We predict how the observed variations in galaxy populations with environment affect the number and properties of gravitational lenses in different environments .
Two trends dominate : lensing strongly favors early-type galaxies , which tend to lie in dense environments , but dense environments tend to have a larger ratio of dwarf to giant galaxies than the field .
The two effects nearly cancel , and the distribution of environments for lens and non-lens galaxies are not substantially different ( lens galaxies are slightly less likely than non-lens galaxies to lie in groups and clusters ) .
We predict that \sim 20 % of lens galaxies are in bound groups ( defined as systems with a line-of-sight velocity dispersion \sigma in the range 200 < \sigma < 500 km s ^ { -1 } ) , and another \sim 3 % are in rich clusters ( \sigma > 500 km s ^ { -1 } ) .
Therefore at least \sim 25 % of lenses are likely to have environments that significantly perturb the lensing potential .
If such perturbations do not significantly increase the image separation , we predict that lenses in groups have a mean image separation that is \sim 0 \farcs 2 smaller than that for lenses in the field and estimate that 20–40 lenses in groups are required to test this prediction with significance .
The tail of the distribution of image separations is already illuminating .
Although lensing by galactic potential wells should rarely produce lenses with image separations \theta \gtrsim 6 \arcsec , two such lenses are seen among 49 known lenses , suggesting that environmental perturbations of the lensing potential can be significant .
Further comparison of theory and data will offer a direct probe of the dark halos of galaxies and groups and reveal the extent to which they affect lensing estimates of cosmological parameters .