The spectroscopic identification of the Extended Medium Sensitivity Survey (EMSS) sources now provides ∼350 X-ray selected QSOs and Seyfert galaxies (AGN in our working definition) and 30 BL Lac objects. Almost all of the AGN are spectroscopically similar to AGN found by other means but a few resemble normal galaxies so closely that they would not be identified as AGN easily by any other method.

The most definitive information presently available on X-ray source populations comes from surveys carried out with the HEAO program (i.e., the all-sky HEAO-1 mission and the HEAO-2 Einstein Observatory) and EXOSAT. The large-sky/low-sensitivity survey (LSS) at high galactic latitudes performed with the HEAO-1 A2 experiment (Piccinotti et al. 1982) is complete and optically identified down to the level of about a millicrab. The LSS overlaps in spectral coverage (2–50 keV) with observations of the cosmic X-ray background (CXB) obtained with the same detectors (Boldt 1987). CXB surface brightness fluctuations observed in this band have been used to set constraints on the Log N–Log S relation for sources an order of magnitude fainter than those resolved in the LSS (Shafer 1983).

A model for the origin of the cosmic X-ray background (hereafter XRB) is presented. The component of the background left after subtraction of the known classes of sources is explained by emission from a population of black hole sources at the redshift of z ∼ 4–5. The model is presented in more detail elsewhere (Zdziarski 1988). Here, we summarize its most important results.

A method is suggested for estimating the beamed and unbeamed components of the X-ray emission from quasars. It is shown that the beamed component correlates well with the core radio luminosity, while the unbeamed component shows no such correlation.

The theory of diffusive acceleration at shock fronts, which applies only if the fluid speed is nonrelativistic, yields a simple formula for the power-law index s of accelerated particles: s = 3r/(r – 1), where r is the compression ratio of the shock front. Although the acceleration process depends on there being effective pitch-angle scattering of the particles in both the upstream and downstream regions, no property associated with this process appears in the formula. Unfortunately, if the velocity of the fluid through the shock front is relativistic, as seems to be the case in the central engines of AGN's, and also in some hot-spots in the outerparts of their jets, this attractive property ceases to hold. To find the index s, it becomes necessary to develop specific models describing the scattering process. The physical reason for this is that the particle distribution close to a relativistic shock is anisotropic. The exact type of anisotropy depends on the properties of the pitch angle scattering and determines both the average energy gain per shock crossing, as well as the probability of escape downstream. In this paper, results are presented for the pitch angle diffusion resulting from scattering in a weakly turbulent plasma with a Kolmogorov spectrum of Alfvén waves moving parallel and antiparallel to the magnetic field. This kind of spectrum has been employed in nonrelativistic models of hot spots [1]. However, the results obtained tend not to vary too dramatically as a function of the turbulence spectrum, being less than about 0.2 in the resulting s (0.1 in the predicted synchroton spectral index) for turbulence spectra between k−1 and k−2 [3].

Many active galactic nuclei and/or associated jets and hot spots show a sharp cutoff in their nonthermal emission near ∼3 1014 Hz. Using the first order Fermi theory for particle acceleration at shocks we demonstrate that synchrotron emission and shock acceleration combine to give a maximum cutoff near to the observed frequency. We explore consequences: First, CCD-observations of the M87 jet demonstrate that the cutoff frequency decreases along the jet. Second, the detailed shape of the cutoff spectrum depends on the precise model for the turbulent wave field. Thirdly, an X-ray Synchrotron contribution from p-p and p-γ interactions exists which can match observed AGN spectra. Finally we apply the model to interpret the origin of very high energy cosmic rays, to the spectral behaviour of quasars and BL Lac's and deduce neutrino emission spectra of active galactic nuclei.

A substantial fraction of the radiation from an active galactic nucleus (AGN) is apparently nonthermal in origin, and is probably produced by ultrarelativistic electrons. How much energy goes into relativistic protons is uncertain, but it is likely to be comparable to the electron energy or larger. Indeed, several authors (Sikora et al. 1987; Kazanas and Ellison 1986; Zdziarski 1986) have shown that proton-photon and proton-proton collisions can be efficient sources of relativistic pairs in the central engine of an AGN. Thus it is not necessary for electrons to be accelerated directly in AGNs, provided that protons are accelerated with high enough efficiency.

A growing body of evidence from stellar dynamics in the nuclei of galaxies indicates that supermassive black holes of 107–109 M0 are common. The two best cases are M31 and M32, for which dark, central mass concentrations are the only straightforward interpretation. M87 continues to be a possible location of an even more massive black hole, but new observations and models by the author and D. Richstone effectively rule out the high black hole mass ∼5 × 109 M0 claimed by Sargent, Young, and collaborators. New data are available for several other nearby galaxies which also show kinematic signatures that could also be due to supermassive black holes. The Hubble Space Telescope will play the key role in strengthening these cases and eliminating, for the best examples, alternative models which do not require supermassive black holes.

The observed evolutionary behavior of active galactic nuclei is compatible with a model in which black holes form in the nuclei of new-born galaxies and then grow at a rate limited by both radiation pressure and the supply of gas. Individual sources become more luminous with time as long as they are being fueled. However, the rapid decrease in the mean rate of supply of gas causes a strong decline in the space density of active objects. Nearby galaxies should harbor modest size (∼ 106 – 108 M⊙) black holes. It is suggested that the gas that fuels high redshift quasars is mostly derived from the host galaxy.

