QUASISTELLAR OBJECTS, HOST GALAXIES ALAN STOCKTON The host galaxy of a QSO is the galaxy in which the QSO is embedded and of which it is, presumably, the nucleus. The original definition of QSOs specified that they should have essentially stellar images. At the time (i.e., the mid-to late-1960s), this statement meant that little or no difference from the images of stars was apparent on deep photographic exposures obtained with large telescopes. Although it was widely believed that QSOs were extremely luminous galactic nuclei, only a few were known to show faint extended luminous material that might be taken as evidence for a surrounding galaxy, the most notable of these being 3C 48. An important early study by Jerome Kristian showed that this general lack of visible extended material was nevertheless consistent with QSOs being in giant elliptical galaxies: Scattering of the QSO light in the atmosphere, telescope optics, and photographic emulsion was sufficient to obliterate any evidence of a normal surrounding galaxy. In fact, somewhat ironically, it appeared that if the faint material seen around 3C 48 was to be interpreted as a galaxy, it would have to be an abnormally large and luminous galaxy. More recently, advances in detector technology and the availability of large telescopes at sites with consistently good atmospheric image quality (i.e., what astronomers call "seeing") have resulted in the resolution of extended luminous material around virtually all QSOs with redshifts z<0.5 and around many at higher redshifts. The focus of studies of this material has shifted from simple detection to elucidating its nature and origin. VARIETIES OF EXTENDED LUMINOUS MATERIAL Applying the term "host galaxy" to the extended luminous material around a QSO presupposes that the extended luminosity is due to a distribution of stars. That such was the case was tacitly assumed in most of the early work; it therefore came as a surprise to find a few objects whose extranuclear luminosity in bandpasses of a thousand angstroms or more was dominated by a single strong emission line from extended ionized gas surrounding the QSO. Although such objects are in the minority, they are by no means uncommon, so it is dangerous to draw conclusions regarding the nature of the host galaxy from images whose bandpass includes one or more of the emission lines known to be strong in gas at low densities and moderately high ionization. The principal relevant lines are Ly-*, [0**]*3727, [O***]**4959, 5007, and H*. In most cases, at least among the low-redshift (z<0.5) QSOs, the distributions of the extended continuum sources and of the ionized gas are quite different, so a misidentification of an emission feature as indicating the structure of the stellar component of a host galaxy could lead to serious errors in interpretation. Even bona-fide continuum features must be viewed with some caution, although problems of interpretation here are less likely. Synchrotron jets producing strong optical radiation occur infrequently on the sorts of scales likely to cause conclusion for ground-based studies of QSO host galaxies, with that associated with 3C 273 being a notable exception. Such jets will presumably be found more frequently in images obtained with the Hubble Space Telescope: They can be distinguished from thermal sources such as stars by their high polarization, their power law spectra, and their coincidence with radio features. For some source geometries and viewing angles, scattered light from the nuclear continuum source may be important, particularly in cases for which direct radiation from the nucleus does not reach the observer. One would generally expect in such cases that the scattered light would show the spectrum of the QSO's broad-line region as well as the nuclear continuum, and that both the continuum and the broad lines would show significant polarization. Scattered nuclear light with these properties has been found associated with the Seyfert galaxy NGC 1068. The surest positive sign that an observed extended distribution of luminosity around a QSO is indeed due to stars is the spectroscopic detection of stellar absorption lines. Because of the faintness of the material and the problem of scattered light from the adjacent bright nucleus, convincing detections exist for only a few QSOs. These few, however, are sufficient to establish the existence of bona-fide QSO host galaxies and make plausible the assumption that most of the extended continuum emission seen in deep images of QSOs is due to stars, even in the absence of spectroscopic confirmation. PROPERTIES OF QSO HOST GALAXIES Assuming that the worries mentioned previously are taken care of, and we are reasonably confident that we are actually dealing with a stellar system, what can we say about the host galaxies of QSOs? As may be expected, it has proved to be quite difficult to say anything very definite about many of the most luminous QSOs, for which the strong nuclear component tends to overwhelm the extended stellar distribution. The usual criteria for distinguishing different classes of galaxies include: (1) radial luminosity profiles; (2) structural features, such as spiral arms; and (3) colors. Of these, the last may be unreliable for QSO host galaxies, for reasons to be discussed shortly the first two are difficult to apply in practice for ground-based observations of normal galaxies with z>0.3, and the presence of a luminous nucleus only aggravates the situation. Considering the large amount of telescope time that has been spent on observations of QSO host galaxies, the results in terms of definitely established properties are relatively meager. It is widely expected that observations with the Hubble Space Telescope will settle many of these remaining questions. Because of these difficulties, much of the ground-based work has concentrated on less-luminous objects and on extrapolations from what we know about related forms of nuclear activity in well-resolved, low-redshift active galaxies. One example is the widely held view that radio-quiet QSOs are simply the high-luminosity tail of the Seyfert galaxy population, which are largely spirals, and that radio-loud QSOs bear a similar relation to the broad-lined radio galaxies, which appear always to be ellipticals (though often peculiar in some way). What evidence there is from QSO imaging surveys tends to support this division, although to date the evidence that many radio-quiet QSOs are in spirals appears to be firmer than that for radio-loud QSOs being in ellipticals. The few cases for which good spectroscopy of the host galaxies have been obtained show stellar spectra in the optical region ranging from a dominantly A-type spectrum (for 3C 48) to