F. Duccio Macchetto
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218. On assignment from the Space Science Department of ESA
The HST results are in broad agreement with the unified theories for AGN. However, these results have also made important modifications to the basic theory, such as the need for additional sources of ionization (e.g., shocks) to explain the observed filamentary structure in the Narrow-Line Region.
The optical counterparts to radio-jets, such as M87, have been studied in detail. Recently, HST observations have revealed, for the first time at optical wavelengths, proper motions in the M87 jet, with Doppler-boosted velocities, as high as 2.5c. Observations over a longer time interval, should reveal the true propagation velocity of the jet.
Finally, HST observations of the complete 3CR sample of radio sources has led to the discovery of several new optical counterparts to radio jets, more than doubling the sample of known objects.
The investigation of the physical properties of the nuclear regions of active galaxies has been the subject of a number of HST observing programs. In the now standard paradigm for AGN, the basic differences between the different classes of objects is simply explained as a result of different orientations to the line-of-sight. In the unified picture, a central energy source, generally assumed to be a massive black-hole, is surrounded by two spatially and kinematically distant regions. The Broad-Line Region (BLR) has scales of the order of a parsec in diameter and emits broad permitted emission lines with widths up to 10000 km s. The Narrow Line Region (NLR) emits narrow (a few hundred km s) permitted and forbidden lines and has sizes of up to a kiloparsec in diameter. The orientation effects and, therefore, the classes of AGN, are determined by an optically thick torus composed of dense molecular clouds, whose inner diameter is comparable to the size of the BLR and can extend for tens of parsecs. The symmetry axis of the torus determines the direction of the kiloparsec scale radio jet and is independent of the orientation of the galactic disc. In the standard model, if the torus is seen face on, we can see the continuum source and BLR directly and the galaxy is classified as a Seyfert 1. If, instead, the torus obscures the central regions, the galaxy is classified as a Seyfert 2.
The HST observations to date seem to confirm the broad validity of this picture. However, the high spatial resolution observations as well as the important new imaging polarimetry results have shed new light on the fundamental physical phenomena that are at play in the nuclear regions of active galaxies.
Closely related to this problem is the study of the optical counterparts to the radio jets. We know that these jets play a fundamental role in transporting energy from the central source to the extended radio lobes. Observations of optical and ultraviolet wavelengths with the HST are essential to obtain spatial resolutions similar to, or better than, those achieved in the radio band and, thus, provide the possibility of directly comparing the sites and mechanisms responsible for the emission at these different wavelengths.
In all cases to date, the emission has been attributed to the synchrotron mechanism, and since the electron lifetime is a strong function of the observed frequency, observations at optical and ultraviolet wavelengths offer the possibility to determine the precise location where particle acceleration occurs. Comparison of the radio and optical morphologies further allows the study of the confinement mechanisms and diffusion processes within the jet. A number of important discoveries and observations that place the theoretical models on firmer observational grounds have been published or are about to be published.
In this review, I will discuss statistical studies of a sample of Seyfert galaxies and then discuss in detail some important examples for each of the main categories of Seyfert galaxies. In the last section, I will highlight important HST discoveries related to the optical counterparts to radio jets.
Nelson et al. (1995) have carried out a SNAP program to observe 52 Markarian Seyfert and 48 Markarian non-Seyfert galaxies with WFPC1. They find that the nuclei of type 1--1.5 Seyfert galaxies are dominated by strong point sources, while those of Seyfert 2 and non-Seyfert Markarian galaxies tend to be resolved, less distinguished, and similar in shape to the luminosity profiles of normal galaxies. This result confirms earlier FOC observations (Boksenberg et al. 1992) of the UV properties of Seyfert nuclei which showed that only Seyfert 1--1.5 nuclei were detectable at UV wavelengths in line with the predictions of the unified model.
There are two possible interpretations of this result for type 2 Seyfert galaxies. The first is that their nuclear continuum sources are undetected, contributing less than 10% of the nuclear light (within 0.5 radius) in all cases. The second is that their nuclear components are resolved and blend in smoothly with the brightness profile of the host galaxy's bulge. Since spectroscopic studies support typical nuclear continuum fractions distinctly greater than 10%, the latter conclusion is clearly preferable.
