QUASISTELLAR OBJECTS, SPECTROSCOPIC AND PHOTOMETRIC PROPERTIES BEVERLEY J. WILLS Historically, QSOs (quasistellar objects) were defined as point-like optical counterparts of radio sources having broad, redshifted emission lines, strong ultraviolet emission (UV excess), and strong time variability of the optical light. Although first discovered by their strong radio emission, more than 90% are only weak radio sources [flux density F*4, and the highest redshift (early 1991) is 4.73, corresponding to a look-back time of 93% of the age of the universe (deceleration parameter q*=0). Ground-based observations can reach a wave-length of about 600 * in the rest frame, satellite observations about 200 *. Optical flux densities range from approximately 38 mJy (apparent visual magnitude V*12.5) to the limit of photographic sky surveys (*6*Jy or V*22) and fainter. As expected for cosmological redshifts, there is a clear trend between V and redshift; however, the distribution of flux density at a given redshift (at any wave-length) is very broad. The implied luminosities of the QSOs are as high as **** a few x 10** erg s** (for a Hubble constant, H*=100 km s** Mpc**, and q*=0). The lower-luminosity limit is often arbitrarily set at Mv=-21.5 (10** erg s**). This is the approximate luminosity at absolute visual magnitude which the nucleus dominates light from the host galaxy. Thus there is no clear distinction between the QSOs and other active nuclei of galaxies, especially at low redshifts (z*0.1). EMISSION LINES There are two kinds of emission lines: relatively narrow lines from both permitted and forbidden transitions, with Doppler widths (FWHM, or full-width at half maximum intensity) from 300-1500 km s**, typically 600 km s**; and broad permitted lines with FWHM from 1000-10,000 km s**, typically 4000 km s**. The full-width at zero intensity for broad lines can approach 50,000 km s**ù The broad-line region and most of the narrow-line region has never been spatially resolved. In some cases, especially at low redshift, some of the narrow-line emission is observed out to many kiloparsecs from the nucleus. Table 1 lists typical relative intensities for broad lines. For some strong lines we also give typical equivalent widths (rest frame), which are a measure of the line strength relative to the local continuum. Unlike other discrete lines the Fe ** emission consists of thousands of blended broad lines forming a pseudocontinuum, mostly between 2000 and 5500 *. In general, the relative strengths among narrow lines, those among broad lines, and the broad-line equivalent widths are similar from one QSO to another, and there are very few for which there is clear evidence for reddening by dust. This may be a result of observational bias as QSOs are often selected by Uv excess and for their luminous quasistellar nuclei. The strengths of the narrow lines can be dominant to very weak compared with the broad lines; for example, the ratio [O ***] *5007/H* is inversely correlated with both optical continuum and x-ray luminosity. The profiles of the narrow lines, where measurable, are generally smooth and symmetric, although there are often small red or blueward asymmetries. The broad-line profiles are also generally quite smooth and symmetric, suggesting many emission clouds rather than a few that dominate the profile. The wings follow an approximately logarithmic intensity distribution. The overall shape may be peaked or stubby. The profiles are generally very similar from one line to another within the same QSO. However, in detail, C ** * 1549 often shows a stronger wing toward short wavelengths; the high ionization lines and Ly * tend to be broader and blue-shifted by about 1000 km s** with respect to the lower ionization lines, which appear to be at rest with respect to the narrow lines, and probably with respect to the systemic velocity as indicated by stellar absorption features. In determining profiles, blending with other weak lines can be a serious problem; for example, [C ***] *1909 can be blended with A] *** *1858, Si ***] *1892, and 1 Fe *** UV34, and weak Fe ** blends can seriously affect many discrete lines. The broad emission lines in the rare class of broad absorption line (BAL) QSOs are different in having larger A] ***/C ***] intensity ratios, probably stronger Fe ** and, compared with NV *1240 and C ***] * 1909, unusually weak C ** * 1549 emission lines whose profiles do not agree well with those of lower ionization lines. CONTINUUM An interesting approximation to the continuum spectral energy distribution is that the power per decade of frequency is roughly constant from radio to x-ray frequencies, except at radio wavelengths for radio-quiet QSOs, where the power may be down by a factor of 10,000 or more. This suggests an underlying relation between energy production processes at different frequencies. However, the same mechanism is not responsible for the whole continuum. After subtracting the contributions from discrete and blended lines and Balmer continuum, there are at least three broad spectral components, even in the near infrared through ultraviolet-the Big Blue Bump, the 3-*m Bump, and a "power law" component. The first two of these seem to be associated with the broad emission line region. BIG BLUE BUMP In most QSOs the optical-ultraviolet region is dominated by the Big Blue Bump, with ******** in the optical and near ultraviolet. An observed minimum in *** near approximately 1 *m suggests that this component declines in the near infrared. Statistically, this power law continues to Ly * in the higher-redshift QSOs, but falls slightly for lower-redshift QSOs. This may be intrinsic to the QSO continua but the results are also consistent with dust "reddening" in lower-luminosity QSOs. (Observational selection in a flux-limited sample results in the least luminous objects being detectable only at small redshifts, introducing a spurious correlation between redshift and luminosity.) Beyond Ly * (*<1216 *), for the continua of intermediate- and high-redshift QSOs the flux density falls markedly, and increasingly with increasing redshift-as steep as ********* for *<912 *, the Lyman limit, sometimes steeper. This is attributed to increasing absorption by intervening neutral hydrogen at lower redshifts, with increasing distance to the QSO. Recent extreme ultraviolet observations of one luminous quasar, HS 1700+6416, suggest an intrinsic continuum that is a simple extrapolation