THE REDSHIFT DISTRIBUTION OF INTERVENING WEAK Mg II QUASAR ABSORBERS AND A CURIOUS DEPENDENCE ON QUASAR LUMINOSITY

We have identified 469 Mg II lambda lambda 2796, 2803 doublet systems having W-r >= 0.02 angstrom in 252 Keck/High Resolution Echelle Spectrometer and UVES/Very Large Telescope quasar spectra over the redshift range 0.1 < z < 2.6. Using the largest sample yet of 188 weak Mg II systems (0.02 angstrom <= W-r < 0.3 angstrom), we calculate their absorber redshift path density, dN/dz. We find clear evidence of evolution, with dN/dz peaking at z similar to 1.2, and that the product of the absorber number density and cross section decreases linearly with increasing redshift; weak Mg II absorbers seem to vanish above z similar or equal to 2.7. If the absorbers are ionized by the UV background, we estimate number densities of 10(6)-10(9) Mpc(-3) for spherical geometries and 10(2)-10(5) Mpc(-3) for more sheetlike geometries. We also find that dN/dz toward intrinsically faint versus bright quasars differs significantly for weak and strong (W-r >= 1.0 angstrom) absorbers. For weak absorption, dN/dz toward bright quasars is similar to 25% higher than toward faint quasars (10 sigma at low redshift, 0.4 <= z <= 1.4, and 4 sigma at high redshift, 1.4 < z <= 2.34). For strong absorption the trend reverses, with dN/dz toward faint quasars being similar to 20% higher than toward bright quasars (also 10 sigma at low redshift and 4 sigma at high redshift). We explore scenarios in which beam size is proportional to quasar luminosity and varies with absorber and quasar redshifts. These do not explain dN/dz's dependence on quasar luminosity.