Using gravitational lenses to study galaxies and cosmology

Gravitational lenses are a powerful new tool for studying galaxies and cosmology. Time delays in multiply-imaged systems offer a new way to determine the Hubble constant {dollar}Hsb0.{dollar} I use detailed models of the four-image lens PG 1115+080 to compute {dollar}Hsb0{dollar} and to understand the systematic uncertainties in the lens models and their effects on {dollar}Hsb0.{dollar}; Multiply-imaged quasars and radio source probe the structure and evolution of galaxies. I model observed lenses to determine the shapes of lens galaxies and show that dark matter halos are generally aligned with the luminous galaxies (except in the presence of known tidal perturbations); this result suggests that the halos must have been modified by interactions with the galaxies. The models also determine lens galaxy masses, which can be combined with luminosities and colors to understand the physical properties of lens galaxies. In particular, I compute lens galaxy mass-to-light ratios and find the first direct evidence of luminosity evolution for a sample of elliptical galaxies in low-density environments.; Lensing can also constrain the structure of spiral galaxies. I construct realistic models of spiral galaxies lenses using both a thin disk and a rounder halo. The disk biases spiral galaxy lenses toward being viewed edge-on, but has little effect on the total number of lenses unless the disk is unreasonably massive. Such massive disks would produce a lensing geometry that has not been observed. A sample of spiral galaxy lenses would offer useful constraints on the masses of disks.; At least five lenses have a lens galaxy that is part of a group or small cluster, which is indicative of the fact that galaxies are found in a variety of environments. I construct semi-analytic models of substructure in dark matter halos in order to study galaxy environments. I show that the models are sensitive to the treatment of substructure, and develop a physically plausible approach combining dynamical friction and tidal stripping that gives good results. When applied to the lensing groups, the models rule out a simple isothermal profile but are consistent with the "universal" dark matter profile.