If the continua observed in Seyfert 2 galaxies originate as nuclear light which is redirected into the line-of-sight by scattering, as predicted by unified models of AGN, then the scattering regions must be extended. Simple simulations suggest that these regions must cover several tens of parsecs, or more, in agreement with estimates of the sizes of the scattering ``mirrors'' in other Seyfert 2 galaxies. However, the similarity of the profiles of non-Seyfert Markarian and type 2 Seyfert nuclei suggests that circumnuclear star formation may also be an important component in the nuclear profiles of the latter.
Korista et al. (1995) have conducted a combined HST/IUE/ground-based spectroscopic monitoring campaign on the Seyfert 1 galaxy NGC 5548. IUE spectra were obtained once every 2 days for a period of 74 days beginning on 1993 March 14. During the last 39 days of this campaign, spectroscopic observations were also made with the FOS on a daily basis. During the HST portion of the monitoring campaign, the 1350Å continuum flux is found to have varied by nearly a factor of 2. In other wave bands, the continuum shows nearly identical behavior, except that the amplitude of variability is larger at shorter wavelengths, and the continuum light curves appear to show more short-timescale variability at shorter wavelengths. The broad emission lines also vary in flux with amplitudes that are slightly smaller than the UV continuum variations and with a small time delay relative to the UV continuum.
On the basis of simple time-series analysis of the UV and optical continuum and emission-line light curves, they find (1) that the ultraviolet and optical continuum variations are virtually simultaneous, with any lag between the 1350Å continuum and the 5100Å continuum amounting to less than about 1 day; (2) that the variations in the highest ionization lines observed, He II 1640 and NV 1240, lag behind the continuum variations by somewhat less than 2 days; and (3) that the velocity field of the C IV-emitting region is not dominated by radial motion.
The optical observations show that the variations in the broad H line flux follow the continuum variations with a time lag of around 2 weeks, about twice the lag for L and C IV, as in their previous monitoring campaign of this same galaxy. However, the lags measured for L, C IV and H are slightly smaller than previous determinations.
They confirm two trends reported earlier, namely, (1) that the UV/optical continuum becomes ``harder'' as it gets brighter and (2) that the highest ionization emission lines have the shortest lags, thus, indicating radial ionization stratification of a broad-line region that spans over an order of magnitude range in radius.
NGC 4151 is the nearest (30 Mpc) example of a (sometime) type 1 Seyfert galaxy although it is rather a low-luminosity example of the class. The broad emission component of H, which is a characteristic of Seyfert 1 galaxies, varies dramatically and, in a low state, can almost disappear. This led to NGC 4151 being reclassified as Seyfert 1.5. Both the permitted and the continuum emission show variations on time scales as short as days. FOC observations in [O III] and nearby continuum have shown that the NLR is resolved into a number of emission-line clouds (sizes 10 pc) with elongated morphology. These are distributed in a biconical structure with apices coincident with the central point source and a cone opening-angle projected on the sky of 75. The cone position angle of is aligned with the extension of the nuclear VLBI radio source, suggesting that the same mechanism may align both the optical ionizing radiation field and the parsec scale radio structure.
Long-slit spectroscopic observations were carried out with the (uncorrected) FOC (Boksenberg et al. 1995). They find a high and a low radial velocity component within the narrow emission lines and identify the low-velocity component with the bright, extended, knotty structure within the cones, and the high velocity component with more confined diffuse emission. Also present are strong continuum emission and broad Balmer emission-line components, which are attributed to the extended point spread function arising from the intense nuclear emission.
Adopting the geometry pointed out by Pedlar et al. (1993) to explain the observed misalignment of the radio jets and the main optical structure they model an ionizing radiation bicone, originating within a galactic disk, with apex at the active nucleus and axis centered on the extended radio jets, and confirm that through density bounding, the gross spatial structure of the emission-line region can be reproduced with a wide opening angle that includes the line-of-sight consistent with the presence of a simple opaque torus allowing direct view of the nucleus. In particular, the modelling reproduces the observed decrease in position angle with distance from the nucleus, progressing initially from the direction of the extended radio jet, through the optical structure, and on to the extended narrow-line region.
