Key points are not available for this paper at this time.
The evolution of galaxies through collapse, merging, and accretion between now and |z = 1| is constrained by its effects on gravitational lens statistics and image separations. If we consider the process of galaxy formation, most galaxies must have collapsed and dynamically resemble the ones seen today by a redshift of 0. 8 (90 per cent confidence limit). If E and S0 galaxies are formed by mergers of spiral galaxies, most of the E/S0s must be formed by a redshift of 0. 4. If large galaxies are formed by mergers of smaller galaxies, then the local rate of mergers must be less than 2. 8 mergers per galaxy in a Hubble time |H₀^-1|. New galaxy populations with velocity dispersion scales of 100, 150 and |200\, km\, s^-1| must have comoving densities less than |0. 15\, h³, 0. 032\, h³| and |0. 017\, h³ Mpc^-3|. These limits do not constrain the faint blue galaxy population because their mass-scale is probably much less than |100\, km\, s^- 1|. In the | = 1| Einstein–de Sitter cosmology, the best model is always the no-evolution model. These limits are consistent with other determinations, but they are derived by a completely independent technique. Cosmological models with a large cosmological constant are consistent with lens statistics given a sufficiently high rate of evolution; for an | = 0. 2, = 0. 8| model, this means that E/S0 galaxies cannot collapse until |z 0. 75|. This contradicts models of galaxy formation, and galaxy number counts, so galaxy evolution does not provide a means of evading the limits on || from the statistics of gravitational lenses. In general, only a dramatic evolution of the E/S0 galaxy population at redshifts less than |z = 1| produces a detectable effect on the existing surveys for gravitational lenses.
Mao et al. (Sun,) studied this question.