Choi, JHKatz, NMurali, CWeinberg, DDave, RHernquist, L2024-04-262024-04-262002-01-01https://doi.org/10.1086/339876https://hdl.handle.net/20.500.14394/3003<p>This is the pre-published version harvested from ArXiv. The published version is located at <a href="http://iopscience.iop.org/0004-637X/571/1/1/">http://iopscience.iop.org/0004-637X/571/1/1/</a></p>We use hydrodynamic simulations to examine how the baryonic components of galaxies are assembled, focusing on the relative importance of mergers and smooth accretion in the formation of ~L* systems. In our primary simulation, which models a (50 h-1 Mpc)3 comoving volume of a Λ-dominated cold dark matter universe, the space density of objects at our (64 particle) baryon mass resolution threshold Mc = 5.4 × 1010 M corresponds to that of observed galaxies with L ~ L*/4. Galaxies above this threshold gain most of their mass by accretion rather than by mergers. At the redshift of peak mass growth, z 2, accretion dominates over merging by about 4 : 1. The mean accretion rate per galaxy declines from ~40 M yr-1 at z = 2 to ~10 M yr-1 at z = 0, while the merging rate peaks later (z 1) and declines more slowly, so by z = 0 the ratio is about 2 : 1. We cannot distinguish truly smooth accretion from merging with objects below our mass resolution threshold, but extrapolating our measured mass spectrum of merging objects, dP/dM M-α with α ~ 1, implies that subresolution mergers would add relatively little mass. The global star formation history in these simulations tracks the mass accretion rate rather than the merger rate. At low redshift, destruction of galaxies by mergers is approximately balanced by the growth of new systems, so the comoving space density of resolved galaxies stays nearly constant despite significant mass evolution at the galaxy-by-galaxy level. The predicted merger rate at z 1 agrees with recent estimates from close pairs in the Canada-France Redshift Survey and Canadian Network for Observational Cosmology Redshift Survey.Astrophysics and Astronomyarticle