Boksenberg et al. (1995) explore the kinematics of the narrow-line, low- and high-velocity components on the basis of the spectroscopy and adopted model structure. For the low-velocity system, both Keplerian rotation and isotropic outflow (or outflow confined to the ionizing cone) give plausible correspondence with the data. If interpreted as rotation, it shows consistency with earlier determinations indicating a central mass concentration of about 10 M. The high-velocity system kinematically conforms to radial outflow with the galaxy disk, although this does not well reproduce the observed intensity structure.
Another interesting example of a Seyfert 1.5 galaxy is Mrk 6, at a distance of 11 Mpc where 0.1 correspond to 53 pc. FOC observations (Capetti et al. 1995a) reveal a jet-like feature in the emission-line which extends about 0.5 or 250 pc, from the nucleus. The optical jet is cospatial with the Southern radio-jet, as revealed by a 6 cm MERLIN image matching the HST angular resolution, and shares with it a similar curved morphology. The emission-line jet shows a low-ionization halo surrounding a higher ionization core, which is cospatial with the radio jet. This suggests that the optical jet and the NLR structure are generated by cylindrical shocks created by the expansion or hot material around the radio-jet, which compress and heat the interstellar gas. The density stratification in the shocked gas is responsible for the inner ridge of higher ionization observed in Mrk 6. Most of the radio and optical emission originates in the innermost 110 pc, which suggests that radio and optical emission is closely associated on this small scale. These results strongly support the interpretation that the structure of the NLR of Mrk 6, and of Seyfert galaxies in general, is dominated by the interaction between the radio-ejecta and the surrounding medium.
A number of observations of the prototypical Seyfert 2 galaxy NGC 1068 (D 22.7 Mpc, 0.1 = 11 pc) have been carried with the FOC (Macchetto et al. 1994, Capetti et al. 1995b, 1996a). These include visible and UV continuum and emission-line observations, as well as FOC imaging polarimetry (F253M, F372M & F501N). These observations show that the inner morphology is very complex. In first approximation, it appears as a ``bi-cone'', with the axis of symmetry changing with distance from the nucleus. This may be due to the different location of the gas as it streams around the inner bar. Alternately, the ionization and scattering cone is rather narrow and tracks the radio jet more closely. The brightest feature in the continuum and line emission images is ``Cloud-B'' which is highly polarized (65%) and contrary to previous claims is not the location of the nucleus. The polarization data shows that the true nucleus is located some 0.6 South of Cloud B and is obscured. This is very consistent with the unified scheme of AGNs. A strange feature, the ``twin crescent'' is also highly polarized (45%). The distance from the active nucleus 0.1 makes it unlikely that it has any physical relations to it and, thus, it remains a mystery. The high-degree of polarization in the circumnuclear region implies that most, if not all, of the observed UV light is scattered light from the nucleus, with either dust or electrons providing the scattering medium.
Recent high precision ground-based and HST astrometry (Capetti, Macchetto & Lattanzi 1996) has allowed the alignment of the optical and radio data to be carried-out. The maps have been registered with a precision of better than 0.030 (Figure 1).
Figure: The FOC image of NGC 1068 aligned to the MERLIN radio contours. The position of the optical nucleus falls between the two peaks of radio emission. The image is 7 high.
With this determination the position of the obscured nucleus, as derived from the polarization measurements, falls within the S1 and S2 peaks of the radio data, but are not coincident with any of these two peaks, whereas it is within 1 of the position of the HO maser source.
Capetti et al. (1996b) have carried out observations of Mrk 573, (z = 0.017, 0.1 = 49pc). The [OII], [OIII] and emission images show an extended NLR. The large scale emission line morphology is dominated by bright arcs at 1.5 from the nucleus and fainter arcs at 4 (200 kpc). The continuum emission with F814W shows a bar with 10 radius (4.9 kpc) superposed on an elliptical bulge. At F550M, there is a fully resolved central knot and a dust lane 1 (500 pc) long and 0.1 (50 pc) wide. As in the other Seyfert 2 galaxies, there is no radiation from the nucleus itself.
