We construct merger trees for galaxies identified in a cosmological hydrodynamical simulation and use them to characterize predicted merger rates as a function of redshift , galaxy mass , and merger mass ratio .
At z = 0.3 , we find a mean rate of 0.054 mergers per galaxy per Gyr above a 1:2 mass ratio threshold for massive galaxies ( baryonic mass above 6.4 \times 10 ^ { 10 } M _ { \odot } ) , but only 0.018 Gyr ^ { -1 } for lower mass galaxies .
The mass ratio distribution is \propto R _ { merg } ^ { -1.2 } for the massive galaxy sample , so high mass mergers dominate the total merger growth rate .
The predicted rates increase rapidly with increasing redshift , and they agree reasonably well with observational estimates .
A substantial fraction of galaxies do not experience any resolved mergers during the course of the simulation , and even for the high mass sample only 50 % of galaxies experience a greater than 1 : 4 merger since z = 1 .
Typical galaxies thus have fairly quiescent merger histories .

We assign bulge-to-disk ratios to simulated galaxies by assuming that mergers above a mass ratio threshold R _ { major } convert stellar disks into spheroids .
With R _ { major } of 1 : 4 , we obtain a fairly good match to the observed dependence of early-type fraction on galaxy mass .
However , the predicted fraction of truly bulge-dominated systems ( f _ { bulge } > 0.8 ) is small , and producing a substantial population of bulge-dominated galaxies may require a mechanism that shuts off gas accretion at late times and/or additional processes ( besides major mergers ) for producing bulges .