Gravitational lensing by rich clusters of galaxies
We present observations and analysis of gravitationally lensed systems designed to test the usefulness of gravitational lensing as a probe into the nature of faint galaxies (B(_j) ≤ 27). The analysis techniques used also provide unique observational constraints on the masses and morphologies of the lensing clusters. Using a catalogue of giant arc candidates constructed with a well determined selection function we initially examine the possible uses of such a catalogue to provide information on the redshift distribution of galaxies to B(_j) ~ 26. We conclude that the extreme sensitivity of the lensing signal to slight differences in the lensing clusters prevents the successful application of this technique. Nevertheless, we show that the tests could be successfully applied to very deep observations of a single rich cluster. We then illustrate the uniquely detailed view of high redshift galaxies which is provided by giant gravitational arcs by presenting near-infrared imaging of a complete sample of spectroscopically-confirmed giant arcs. Spectral modelling of the multi-band colours of these arcs confirms that the most distant arcs (z ~1) are remarkably blue at all wave lengths. This observation is incompatible with models for the arc's colours which contain recent star bursts superimposed upon even moderate fractions of underlying evolved stel-lar populations. We conclude that a constant star formation rate in marginally sub-L* systems is most consistent with the observations of the z ~ 1 arcs. Using very deep optical images of three moderate and distant redshift clusters selected on either X-ray luminosity or optical richness (C11455+-22; z = 0.26, C10016+16; z = 0.55 and C11603-+43; z = 0.89) we create robust samples of faint galaxies (B(_j) ≤ 27) suitable for lensing studies. These samples are used in two separate analytic methods to study both the redshift distribution of the faint field galaxy population and the mass in the lensing clusters. Relatively simple tests rule out redshift distributions for the faint field galaxies which depart significantly, at either low or high redshift, from the no evolution prediction. The mass analysis uses newly-developed techniques to reconstruct for the first time the two dimensional mass distribution in the lens from the gravitationally distorted images of galaxies seen through the clusters. We conclude that the morphology of the mass distribution is accurately traced by the cluster galaxies. Finally, we present the discovery of a very wide separation multiply-imaged system recently found in deep Hubble Space Telescope images of the rich cluster AC114 (z = 0.31). Spectroscopic observations of the multiply imaged system confirm its lensed nature and provide a probable redshift of z = 1.87 for the source. Preliminary modelling of the cluster mass using all the gravitationally lensed galaxies in the field support the conclusions reached above - the cluster galaxies appear to be fair tracers of the mass.