The difficulties we have described are sufficiently common that almost every lensed quasar is subject to at least one of them. In general (though there are important exceptions) the double systems suffer from a paucity of constraints. Many quadruples suffer either from the mass sheet degeneracy (due to a nearby cluster) or from multiple lensing galaxies. Only a few systems (mostly radio loud quasars) have multiple sources permitting unambiguous determination of a radial density exponent. Are there any lenses which are not obviously subject to one or another of these difficulties?
A decade ago we imagined finding a "perfect" [Press (1996)] or "golden" lens [Williams & Schechter (1997)] that would permit determination of H0 with small uncertainty. A lens that was not subject to any of the above difficulties might reasonably qualify for certification as a golden lens.
The system JVAS0218+357 [Patnaik et al. (1992)] is such a system. It is a radio double that, at VLBI resolution, has a core and a blob [Biggs et al. (2003)] of the sort that radio astronomers fancifully call jets. The double source breaks the concentration degeneracy, and implies an isothermal potential. [Wucknitz, Biggs, & Browne (2004)] have modeled the radio data, finding the radial exponent and the center of symmetry of the lensing potential (see also WUCKNITZ' contribution to the present proceedings). The time delay has been measured with better than 5% accuracy [Biggs et al. (1999)].
But even this system is not 24 carat gold. The pairs of images are separated by only 0."33, implying a low luminosity lens, almost certainly not an elliptical. [York et al. (2004)] have co-added data from a large number of HST orbits (see also JACKSON AND YORK'S contribution to the present proceedings). After careful PSF subtraction of the two quasar images, their data clearly shows an M101-like spiral. The residuals from PSF subtraction crowd the nucleus of the galaxy, making it difficult to measure its centroid accurately. The position of the center is crucial because of the strong dependence of the differential delay on the image distances. The agreement between the York et al. optical position and the Wucknitz center of symmetry is excellent (figure 7). The radio position gives H0 = 78 ± 6 km/s/Mpc (2 uncertainty).
Figure 7. The shading of the pixels indicates relative likelihood for the optical position of the galaxy lensing JVAS0218+357 [York et al. (2004)]. The solid line is drawn from the B image toward the A image, which lies beyond the plot boundary. The diagonal dashes indicate loci of constant Hubble constant, starting with H0 = 90 km/s/Mpc on the left and decreasing in steps of 10 km/s/Mpc. The elliptical contour gives the 2 confidence region for the center of the lensing potential, as derived from radio observations of the lensed radio source by [Wucknitz, Biggs, & Browne (2004)].
The work on JVAS0218+357 is remarkable for the range of astronomical techniques and resources that have been brought to bear on the system. It was discovered [Patnaik et al. (1992)] and monitored [Biggs et al. (1999)] at radio wavelengths using the VLA and Merlin. The global VLBI data [Biggs et al. (2003)] are crucial to the modeling. The redshift for the lens [Browne, Patnaik, Walsh, & Wilkinson (1993)] required large ground based optical telescopes. Measuring the position of the lensing galaxy required HST [York et al. (2004)], as is the case for most lenses.