We present new arcsecond-scale radio images of the gravitational lens system 0957+561 A, B. Observations at 1.6GHz were made in 1991 October with the VLA in A/B configuration with a resolution of 1.5 arcsec (Fig. 1, left). The lowest contour is 0.37 mJy/beam. In addition to the compact A and B image components, and the familiar NE/SW radio double-lobe structure surrounding image A, this map shows two interesting new features:

1. (a) a long, thin feature extending south and west from the NE lobe, some 10 arcsec in extent, confirming the detection by Avruch et al. (1994). It is reminiscent of the ‘arc’ features seen in optical images of lensing clusters, and hopefully can be used to constrain models of the cluster mass distribution. (For the superstitious, one can note that the position angle of the arc, 17°, is identical to that of the VLBI jet in image B, a few arcseconds to the west.)

2. (b) a low-level extension of the B image in the NW direction. We are not aware that this has been seen before. A possible interpretation of this feature is a continuation of the jet emission seen in VLBI images of B, whose counterpart in A is seen in higher resolution maps. The change of position angle between the VLBI and arcsec-scale jets would indeed have opposite parities in the A and B images.

CCD astrometry makes possible milli-arcsecond precision relative astrometry at faint magnitudes (Tinney 1994a). The measurement of proper motions for the nearest galaxies is therefore a project requiring only years – not decades. A prerequisite for such measurements, however, is a set of unresolved, extragalactic reference objects – ie. QSOs. While searching for such objects behind the Fornax dSph galaxy, Tinney (1995b) discovered the QJ0240-343AB system as a pair of bright, UV-excess objects (U-B = −1.0, B = 19.0 & 19.8) with a separation of 6.1″. Spectra obtained with the 3.9 m Anglo-Australian telescope and the RGO+FORS spectrographs (Fig. 1) have shown both components of the pair to be at a redshift of z = 1.4, with a rest velocity difference consistent with zero at a 1-σ limit of 180 km/s. The spectra also show a definite metal-line absorption system at z = 0.543 and a possible system at z = 0.337. No bright lens is seen, but the strong similarities in the spectra suggest a gravitational lens nature.

The source of the gravitational mirage MG J0414+0534, found in a radio survey (Hewitt et al. 1992), is an extremely reddened quasar at z ≃ 2.63 (Lawrence et al. 1995) but the redshift of its lensing galaxy is still unknown (≃ 0.8 ?). Whether the extinction occurs mainly in the lens galaxy or in the source itself is a matter in debate. Also, a discrepancy is observed between the radio and optical flux ratios for images A1 and A2, which could be due either to differential magnification near a caustic or to differential extinction (Schechter et al. 1992, hereafter SM92; Angonin-Willaime et al. 1994, hereafter AW94). The new IR and optical data discussed here could throw some light on these obscure issues.

Some ways in which gravitational lenses can act as crude telescopes are reviewed. Magnification limits associated with finite source size, corrugation of the gravitational potential and finite wavelength are specified. It should be possible to obtain rotation curves for a small sample of ultra-faint galaxies imaged as giant arcs. There is an appreciable chance of eventually observing a radio jet component cross a caustic. The successful observation of microlensing in 2237+031 suggests that the continuum emission from this source at blue wavelengths is at least partly non-thermal. The maximum possible magnification observable, granted existing constraints is argued to be from γ-ray pulsars crossing caustic sheets formed by galaxies or a hypothetical population of intergalactic massive objects.

Pixel lensing, the gravitational microlensing of unresolved stars, is potentially a powerful method for detecting and measuring microlensing events. Two groups are currently refining this method in observations toward M31. I show that the technique has wide application, from searching for intra-cluster Machos in the Virgo cluster, to improving the accuracy of follow-up observations of Galactic microlensing events, to measuring the star-formation history of the universe. I derive the equation for the pixel lensing event rate where Qmin is the minimum signal to noise for detection, τ is the optical depth, Nres is the number of telescope resolution elements in the field, is the photon detection rate from a fluctuation magnitude star, and ξ is a suppression factor.