Capetti et al. (1996b) have observed Mrk 348 (z = 0.015; 0.1 = 43pc) in [OII], [OIII] and continuum. In this case, the line emission is confined to a linear structure 0.45 long. The northern section is highly collimated, extends to 0.2 then bends and terminates in an orthogonal feature. The southern component appears as an arc-like lobe with a bright knot superposed. The continuum image is very similar to the [OIII] image. There is no emission from the nucleus itself.
Capetti et al. (1996b) have observed Mrk 78, (z = 0.038, 0.1 = 110pc). The [OII] and [OIII] images show that the NLR is very extended at low surface brightness and has a double-lobed structure. The Eastern lobe is relatively bright and compact (0.9 1.1), while the Western lobe is fainter and shell-like (1.5 2.3). The ionization ratio [OIII]/[OII] is rather high (2.9) in the eastern lobe and lower (1.4) in the Western lobe. At F550M, the continuum is elongated and has two diffuse patches cospatial with the high excitation regions. There is no point-like source at the position of the nucleus and the faint diffuse emission is fully accounted for by a normal starlight distribution in the bulge.
Capetti et al. (1995c) have observed Mrk 3 (z = 0.0133, 0.1 = 39pc), in [OII], [OIII] and continuum. The emission-line shows an ``S'' shaped structure with sheets and filaments. The western end is coincident with the radio hot-spot. The narrow absorption lane perpendicular to the NLR is very evident and hides the nucleus. Continuum observations show a small scale bar, nearly perpendicular to the bulge (Figure 2).
Figure: The FOC image of Mrk 3 superposed with the radio contours shows the alignment of the optical emission and radio jet.
These and other HST observations of the nuclear regions of Seyfert 2 have been discussed by Capetti et al. (1996b). They show that in all cases the physical structure of the NLR in Seyfert 2's is closely related to the radio-emission. The optical morphology is dominated by the interaction with the radio ejecta. The NLR appears to take a different form depending on the structure of the radio emission. Where there are radio lobes, there are shell-like emission-line structures. For Mrk 573, Mrk 78 and Mrk 348, the emission-line structures are bow-shocks. Where a collimated radio jet is present, the morphology is different. In Mrk 3, the NLR follows the jet morphology. Capetti et al. (1996b) show that this dichotomy implies that bow-shock emission-line structures are produced by the sweeping-up of gas at the advancing working surface of the ejected radio plasma. The corrugated structure indicates that instabilities have developed in the compressed gas. Where a jet is apparent, it is surrounded by a halo of hot gas which expands radially from the jet axis and the emission line forms a cylindrical cocoon on the outer cooling surface. In all cases, the ionization conditions, as determined for example from the emission-line ratios, are such that the ionization parameter increases with distance from the nucleus. This requires a source of ionization in addition to the unclear ionizing flux.
Capetti et al. (1996b) show that shock mechanisms are the best candidates to produce the relatively small amount of locally produced excess radiation. They are fully consistent with both the measured ionization conditions and the observed filamentary morphology.
PKS 0521-36 is a relatively isolated radio galaxy at a redshift z=0.055 which also harbors a bright V = 16 BL Lac nucleus and extended optical line emission. Sparks et al. (1990) reported optical polarization measurements of the jet and nucleus, which confirmed that the emission is due to synchrotron radiation.
The FOC images obtained in 1990 with the FOC (Macchetto et al. 1991a) show a bright knot located 1.8 to the NE and clearly resolved as in the VLA data (Keel 1986). The width of the knot is 0.8. Beyond this bright knot, the jet has approximately constant surface brightness and a morphology similar to the VLA image with a total length of 6.5. The jet is also resolved in width, 0.6 wide in the fainter regions of the jet, with little or no evidence of structure on a scale of 0.1. The FOC data appears to show more flux than the VLA data in the region at slightly larger radius from the nucleus but close to the southern tip of the knot. A large bright knot further along the jet is a clear counterpart to the radio knot. The bright knot is an important site where particle acceleration is occurring.
