AGN have unusually broad P-Cygni-like absorption features in their optical and ultraviolet spectra, and so are known as BAL (Broad Absorption Line) quasars (Turnshek 1984). If BAL spectral features are caused by polar outflows at small angles to the line of sight, they too are Type 0 AGN as indicated in Table 1; alternatively, they may have edge-on disks with winds instead (Turnshek 1988). There are no known radio-quiet BL Lacs.

A subset of Type 1 quasars, including those defined variously as Optically Violently Variable (OVV) quasars, Highly Polarized Quasars (HPQ) (3) Core-Dominated Quasars (CDQ) or FSRQ, are probably also found at a small angle to the line of sight. Their continuum emission strongly resembles that of BL Lac objects (apart from the presence of a blue ``bump'' in a few cases) and, like BL Lac objects, they are characterized by very rapid variability, unusually high and variable polarization, high brightness temperatures (often in excess of the Compton limit T ~ 1012 K; Quirrenbach et al. 1992), and superluminal velocities of compact radio cores (Sec. 4). Although the names OVV, HPQ, CDQ, and FSRQ reflect different empirical definitions, evidence is accumulating that they are all more or less the same thing - that is, the majority of flat-spectrum radio quasars tend to show rapid variability, high polarization, and radio structures dominated by compact radio cores, and vice-versa (Fugmann 1989; Impey et al. 1991; Valtaoja et al. 1992; Wills et al. 1992a) - so hereafter we refer to them simply as FSRQ. Collectively, BL Lacs and FSRQ are called blazars. Even though the FSRQ have strong broad emission lines like Type 1 objects, they are noted in the ``Type 0'' column in Table 1 because they have the same blazar-like continuum emission as BL Lac objects.

We have described an empirical division of AGN according to radio and optical/ultraviolet properties. Table 1 is analogous to the Periodic Table of the Elements developed by chemists a century or so ago, when many chemical elements had been discovered and studied but the relations among them were not entirely clear. Where in chemistry it was eventually recognized that valence electrons dominate the horizontal relations and nuclear mass the vertical relations, we will argue that the categories in this ``Periodic Table of the AGN'' are distinguished primarily by orientation effects along the horizontal direction and by as-yet unknown physics in the vertical direction.

Whether AGN are classified Type 1 or Type 2 depends on obscuration of the luminous nucleus, and whether a radio-loud AGN is a blazar or a radio galaxy depends on the alignment of the relativistic jet with the line of sight. These two causes of anisotropy, obscuration of infrared through ultraviolet light by optically thick gas and dust, and relativistic beaming of radio jets, both important in radio-loud objects, are discussed in turn in the next two sections, which can be skipped by readers interested only in the actual unification schemes. Note that the unification of quasars with high-luminosity radio galaxies requires both kinds of anisotropy while the unification of BL Lac objects with low-luminosity radio galaxies requires (at present) only relativistic beaming.

3 We refer here only to radio-loud HPQ. While a few radio-quiet quasars also have highly polarized optical emission, and thus fit the HPQ definition, their polarization is almost certainly caused by scattering rather than intrinsic emission processes. Back.

Next Contents Previous