A small volume of space, nearly on-axis behind a gravitational lens with respect to a given source, will receive a greatly increased radiation flux. In the idealized case of a point mass lens acting on a point source in complete isolation, the volume will approach zero only as the flux tends to infinity; in fact, the volume weighted rms flux is divergent. In realistic cases, finite source size and the effects of other gravitational deflections (i.e., non-zero shear) limit the maximum flux and considerably complicate the physics, but very large fluxes are still produced in small volumes. We consider the physics and statistics of these Extreme Gravitational Lensing Events (EGLE) and present an initial examination of their possible astrophysical effects for various known and putative populations of lensing objects and sources, with particular attention to the case in which finite source size is important but shear is not.

I briefly summarize all that is known about gravitational lensing by cosmic strings.

If a compact double (CD) is caused by lensing, the orientation and/or flux ratio of its components can substantially change in ≈ 1 year.

Recent results at high redshifts are briefly reviewed. An analysis of the evolution of the quasar luminosity function based on the recently published Hawkins & Véron (1995) variability-selected quasar sample is presented. The results support the conclusions of earlier work that suggested the simple two power law luminosity function, of invariant shape as a function of redshift, requires some modification in order to match the data. However, the contention of Hawkins & Véron that a single power law model provides an adequate fit to the variability-selected quasar sample and that there is no requirement for curvature in the form of the quasar luminosity function is not borne out by the analysis.

In the context of our studies on gravitational lensing effects among Highly Luminous Quasars (HLQs), we are presently compiling at STScI an archive of direct CCD frames for more than 1000 bright quasars observed with HST and ground-based telescopes. This archive will soon become publicly accessible through the Internet. On the basis of these observations, we are pursuing in a systematic way the analysis (subtraction of numerical PSFs and/or deconvolution) of the HLQ images in order to detect multiple QSO images and/or nearby foreground galaxies at very small angular separations. Residual images corresponding to several new possible multiply imaged HLQs are presented here. From the observed number and image configuration of gravitational lens candidates identified in this large sample of HLQs, it is possible to infer realistic values for parameters characterizing the galaxy deflectors, the number counts of quasars, etc. (cf. Claeskens et al. 1995ab), and also to set interesting constraints on the cosmological density of compact objects in the mass range 1010 ‒ 1012M⊙.

We are undertaking an extensive survey of a sample of radio-selected flat-spectrum sources, which are predominantly quasars. The sample will be used for a range of gravitational lensing studies as well as studies of the generic properties of radio quasars. As yet we have not found any instances of multiply-imaged quasars, however we have found evidence for dust which we believe is reddening the observed spectra of many of the quasars. Samples which are selected to be complete in the radio are the best way to account for the effects of dust.

We are searching for small lens systems (50-250 mas or 108–109M⊙) in a sample of ∼ 1800 flat spectrum radio sources. This is the first time a systematic search has been made “between” the VLA and VLBI resolutions. Finding any would indicate the existence of other than the “conventional” spiral/elliptical lenses (only ∼ 0.01% chance - Turner et al. (1984)). For example, faint galaxies are numerous (∼ 106 gal/deg2 - Lilly (1993), Glazebrook et al. 1995), compact (HST Medium Deep Survey (MDS) - Griffiths et al. 1994) and ideally placed for lensing (< z >∼ 0.6 - MDS, Smail et al. (1994); c.f. Turner et al. 1984). Early-type dwarf galaxies (dE,N and cE), if extant at intermediate-z as favored by MDS are also obvious lens candidates. If no lenses are found, a limit 400 times better than the current one (Surdej et al. 1993), ΩL < 0.001, will be placed on the cosmological density of compact objects (e.g. black holes) for the above mass range.

Selection criteria for 1800 MIT-VLA snapshots of PMN radio sources are described, and 6 new MG & PMN lens candidates are presented.

We are undertaking an imaging survey with the Australia Telescope Compact Array (ATCA) to find gravitational lens candidates in flat-spectrum Parkes Catalogue radio sources. Flat-spectrum radio sources typically possess a single high brightness temperature nucleus of milliarcsecond size. Such sources, if lensed, will show multiply imaged nuclei with separations that are large compared to their milliarcsecond sizes. Our flat-spectrum sample was selected using the criteria α2.7/5.0 > −0.5 (S(v) ∝ vα), S2.7 > 0.34Jy and δ ≤ −20°, and comprises a total of 461 sources.

