A radio-based search for strong gravitational lensing , with image separations in the range 160–300 milliarcsec ( mas ) , has yielded a null result for a sample of 1665 sources ( Augusto , Wilkinson & Browne 1998 ) whose mean redshift is estimated to be \sim 1.3 .
The lensing rate for this previously-unexplored separation range , < 1 : 555 at the 95 % confidence level , is less than on arcsecond-scales—as expected from models of lensing galaxy populations .
Lensing on 160–300 mas scales is expected to arise predominantly from spiral galaxies at a rate dependent on the disk-halo mass ratio and the evolving number density of the population with redshift .
While the present sample is too small for there to be a high probability of finding spiral galaxy lenses , our work is a pilot survey for a much larger search based on the full CLASS database which would provide useful information on galactic structure at z \sim 0.5 .
We examine other possible lens populations relevant to our present search , in particular dwarf galaxies and supermassive black holes in galactic nuclei , and conclude that none of them are likely to be detected .
Our null result enables us formally to rule out a cosmologically significant population of uniformly-distributed compact objects : \Omega _ { CO } < 0.1 ( 95 % confidence ) in the mass range 10 ^ { 9.5 } – 10 ^ { 10.9 } M _ { \odot } .