Astron. Astrophys. 329, 827-839 (1998)

7. Discussion, prospective, and conclusion

The complete multi-wavelength analysis of the Cloverleaf reveals that this is probably a complex lens which includes a lensing galaxy and an additional distant lensing cluster of galaxies. The reality of the cluster toward the Cloverleaf has still to be confirmed independently. Even at the level of a 4 detection, clustering and projection effects can not be readily discarded to explain the observed galaxy enhancement therefore measuring the redshift of these galaxies is of high priority to position them in redshift space.

Yet, most of the galaxies around the Cloverleaf are found in the same magnitude and size ranges, as expected if they indeed belonged to a cluster. If the cluster is at a very large distance, the shift of its galaxy luminosity function up to higher apparent magnitude would explain why the number-density contrast of the cluster with respect to faint field galaxies is lowered down to only a 4 level.

This interpetation implies that the lensing galaxy may not be very massive and consequently may not be very luminous, helping to explain the mystery of the lensing galaxy not having been detected so far. The drawback is that, despite the constraint that the shapes of the CO spots provide on the orientation of the mass density distribution, it mandates sharing the mass between the lensing-galaxy and the lensing-cluster which increases the number of possible lens configurations. Hence, it considerably reduces the chances to infer a secure measure of the Hubble constant from the time delay measurements of lightcurves between the four spots (a thorough report of the variability of the quasar is given in Ostensen et al. 1997). A measure of time-delay would bring additional constraint on the mass distribution of this system. Thus, we need to confirm, probably from ultra-deep visible images and spectra and additional near infrared photometry, that the cluster is present and that it is at a large redshift.

Ultra-deep visible and K band images might also reveal the position and the shape of the light distribution of the lensing galaxy. Such information would be useful to improve the mapping of the CO source, as already emphasised by Yun et al. (1997) and Alloin et al. (1997). With the present-day data, the CO source is found to be a disk- or ring-like structure with typical radius of 100 pc, under the lens model of a galaxy and a cluster at z=1.7, leading to a central 10⁹ M object, typical of a black-hole. It is amazing to see that a disk with such a small intrinsic size can be spatially "resolved" even at an angular distance as large as 1.6 Gpc. However, we are aware that the CO source could have a more complex geometry and as long as the exact redshift and mass distribution of the lenses will not be known in a direct way, uncertainties on the CO source size and structure will persist.

Although this remains to be confirmed independently, the discovery of a distant cluster of galaxies on the line of sight to the Cloverleaf is remarkable because it reinforces the suspicion that many bright high redshift quasars are magnified by cluster-like systems at large distances. This was already reported from analyses in the fields of the doubly imaged quasar Q2345+007 (Bonnet et al 1993; Mellier et al 1994; van Waerbeke et al 1997), where a cluster candidate is expected to be at z 0.75 (Pelló et al 1996) and of MG2016 where the X-ray emission of the intra-cluster gas has been observed and for which the Iron line (from X-ray spectroscopy) gives a redshift z 1 (Hattori et al 1997).

These cases of strong lensing may substantially change the intrinsic bright-end luminosity function of quasars. In fact, there is now convincing evidence that magnification biases play an important role and this should draw some important cosmological issues. Early observational evidence was emphazised by the galaxy-quasar associations detected by Fugmann (1990). They have been re-investigated and confirmed by Bartelmann & Schneider (1994) and Beñitez & Martnez-Gonzalez (1997). According to Bartelmann & Schneider (1992) the galaxy-quasar association on arcminute scale cannot be explained by lensing effects from single galaxies only, but must involve lensing effects by large-scale structure or rich clusters of galaxies. Since all these sources, including the Cloverleaf, are very bright, they correspond to preferentially selected fields which probe the bright-end of the magnified areas where we expect that the lensing agent responsible for magnification must be strong. If so, many bright quasars are magnified by massive gravitational systems. It would therefore be important to collect ultra-deep visible and near infrared high-resolution imaging on a large sample of overbright quasars in order to compute what fraction is surrounded by high-redshift foreground clusters of galaxies.

Though these fields correspond to biased lines of sight, they are typical fields showing the strong-end of cosmic shear events expected on arcminute scales from the predictions of Jain & Seljak (1997). By using the non-linear evolution of the power spectrum, these authors have shown that the rms cosmic shear on such scales is more than twice the values predicted from the weakly non-linear regime. Arguing that the cosmic shear should therefore be observable even with present-day ground based telescopes, Schneider et al (1997) have re-analysed the Fort et al (1996) data and have shown that the cosmic shear may have already been detected in at least one quasar field (PKS1508). Very deep observations of the Cloverleaf in the near IR, to confirm that a shear pattern with significant shear amplitude of order of 5% is present around the four spots, would provide independent data in favor of the this interpretation.

The observation of cosmic shear or the detection of distant clusters puts strong constraints on the density parameter and on the initial power spectrum of density fluctuations. If most of the bright quasars are magnified by high-redshift clusters, then the density of distant clusters of galaxies must be large (this may conflict with the standard CDM model for example). However, the impact on cosmological scenarios is not clear since we do not understand yet the selection biases.

Though our multiwavelength analysis, using the best visible and CO images, provides a new interpretation of the Cloverleaf and a good understanding of the lens configuration, the discussion raises new questions about the CO source and the lenses. In order to address these questions thouroughly, additional photometric and spectroscopic data are required. Since the members of the cluster are extremely faint their spectroscopy and the measurement of the cluster velocity dispersion will be technically difficult. In addition to ultra-deep visible and near infrared imaging to probe the light distribution of the lenses (individual galaxy and cluster), the mass model representing the cluster could be considerably improved by using the shear pattern generated by the lensed background sources and from the redshift measurements of the `brightest' cluster members. Last but not least, the photometric monitoring of the four spots (Ostensen et al 1997) will provide the time delay of the light curves which is also an important and independent constraint for the mass modelling.

© European Southern Observatory (ESO) 1998

Online publication: December 16, 1997
helpdesk.link@springer.de