Annu. Rev. Astron. Astrophys. 1982. 20: 431-468 Copyright © 1982 by Annual Reviews. All rights reserved

2.2. Near-Neighbors and Interacting Systems

Stocke (1979), Dressler (1980a), and others have argued convincingly that the immediate environment of a galaxy has a far more important influence on its properties than membership or nonmembership in a cluster. Such a suggestion and the relevant observations are explored in this section.

A pair of galaxies with a small ( 100 kpc) projected separation is almost certain to be a physical rather than optical pair. Interacting galaxies are those for which one or more companions are nearby, and evidence suggestive of tidal "debris" (Toomre & Toomre 1972) can be identified. Probably closely related are disturbed galaxies that show the usual signs of tidal stress, yet no companion is readily identifiable. We further consider galaxies with double nuclei since such galaxies may be in an advanced stage of an interaction leading to a merger. All such types of galaxies often show signs of activity.

Galaxy-galaxy interactions or mergers may be a very important mechanism in triggering and sustaining galactic activity, and there are a large number of radio galaxies whose optical morphology supports this (see below). Begelman et al. (1980) argue that the observed structure of some radio sources can be explained in terms of a merger which produces two orbiting, precessing, supermassive black holes in the merged galaxy's center. Whitmire & Matese (1981) have extended this suggestion; by analogy with SS 433 they argue that supermassive binary systems can explain many phenomena in radio galaxies. Roos (1981b) has detailed how a merger can lead to stars being scattered into the "loss cone" of a central black hole and, through subsequent stellar disruption and accretion, can power the nuclear activity.

RADIO SOURCES     A number of recent surveys of radio emission have convincingly shown that the presence of nearby or interacting galaxies enhances the probability of finding a radio source associated with at least one of the galaxies (Sulentic 1976, Stocke 1978, Condon & Dressel 1978, Dressel 1981, Condon et al. 1982, Hummel 1980, 1981b, Adams et al. 1980). This result holds for both spiral and elliptical galaxies, and is of unquestionable significance as a clue to the origin of galactic activity.

Some debate attends specifics of the radio morphology and its implications concerning the onset of activity. Hummel (1981b) has shown that the enhanced radio emission associated with close pairs of galaxies comes from the nucleus of a galaxy rather than its disk. However, Condon (1980) and Condon et al. (1982) find that the radio emission is not strictly nuclear, and that it preferentially arises within 1 kpc of the nucleus in areas of the greatest stellar projected density. They argue that the radio emission comes from supernovae and that the nucleus plays only a minor role, if any (see also Feldman et al. 1981).

The situation is somewhat less complicated for interacting elliptical and SO galaxies where, in general, the radio sources are clearly not associated with starbursts inside the galaxy. In most such cases these sources qualitatively resemble the much more powerful radio sources in quasars. Surveys of radio emission from relatively bright, nearby early-type galaxies have been made by Condon & Dressel (1978), Dressel (1981), Hummel (1980), and Kotanyi (1981). For the low-luminosity radio sources it is fairly clear that the luminosity of radio emission is enhanced by the presence of close companions.

At the high levels of radio luminosity typical of sources selected from radio flux-limited samples (e.g. 3C, 4C, Parkes), there is little information concerning the effects of neighboring galaxies on the production of radio emission. The occurrence of strong radio emission associated with double-nucleus galaxies has long been recognized (e.g. the "db" classification of Mathews et al. 1964). Stocke (1979) pointed out that most classical radio galaxies have a companion galaxy of comparable brightness within ~ 50 kpc, but no correlations were found between the attributes of the companions and the radio source according to Kingman (1981).

There is growing evidence, though not as yet of a statistical nature, to suggest that galaxy mergers may be important in triggering the production of a radio source. The two nearest radio galaxies, NGC 1316 = Fornax A (Schweizer 1980) and Cen A (Tubbs 1980), have both been interpreted as cases of a recent galaxy merger. Other noteworthy radio galaxies in which a recent merger has been proposed include NGC 1052 and NGC 4278 (Gunn 1979), 3C 305 (Heckman et al. 1981a), 4C 29.30 (van Breugel et al., in preparation), and NGC 6240 (Fosbury & Wall 1979). The latter authors have drawn attention to the fact that Toomre's (1977) merger candidates from the NGC catalog may have higher-than-normal radio luminosities.