The formation and evolution of massive binary black holes in galactic nuclei is discussed in a galaxy merger model for the activation of galactic nuclei. Mass flow into a galactic nucleus induced by the merger causes the simultaneous tightening of a preexisting binary and an increase in the fueling level of the central hole(s). The model predicts that during the final, short and most luminous phase of an AGN the precession period of the central hole will rapidly become as short as 103 yr. Some observational implications are mentioned.

We have refined our calculation of the line profile of a relativistic, Keplerian disk by incorporating a variety of emissivity laws, as well as broadening due to turbulence or electron scattering. The significant improvement in the fit to the double-peaked Ha line profile of the elliptical radio galaxy Arp 102B provides the most convincing direct evidence for an accretion disk in any AGN. Arp 102B appears to be a low-luminosity analog of 3C 390.3, and several lines of evidence point to the existence of small, hot ion tori illuminating an outer thin disk in both of these galaxies. The rarity of these emission-line profiles might be understood if this particular combination of ion torus/thin disk occurs only for a narrow range of ṁ(= Ṁ/ṀEdd).

We have made detailed calculations of the structure and the spectrum of massive, geometrically thin “bare” accretion disks. The calculations are for an α-disk with various assumptions on the viscosity law. The radiative transfer was treated with the Eddington approximation for an atmosphere with a vertical temperature gradient. All significant sources of opacity, for T>104K, are included, and all models are found to be optically thick throughout. Spectral modifications due to electron scattering (modified blackbody and comptonization) are not significant in most cases. The requirement of a geometrically thin accretion disk forces a limit of L<0.3Led on the accretion rate. Several previous disk calculations violate this limit (Malkan 1983, Czerny & Elvis 1987, Bechtold et al. 1987) and their results are questionable. The surface temperature is close to the effective temperature, even for regions where electron scattering effects are significant. This is due to the vertical temperature gradient and is in contradiction to earlier findings. The angular distribution of the ionizing flux is strongly influenced by general relativistic effects, and can be very different for various disks.

We model the infrared through gamma ray AGN spectrum with a combined accretion disk and nonthermal source model and fit the spectra of a number of sources. The accretion disk model of Sun and Malkan (Sun 1987) adds relativistic effects to the emission from an optically thick, geometrically thin accretion disk; further details can be found in Sun and Malkan (1988) in these proceedings. In the nonthermal source model of Band and Grindlay (1986) relativistic electrons radiate the infrared continuum by synchrotron emission, and the X-ray spectrum by inverse Compton scattering of ultraviolet photons from the accretion disk. The electron distribution consists of a flat (n ∝ γ∼−2.4) low-energy component, and a steeper (n ∝ γ∼−3.4) high-energy component.

If accretion disks are present in AGN and extend to large radii they should contribute substantially to the Broad Line emission. The outer regions of the disk are indeed illuminated by a small amount of ionizing radiation. X-rays are emitted by the central inner region near the black hole, and they are either received directly by the outer disk, owing to its “flaring” shape (Cunningham, 1976), or partly reflected towards the disk by a hot Compton thin medium (Begelmann and McKee, 1983). X-ray photons are also produced through the Inverse Compton mechanism in compact radio sources located above the disk(“jet model”).

After the suggestion (Shields 1978) that some AGN emission might arise in an opaque accretion disk around a supermassive compact object, several papers (e.g., Malkan and Sargent 1982; Malkan 1983; Bechtold et al. 1987) have interpreted the flat ultraviolet continuum (“big blue bump”) observed in many AGN spectra, in terms of such a model. The early calculations approximated the radiation locally emergent from the accretion disk as blackbody; the more recent calculations (e.g., Czerny and Elvis 1987; Wandel and Petrosian 1988) have treated this emission as (electron-scattering) modified (possibly comptonized) blackbody.

From continuum and line intensity light-curves in the ultraviolet range, we find that the CIV peak-emission lies at about 2 ± 0.5 light-month from the centre. Line profile variations in the Ha and CIV emission suggest that the BLR is made of two regions with distinct kinematical properties:

1. • a spherically distributed set of clouds, responsible for the broad HIL emission and responding to ionizing flux variations with a time-lag of 2 light-months at most

2. • a disc-like structure from which most of the LIL emission arises, with a slow response to ionizing flux changes: hence larger is size.

Geometrically thin and optically thick accretion disk models which use non-LTE stellar atmospheres predict a Lyman absorption edge near 912Å in the quasar rest frame (Sun and Malkan 1987; Sun 1987). The detailed spectral shapes around the Lyman limit have been calculated for both Kerr and Schwarzschild metrics. The Lyman absorption edge, especially in the (more realistic) Kerr case, is markedly distorted and displaced by the relativistic effects of the strong gravitational field of the black hole and the large Doppler motions of the disk surface. Figure 1 shows a Kerr disk model accreting at 10% of its Eddington luminosity (LEdd), typical for quasars, viewed from five different angles.

In this paper we present recent observational data on the flat spectrum double (1 arcsec.) radio source 1830-211 (Rao and Subrahmanyan, 1988, MNRAS, 231, 229) and discuss why gravitational imaging appears to be the most plausible explanation for its unique nature.

Gravitational micro-lensing can be used to obtain information 1) on the inner structure of active galactic nuclei and 2) on the distribution of compact masses in the range 10−4 … 102 solar masses. A regular monitoring of the most promising candidates for micro-lensing should be carried out. The strongest candidate is the macro-lensed quasar 2237+030, for which we expect a change in the luminosity ratio between the images of about 5 per cent per year.

We are conducting an optical imaging search for gravitational lenses. A sample of known quasars was selected on the basis of high redshifts and apparently large absolute luminosities; these selection criteria significantly enhance the probability that a given QSO is gravitationally magnified. Our search has yielded so far at least one lensed QSO, and several promissing candidates.