Using standard formulae and values for the relevant parameters, we derive a mean lifetime for the electrons of 600yr. This implies that there must be continuous acceleration along the jet of the electrons responsible for the optical emission, since electron diffusion from the bright knot could not account for the observed optical extent.
3C 66B is a relatively nearby bright radio source associated with a 13th magnitude galaxy at a distance of 86 Mpc. At this distance an angular scale of 0.1 corresponds to a projected linear size of 41 pc. Images of 3C 66B were obtained with the FOC and compared with the best VLA map of 3C 66B. Several conclusions can be drawn from these observations (Macchetto et al. 1991b).
On the scale of the HST resolution, the jet of 3C 66B is filamentary. Two distinct ``strands'' can be traced from 3.7 (1.5 kpc) from the nucleus out to a distance of 7.6 (3 kpc), where they disappear into the noise. The separation between the strands varies between about 0.3 and 0.4, that is about 150pc, and they appear to undergo sharp ``kinks'' at distances of 2.5 (1.0 kpc) and 6.2 (2.5 kpc) from the nucleus.
The origin of these kinks is unclear. The fact that they are mimicked in more than one filament suggests that they are not due to an instability mode in an individual filament. They may well trace out irregularities in the ISM of the galaxy and/or be due to time-dependent variations in the power of the nuclear source responsible for producing the jet.
3C 273 is one of the nearest and brightest quasars known. Its jet has been extensively studied at radio and optical wavelengths. Using the FOC, Thomson, Mackay & Wright (1993) have carried out high-resolution imaging polarimetric observations of the jet. More recently, Bahcall et al. (1995) have obtained WFPC2 as well as Merlin observations with matching resolution. The projected jet length is more than 70 kpc (H Mpc. The width is only a few tenths of an arcsecond (. The optical emission is highly confined to the core of the radio jet. It runs along the ridge of the radio emission and is asymmetric compared to the radio. All oblique radio features coincide with optical knots, though there are some optical features without radio counterparts.
Bahcall et al. (1995) suggest that the `radio jet' consists of two components. The first is the fast-moving jet, shown by the coincident oblique radio and optical features. The second consists of the emission from a surrounding, slow-moving ``cocoon''. The radio data suggests that the oblique radio features coincident with the optical knots may be in the form of a helix. If the velocity is relativistic, the emission will appear brightest where the velocity vector is closest to the line-of-sight, the enhancement being independent of wavelength. This would explain the close correspondence between the optical and radio knots. The helical form may arise from a driven Kelvin-Helmholtz instability.
NGC 3862 (3C 264) at a distance of 86.2 Mpc is an FR1 source. This galaxy was observed in two bands with the FOC. The images showed a prominent jet-like feature emanating from the nucleus (Crane et al. 1993) morphologically similar to the jets seen in M87 and 3C 66B. Indeed, the bifurcation seen at the end of the jet is reminiscent of the ``double-stranded'' feature seen in the HST image of 3C 66B (Macchetto et al. 1991b).
Broad-band (F702W) imaging data taken with the WFPC2 shows a very intriguing feature in this galaxy (Sparks et al. 1996). In addition to the very evident optical jet, an almost perfectly circular ring is observed. It is not possible with a single band observation to decide whether this is an emission-line ring, or even if it is a true emission feature or the result of dust absorption closer to the nucleus. The fact that the jet appears to stop suddenly at the position of the ring may be indicative of some physical association between the two features. Observations in other bands will, hopefully, help clarify this mystery.
The giant elliptical galaxy M87 contains the closest extragalactic jet, which makes it a prime target for studies of jet structure and kinematics. Optical and ultraviolet observations of M87 have been carried out with HST and have been extensively reported (Macchetto 1991, Boksenberg et al. 1992, Macchetto, Biretta & Sparks 1992).