of the optical-ultraviolet continuum to 300 *, although Lyman line and continuum absorption severely depresses this continuum between 850 and 450 *. This same Big Blue Bump component may be responsible for the soft x-ray excess above an extrapolation of the hard x-ray spectrum. POWER LAW COMPONENT It is often assumed that there is a power law component (*** constant) dominating in the near infrared that can be extrapolated into the optical-ultraviolet and even to the x-ray region. Quasi-power-law continua do indeed dominate in QSO-like objects called Blazars, and appear to be a smooth continuation of their high radio frequency continuum. Actually, this component in the Blazars is only approximately power law, typically varying with wavelength from **** constant in the near infrared, generally steepening to as much as ***** in the ultraviolet. Flux densities vary rapidly, often by factors of 2, occasionally by factors of approximately 100. The time scales are hours to weeks for smaller variations, and weeks for larger amplitude changes. This component also has high linear polarization, often extremely variable in degree and position angle. This component, attributed to electron synchrotron radiation, is strongly correlated with the presence of a luminous, compact radio core. It has been shown recently that any QSOs with luminous compact cores, not just the traditional Blazars, are likely to have such a power law component whose strength probably depends on the luminosity of the radio core. Thus a power law component has not yet been demonstrated in QSOs with weak radio cores, or in the radio-quiet QSOs. TIME VARIABILITY For the Big Blue Bump component (contributing the entire optical- ultraviolet continua of radio-quiet and weak-radio-core QSOs), the amplitudes of variability range from less than 10% to factors of 2 or more, over time scales of months to years. Variability seems to be of larger amplitude toward shorter wavelengths. For the power law component, the amplitude and time scale of variability of both flux density and linear polarization appears to be similar to those observed for the Blazars. Broad emission line variability has been claimed in a few QSOs, with a lag behind the continuum variations of less than a few weeks, suggesting light crossing times for the line-emitting regions much smaller than predicted by standard photoionization models. LINEAR POLARIZATION The high and variable polarization of the synchrotron-emitting core-dominant quasars contrasts with that in other QSOs, where the mechanism may be dust or electron scattering, or transmission by aligned grains. In the weak-core, radio-loud quasars, there is weak linear polarization (p*1-2%) that tends to be aligned with the radio structure. In a few well-observed cases it is the continuum, not the emission lines, that is polarized. One exceptional weak-core, radio-loud quasar, OJ 287 (0752+258), shows p*8%, independent of wavelength and aligned in the direction of the radio lobes. The rare broad absorption line QSOs and infrared selected QSOs (i.e., detected by the IRAS satellite) are the only radio-quiet QSOs to show high polarization (p up to 20%, and steady), and a large fraction have measurable polarization (p>1%), suggesting a possible relation between these classes. RELATIONS BETWEEN OBSERVED QUANTITIES The most important observed relationship is probably the strong, almost linear relation among the line and continuum luminosities. When combined with AGN of lower luminosity, this relation extends over a factor of 10* in luminosity. This is most readily explained in terms of photoionization of the emission line regions by the QSO ultraviolet continuum. The next most important is the small but definite departure of the previous relation from a linear one-the "Baldwin effect." The equivalent widths of the broad C ** *1549 emission line become smaller with increasing continuum luminosity. Some other lines show slightly different dependences on luminosity. There is a possible trend for higher-luminosity QSOs to have broader C ** lines (FWHM). The trend may be most marked, however, for the full width of H*, measured at zero line intensity, and has been used to deduce a mass-luminosity relation, consistent with accretion near and below Eddington luminosities. EMISSION LINES AND RADIO PROPERTIES An important radio parameter is R, the ratio of the radio core luminosity to the luminosity in the extended radio emission (usually double lobes). In the relativistic beaming hypothesis, high values of R imply strong beaming toward the observer. There is an anticorrelation between widths of broad lines (at least H* and C ** *1549) and R, the dominance of the radio core-suggesting dominant motions of emission line gas in a plane perpendicular to the direction of radio beaming. There are anticorrelations between R and the ratios of emission line strengths to local or x-ray continuum intensities, supporting beaming of these continua. Also relevant to the geometry and dynamics of the emission line and continuum emission regions are positive correlations between broad line widths and equivalent widths, although these are not well understood. The only clear dependence on radio properties is in the strength of the optical power law continuum mentioned previously. It has been suggested that the blended optical Fe ** lines are stronger in core-dominant (high R) than in lobe-dominant quasars, and in radio-quiet compared with radio-loud QSOs. The greater strength of [O ***] emission in lobe-dominant quasars is also controversial. It remains to be seen whether these are real differences, or a dependence on some other property, such as intrinsic luminosity. Additional Reading Burbidge, C. and Burbidge, M.(1967). Quasi-Stellar Objects. W.H. Freeman, San Francisco. Courvoisier, T. and Mayor, M. eds.(1991). Active Galactic Nuclei. Springer, Berlin. Elvis, M.(1989). The ultraviolet continua of active galactic nuclei. Comm. Ap. 14 177. Gondhalekar, P.M., ed.(1987). Emission Lines in Active Galactic Nuclei. Seventh Workshop on Astronomy and Astrophysics, RAL-87-109. Rutherford Appleton Laboratory, Chilton. Osterbrock, D.E. and Miller, J.S., eds.(1989). Active Galactic Nuclei. IAU Symp. Kluwer Academic Publishers, Dordrecht, 134. Strittmatter, P.A. and Williams, R.E.(1976). The line spectra of quasi-stellar objects. Ann. Rev. Astron. Ap. 14 307. Weedman, D.W.(1986). Quasar Astronomy. Cambridge University Press, Cambridge.