Gravitational Lens Image Separations Versus Redshift as a Cosmological Test

It has recently been noted by Park & Gott (1997) that there is a statistically significant negative correlation between gravitational lens image separations and source redshifts. This negative correlation is puzzling since is cannot be explained with any realistic models both in flat and open cosmologies, while ususal lens statistics, such as number density and distribution of image separation, are in agreement with such models. Using an improved sample of candidate gravitational lenses, we show that there is no such a negative correlation in present lens data and that the data are in agreement with general flat and open cosmologies. However, given the strong dependence of the observed result on the selection effects in lens samples, we suggest that results from a well defined lens search program are necessary to more accurately test the curvature of the universe using this relation. Subject headings: cosmology: miscellaneous — gravitational lensing — quasars: general