While the radio and optical images present a remarkable degree of similarity, there are nevertheless significant differences. The optical/UV images show intrinsically higher contrast than the radio, with compact regions of emission localized within the knots. The jet is narrower in the optical/UV, and more concentrated to the jet center in the optical/UV than in the radio. The radio-to-optical spectral index of the inter-knot regions is steeper than that of the knots themselves. There are also differences in the detailed knot structure of the optical emission compared to the radio, and there is a weak overall spectral steepening with distance from the nucleus beyond knot A.
Capetti et al. (1996c) have analyzed polarization observation of the M87 jet taken with the FOC in the ultraviolet and with the WFPC1 in the visual. The degree of polarization is typically 30% over most of the jet. At the edges of the jet the polarization is as high as 60%, requiring a highly ordered magnetic field. In the center of the jet the small scale structure of the magnetic field produces significant cancelation reducing the polarization to 10%. The degree of polarization and the polarization pattern are very similar at radio and optical wavelengths. No significant depolarization or Faraday rotation have been detected, in agreement with previous radio determinations. However, the morphology of knot D is considerably different in the VLA observations by Owen et al. (1989) and these HST observations. Knot D1 appears to be relatively brighter and closer to the nucleus in the optical than in the radio images. Capetti et al. (1996c) conclude that this component is associated with a shock front. At the location of a shock front acceleration of relativistic electrons occurs, enhancing the synchrotron emission at shorter wavelengths, and the transverse component of the magnetic field is amplified by the compression produced by the shock.
Figure 3 shows FOC images of the nucleus obtained in August 1994 and July 1995 by Biretta et al. (1996). The images, which have been aligned at the bright point source show a feature approximately 150 milliarcseconds (12 pc) from the core, which corresponds roughly to knot M in the 18 cm VLBI image of Reid et al. (1989). Comparison of the two images shows that this feature moves outward by 8.5 mas or 0.66pc, for a velocity of 2.3c+0.3c. From this result it is clear that the jet is already relativistic within the core, and that acceleration is probably unimportant on these scales. Velocity measurements for three features in the region of knot D were also obtained. We find a speed of 2.6c+0.4c for feature DW, and a somewhat larger 3.7c+0.5c for DM. A third feature, FOC1, which was not seen in the radio data, has a speed of 2.1c+0.3c. Two additional features, labeled FOC2 and FOC3, appear to fade from view between 1994 and 1995. This fading rate is much higher than typical synchrotron lifetimes in the jet (100yr) suggesting that the adiabatic expansion of the emitting region is likely to play an important role in the evolution.
Figure: FOC images of the M87 jet. The lower panels show evidence for proper motion over an 11 months baseline. See the text for details.
Sparks et al. (1996) have very recently reported the discovery of an optical synchrotron jet emanating from the nucleus of NGC 1218, the galaxy hosting the radio source 3C 78. The jet is similar to that of M87, although smaller in projected length. The observations were obtained as part of the HST ``snapshot'' program to acquire broad-band images with high spatial resolution of 3CR radio sources with the WFPC2 F702W filter.
The visible jet has a total length of 1.37 which is 0.75 kpc in projection for = 75 Mpc. Within 0.5 of the nucleus, the jet is essentially unresolved across its width even at PC resolution, although it is slightly curved and there are distinct knots within that portion of the jet. The gap between the brightest part of the jet and the nucleus appears to be real. At 0.5 , the jet fades dramatically, and fans out into a broad plateau of emission. There are also two discrete knots, and a possible third fainter one, within the plateau portion of the jet, seen in the optical data.
The HST observations conducted so far have already led to the discovery of two new optical jets (3C 264 and 3C 78) and the identification of at least six other candidates, which will be further studied in the future.
Although few in number, there are common features shared by the optical jet radio sources: (i) all have relatively prominent nuclei in both radio and optical domains; (ii) the nuclei all have flat radio spectra; (iii) the jets are small compared to typical radio jet dimensions, with the arguable exception of 3C 273; (iv) there are no two-sided optical jets---they are all one-sided with a large jet to counterjet lower limit; (v) there is noticeable jet curvature; (vi) in addition, where measured, the optical emission is more localized than the radio and the optical jet is narrower than the radio jet.
