Image separation versus redshift of lensed QSOs: Implications for galaxy mass profiles

Recently, Park and Gott reported an interesting observation: image separation of lensed QSOs declines with QSO redshift more precipitously than expected in any realistic world model, if the lenses are taken to be either singular isothermal spheres or point masses. In this letter I propose that the observed trend arises naturally if the lensing galaxies have logarithmic surface mass density profiles that gradually change with radius. If the observed lack of central (odd) images is also taken into account, the data favour a universal dark matter density profile over an isothermal sphere with a core. Since the trend of image separation versus source redshift is mostly a reflection of galaxy properties, it cannot be straightforwardly used as a test of cosmological models. Furthermore, the current upper limits on the cosmological constant may have to be revised.