The properties of these systems that are suggestive of a merger include the following: unusually large amounts of gas and/or dust for their early Hubble type, misaligned stellar and dust/gas dynamical axes, morphological or kinematic peculiarities in the ionized gas, morphological peculiarities in the stellar body of the galaxy, and shock-heated extranuclear emission-line gas.

SEYFERT GALAXIES     Considering the avalanche of literature published on Seyfert galaxies, it is astonishing that no methodical statistical study of the effect of interactions, mergers, or even nearby galaxy density exists for these objects. Stauffer (1981) has found that galaxies with emission lines indicative of low-level nuclear activity (most of which are not classical Seyferts) occur more frequently in groups than in more isolated environments.

Studies of the morphologies of classical Seyfert galaxies have been conducted by Simkin et al. (1980), Adams (1977), and Wehinger & Wyckoff (1977). The Iwo earlier surveys give the impression (not rigorously statistically substantiated) that an anomalously large fraction of Seyfert galaxies are interacting and disturbed.

To the list of active interacting galaxies can probably be added two "ring galaxies" with Seyfert nuclei: NGC 985 (de Vaucouleurs & de Vaucouleurs 1975) and "The Carafe" (Hawarden et al. 1979a). Such objects may be produced as a consequence of a violent collision (Theys & Spiegel 1976).

Often both members of an interacting pair of galaxies are active, or at least have strong emission-line nuclei (ELNs) (Ward et al. 1978, Ward & Wilson 1978, Burbidge et al. 1963). Petrosyan et al. (1979, 1980a, b) have identified several double-nucleus Seyfert or Seyfert-like galaxies, all of which may be interactions in an advanced stage of evolution. Further studies of these types of active galaxies, including neutral hydrogen and deep optical imaging, would be of considerable importance in establishing mergers as a cause of activity in Seyfert galaxies.

Tifft (unpublished) has discovered a strong correlation between the spectroscopic properties in pairs of galaxies, in that both members tend to have a comparable degree of line emission. Furthermore, he finds that the strength of the emission lines correlates with the relative radial velocities of the nuclei; strongest emission is found in low-velocity interactions (v ~ 10² km s^-1). Both the emission-line strength and the relative velocity of galaxy pairs could easily depend on the total masses of the galaxies, and so the correlation between them could be secondary. Little correlation between nuclear activity and pair separation has been found. This last result is perhaps surprising and in need of further examination. Further work on this sample would be most valuable.

QUASARS     There is little information on the local extragalactic environment of quasars. Weymann et al. (1978) showed that the positions of galaxies are more strongly correlated with quasars than with other galaxies. Stockton (1980) similarly concluded that (radio-loud) QSOs occur in regions where the density of galaxies is significantly higher than average. Stockton emphasized that quasars are found in galaxy groups rather than rich clusters - a conclusion consistent with Seyfert galaxies and low-luminosity radio galaxies. However, Seldner & Peebles (1979) find a statistically significant excess of galaxies around quasars at all redshifts, which, if not accidental, may argue that quasars are local and their redshifts are noncosmological.

Hutchings et al. (1981) and Wyckoff et al. (1981) have shown that the fuzz - presumably starlight - surrounding low-redshift QSOs is sometimes asymmetric. This might suggest that a substantial number of QSOs originate in closely interacting systems. Stockton (1982) reports the discovery of three cases of nearly stellar objects located within a few kpc (in projection) and ~ 200 km s^-1 of a low-redshift quasar. He suggests that these objects are close companions whose interaction with the quasar host galaxy has triggered the quasar activity.

In concluding this section, we must note that available observations of active interacting galaxies are consistent with the theoretical notions of interactions as one of the many causes of galactic activity. As yet the observational tests are not critical; many details remain to be explored and a more secure statistical foundation for the models is still needed.