The clustering of galaxies relative to the underlying mass distribution declines with cosmic time for three reasons .
First , nonlinear peaks become less rare events as the density field evolves .
Second , the densest regions stop forming new galaxies because their gas becomes too hot to cool and collapse .
Third , after galaxies form , they are subject to the same gravitational forces as the dark matter , and thus they tend to trace the dark matter distibution more closely with time ; in this sense , they are gravitationally “ debiased. ” In order to illustrate these effects , we perform a large-scale hydrodynamic cosmological simulation of a \Lambda CDM model with \Omega _ { 0 } = 0.37 and examine the statistics of \delta _ { \ast } ( { \bf r } ,z ) , the density field of recently formed galaxies at position { \bf r } and redshift z .
We find that the bias of recently formed galaxies b _ { \ast } \equiv { \langle { \delta _ { \ast } ^ { 2 } } \rangle } ^ { 1 / 2 } / { \langle { \delta ^ { 2 } } % \rangle } ^ { 1 / 2 } , where \delta is the mass overdensity , evolves from b _ { \ast } \sim 4.5 at z = 3 to b _ { \ast } \sim 1 at z = 0 , on 8 h ^ { -1 } Mpc comoving scales .
The correlation coefficient r _ { \ast } \equiv { \langle { \delta \delta _ { \ast } } \rangle } / { \langle { \delta ^ { 2 } } % \rangle } ^ { 1 / 2 } { \langle { \delta _ { \ast } ^ { 2 } } \rangle } ^ { 1 / 2 } evolves from r _ { \ast } \sim 0.9 at z = 3 to r _ { \ast } \sim 0.25 at z = 0 .
That is , as gas in the universe heats up and prevents star formation , the star-forming galaxies become poorer tracers of the mass density field .
We show that the linear continuity equation is a good approximation for describing the gravitational debiasing , even on nonlinear scales .
The most interesting observational consequence of the simulations is that the linear regression of the galaxy formation density field on the galaxy density field , b _ { \ast g } r _ { \ast g } = { \langle { \delta _ { \ast } \delta _ { g } } \rangle } / { \langle { \delta% _ { g } ^ { 2 } } \rangle } , evolves from about 0.9 at z = 1 to 0.35 at z = 0 .
Measuring this evolution , which should be possible using the Sloan Digital Sky Survey , would place constraints on models for galaxy formation .
In addition , we evaluate the effects of the evolution of galaxy formation on estimates of \Omega from cluster mass-to-light ratios , finding that while \Omega ( z ) increases with z , the estimated \Omega _ { \mathrm { est } } ( z ) actually decreases .
This effect is due to the combination of galaxy bias and the relative fading of cluster galaxies with respect to field galaxies .
Finally , these effects provide a possible explanation for the Butcher-Oemler effect , the excess of blue galaxies in clusters at redshift z \sim 0.5 .