An obvious candidate explanation for most of these characteristics is relativistic `beaming', in which the jet becomes visible in the optical only when pointing towards the observer. In the beaming picture, the jet appears brightened, foreshortened and with a prominent active nucleus. Beaming also blueshifts the synchrotron `break' frequency.
As an alternative to relativistic beaming, environmental effects may be considered. If the pressure is higher in the vicinity of the optically emitting sources, then the additional confinement may act to enhance their radiative luminosity while suppressing the growth of the source, thereby giving rise to the correlation between size and power. This does not immediately suggest an explanation for the core dominance and one-sidedness; however there may be instabilities which cause jet disruption and optical emission that are sufficiently rapid that only one side is visible at a given time. `Age' may provide yet a third alternative, with the optical jet sources being young, in the process of forcing their way out through the interstellar medium, and ram pressure playing a similar role in enhancing the visibility.
There are statistical uncertainties at present, however an extensive analysis of many more optical jets should provide results that will be essential in elucidating the nature of extragalactic synchrotron jets.
Bahcall, J.N., Kirhakos, S., Schneider, D.P., Davis, R.J., Muxlow, T.W.B., Garrington, S.T., Conway, R.G., & Unwin, S.C. 1995, ApJ, 452, L91
Biretta, J.A., Sparks, W.B., Macchetto, F., & Capetti, A. 1996 ApJ, submitted
Boksenberg, A. et al. 1992, A&A, 261, 393
Boksenberg, A., Catchpole, R.M., Macchetto, F. et al. 1995, ApJ, 440, 151
Capetti, A., Axon, D.J., Kukula, M., Macchetto, F., Pedlar, A., Sparks, W.B., & Boksenberg, A. 1995a, ApJ, 454, l85
Capetti, A., Axon, D.J., Macchetto, F., Sparks, W.B., & Boksenberg, A. 1995b, ApJ, 446, 155
Capetti, A., Macchetto, F., Sparks, W.B., & Boksenberg, A. 1995c, ApJ, 448, 600
Capetti, A., Macchetto, F., Axon, D.J., Sparks, W.B., & Boksenberg, A. 1996a, ApJ, in press
Capetti, A., Macchetto, F., & Lattanzi, M. 1996, A&A, submitted
Capetti, A., Axon, D.J., Macchetto, F., Sparks, W.B., & Boksenberg, A. 1996c, ApJ, in press
Capetti, A., Macchetto, F., Sparks, W.B., & Biretta, J.A. 1996c, ApJ, in press
Crane, P. et al. 1993, ApJ, 402, L37
Keel, W.C. 1986, ApJ, 302, 296
Korista, K. et al. 1995, AJ, 97, 285
Macchetto, F. 1991, Proceedings, Physics of AGN, ed. S.J. Wagner, W.J. Duschl, Springer-Verlag: Berlin, 325
Macchetto, F. et al. 1991a, ApJ, 369, L55
Macchetto, F. et al. 1991b, ApJ, 373, L55
Macchetto, F., Biretta, J.A., & Sparks, W.B. 1992, Proceedings 182nd AAS Meeting, 24, 1183
Macchetto, F., Capetti, A., Sparks, W.B., Axon, D.J. & Boksenberg, A. 1994, ApJ, 435, L15
Nelson, C.H., Mackenty, J.W., Simkin, S.A., & Griffiths, R. E. 1996, ApJ, 466, in press
Owen, F.N., Hardee, P.E. & Cornwell, T.J. 1989, ApJ, 340, 698
Pedlar, A., Howley, P., Axon, D.J., & Unger, S.W. 1993, MNRAS, 259, 369
Reid, M.J., Biretta, J.A., Junor, W., Spencer, R., & Muxlow, T. 1989, ApJ, 336, 125
Sparks, W.B., Miley, G., & Macchetto, F. 1990, ApJ, 361, L41
Sparks, W.B. et al. 1996, ApJ, submitted
Thomson, R.C., Mackay, C.D., & Wright, A.E. 1993, Nature, 365,133