We have embarked on a 15 GHz VLBA survey of 1000 flat spectrum sources. We present the results from a 24 hour pilot observing run in which 72 sources were mapped. The primary aims of this project are:

1. – to search for small separation (1-150 mas) gravitational lens systems

2. – to identify targets for current mm and anticipated Space VLBI programs

3. – a morphological classification of compact radio sources at relatively high frequency with sub-mas resolution.

About 20 years elapsed between my first and second papers on gravitational lenses (Press & Gunn 1973; Press, Rybicki & Hewitt 1992ab). Therefore, the conference organizers have asked me to prognosticate on the future of gravitational lenses. Their reasoning, if I understand it correctly, is that I will likely go to sleep for another 20 years immediately following this conference. So, if I simply look ahead to my very next paper on the subject, I will be in effect prognosticating over two decade's span.

We present a non-exhaustive bibliography on “Gravitational Lensing” (GL), totalizing nearly 1500 titles (see Pospieszalska-Surdej et al. 1993 for an earlier version of this bibliography). It also includes recent abstracts of papers dealing with gravitational lensing and submitted to astro-ph@babbage.sissa.it. The GL bibliography is accessible through a World-Wide-Web page at the URL (cf. next page) http://www.stsci.edu/ftp/stsci/library/grav_lens/grav_lens.html It is also possible to retrieve -via FTP- the latex file ‘grav_bib1.tex’ or the postscript file ‘grav_bib1.ps’ containing the most recent version of this bibliography. Please proceed as follows: ftp stsci.edu, Name: anonymous, Password: your E-mail address, cd stsci/library/grav_lens, get grav_bib1.tex or get grav_bib1.ps, and finally bye).

We are aware that the present compilation is not complete, not always uniform and that there are still errors, typos, etc. We would appreciate very much if you could communicate to us (via E-mail: surdej@stsci.edu) missing or new titles (published or accepted for publication), errors, changes and suggestions. It would also be nice for us to receive in the future your preprints or reprints related to ‘Gravitational Lensing’ studies. Please send your material to: A. Pospieszalska-Surdej, STScI, 3700 San Martin Drive, Baltimore MD 21218, USA.

It is well known that simple optical lenses simulating light deflection due to a foreground object may easily reproduce all types of image configurations observed among known multiply imaged QSOs and AGN (cf. simulations in Refsdal and Surdej 1992, 1994, Surdej et al. 1993). Such an optical lens experiment can also be used to simulate the formation of multiple giant luminous arcs and arclets seen near massive foreground galaxy clusters. The optical setup used to make this experiment is shown in Fig. 1 (a similar setup was presented during the Melbourne conference). A compact light source is located on the right side (not very clearly seen), then comes to the left a ‘spiral galaxy disk’ optical lens (manufactured by the authors) and a cardboard perforated by multiple pinholes. In the absence of light deflection (i.e. by simply removing the optical lens), direct (non distorted) images of the background galaxies are seen by an observer, like those projected on the distant white screen (see Fig. 2a). When the optical lens is inserted between the light source and the multiple pinholes screen (cf. Fig. 1), images of the background galaxies get distorted (arclets) and/or transformed into multiple images, including giant luminous arcs (see Figs. 2bc for examples). We showed during the conference that by covering the pinholes with various transparent colored filters and translating the multiple pinholes screen, it was possible to reproduce, in a very inexpensive way, distorted images of background galaxies like the very impressive strong lensing artifacts seen behind the rich cluster Abell 1689 (cf. the color videotape of simulations by Tony Tyson).

We present an extensive summary of the position, redshift, and flux data on the secure, multiply imaged, quasar and radio lenses. It includes neither the less-promising (in the opinion of the author) lens candidates nor the cluster lenses (which are too hard to reduce to short tables of numbers). A broader listing of suggested lenses and also of the cluster lenses is available at (see Pospieszalskia-Surdej et al. in this volume): http://www.stsci.edu/ftp/stsci/library/grav_lens/grav_lens.html