those typical of a smaller admixture of young stars in a predominantly older population. This spectroscopic evidence is in general agreement with what color information exists: Those host galaxies that morphologically look most like ellipticals nevertheless have bluer colors than elliptical galaxies generally, indicating recent star formation. EVIDENCE FOR INTERACTIONS Essentially all of the large surveys have resulted in the conclusion that a large fraction of QSO host galaxies have distorted morphologies or are otherwise peculiar. An unusual number seem to have close companions. Even though the statistical bases for these claims have not been thoroughly established, this evidence, together with the color information indicating recent widespread star formation is not unusual in QSO host galaxies, have led to the common supposition that galaxy interactions have played a major role in the QSO phenomenon. Even before there was any observational evidence to support such a conjecture, some of the components of a link between galaxy interactions and nuclear activity were being discussed. In a classic paper on interactions, Alar and Juri Toomre included a speculative section (entitled "Stoking the Furnace"), in which they suggested that interactions might often bring a fresh supply of gas deep into the centers of galaxies. Although the Toomres spoke only in terms of enhanced star formation in galaxy nuclei, this gas could also presumably fuel a black hole (it was about this time that an effective consensus had been reached that QSOs were very likely powered by energy released as matter fell onto a compact supermassive object, presumably a black hole). These thoughts were made explicit in a paper ("Feeding the Monster") by James E. Gunn, which primarily dealt with the problem of removing angular momentum from the gas so that it can come within reach of the central black hole. This issue remains a thorny one to this day. The most obvious signatures of tidal interactions are large-scale distortions in the luminosity distributions of galaxies, particularly the tidal bridges and tails that formed the main subject of the Toomres' paper. Such features are likely to be really prominent only for disk galaxies. In the absence of these or other morphological clues, the presence of a close companion galaxy may suggest an interaction, but cannot be conclusive in any individual case because of the possibility of projection effects. However, a statistical excess of close companions can be strong circumstantial evidence in favor of the importance of interactions. A final piece of evidence is enhanced star formation in QSO host galaxies, because it is well known that nearby interacting galaxies often show galaxy-wide enhanced star formation. In some cases, apparent tidal tails occur in situations where only one galaxy is visible. These can be explained as cases where the two interacting galaxies have merged or where one of the galaxies is hidden in the region obscured by the brilliant QSO nucleus. From the evidence available so far, it appears that a significant fraction of QSOs are in interacting systems. Because of the observational difficulties involved and the subtlety of some kinds of interaction signatures, it is remarkable that there is as much evidence as there is. Whether essentially all QSOs are due to interactions will, once again, be a program for the Hubble Space Telescope. ULTRALUMINOUS IRAS GALAXIES One of the major discoveries of the Infrared Astronomical Satellite (IRAS) was the detection of a population of galaxies emitting most of their energy in the far-infrared region of the spectrum. The most luminous of these emit as much power as do many QSOs, and the most plausible interpretation for some of them, at least, is that an active QSO-like nucleus is present at their centers but hidden by massive amounts of dust. The dust absorbs the optical and ultraviolet radiation from the nuclear continuum source and reradiates it in the far infrared. These ultraluminous IRAS galaxies are virtually all members of strongly interacting or merging disk systems, and the dust is a result of the unusually vigorous, galaxy-wide star formation that has been induced by the interaction. A speculative scenario is that some fraction of such galaxies, once the star-formation rate moderates and the dust is reduced, will be visible as normal QSOs (not all can be, because the number density of the ultraluminous IRAS galaxies is larger than that of QSOs, and it seems unlikely that the dust enshrouded stage would last significantly longer than the "normal" QSO stage). If this connection can be demonstrated, it would provide another link between interactions and QSO activity. QSOs IN GROUPS AND CLUSTERS The host galaxies of QSOs at low redshifts are usually found in small groups, but almost never in rich clusters. At larger redshifts, however, QSOs are found in richer environments and possibly in more luminous galaxies. Most of the conclusions regarding QSO host galaxies mentioned previously are based on samples of QSOs strongly biased toward low redshifts and may not apply to those at higher redshifts. In particular, the effect of interactions is likely to be more important in small groups, where relative velocities of galaxies are low, than in rich, relaxed clusters. On the other hand, it has been suggested that nuclear activity may sometimes be fueled by gas from cooling flows resulting from thermal instabilities in the hot intracluster medium, and this process may be more important for QSO host galaxies that are central galaxies in rich clusters. FUTURE PROSPECTS When the Hubble Space Telescope is optically corrected to perform as planned, many of the present uncertainties regarding QSO host galaxies should be cleared up. The fivefold increase in spatial resolution over the best ground-based images, the ability to obtain high-quality images in the ultraviolet, and the absence of the strong airglow background in the near infrared-all of these should make a substantial difference in our ability to interpret these elusive objects. Additional Reading Balick, B. and Heckman, T.M.(1980). Extranuclear clues to the origin and evolution of activity in galaxies. Ann. Rev. Astron. Ap. 20 431. Courvoisier, T. and Mayor, M., eds.(1991). Active Galactic Nuclei. Springer, Berlin. Heckman, T.M.(1990). Galaxy interactions and the stimulation of nuclear activity. In Paired and Interacting Galaxies, IAU Colloquium 124, J.W. Sulentic, W.C. Keel, and C.M. Telesco, eds. NASA CP-3098, Washington D.C., p. 359. Soifer, B.T., Beichman, C.A., and Sanders, D.B.(1989). An infrared view of the universe. American Scientist 77 46. Stockton, A.(1986). The environments of QSOs. Ap. Space Sci. 118 487.