We show that a model consisting of individual, log-normal star formation histories for a volume-limited sample of $zapprox0$ galaxies reproduces the evolution of the total and quiescent stellar mass functions at $zlesssim2.5$ and stellar masses $M_*g
eq10^{10},{rm M_odot}$. This model has previously been shown to reproduce the star formation rate/stellar mass relation (${rm SFR}$--$M_*$) over the same interval, is fully consistent with the observed evolution of the cosmic ${rm SFR}$ density at $zleq8$, and entails no explicit quenching prescription. We interpret these results/features in the context of other models demonstrating a similar ability to reproduce the evolution of (1) the cosmic ${rm SFR}$ density, (2) the total/quiescent stellar mass functions, and (3) the ${rm SFR}$--$M_*$ relation, proposing that the key difference between modeling approaches is the extent to which they stress/address diversity in the (starforming) galaxy population. Finally, we suggest that observations revealing the timescale associated with dispersion in ${rm SFR}(M_*)$ will help establish which models are the most relevant to galaxy evolution.
The slope of the star formation rate/stellar mass relation (the SFR Main Sequence; ${rm SFR}-M_*$) is not quite unity: specific star formation rates $({rm SFR}/M_*)$ are weakly-but-significantly anti-correlated with $M_*$. Here we demonstrate that th
is trend may simply reflect the well-known increase in bulge mass-fractions -- portions of a galaxy not forming stars -- with $M_*$. Using a large set of bulge/disk decompositions and SFR estimates derived from the Sloan Digital Sky Survey, we show that re-normalizing SFR by disk stellar mass $({rm sSFR_{rm disk}equiv SFR}/M_{*,{rm disk}})$ reduces the $M_*$-dependence of SF efficiency by $sim0.25$ dex per dex, erasing it entirely in some subsamples. Quantitatively, we find $log {rm sSFR_{disk}}-log M_*$ to have a slope $beta_{rm disk}in[-0.20,0.00]pm0.02$ (depending on SFR estimator and Main Sequence definition) for star-forming galaxies with $M_*geq10^{10}M_{odot}$ and bulge mass-fractions $B/Tlesssim0.6$, generally consistent with a pure-disk control sample ($beta_{rm control}=-0.05pm0.04$). That $langle{rm SFR}/M_{*,{rm disk}}rangle$ is (largely) independent of host mass for star-forming disks has strong implications for aspects of galaxy evolution inferred from any ${rm SFR}-M_*$ relation, including: manifestations of mass quenching (bulge growth), factors shaping the star-forming stellar mass function (uniform $dlog M_*/dt$ for low-mass, disk-dominated galaxies), and diversity in star formation histories (dispersion in ${rm SFR}(M_*,t)$). Our results emphasize the need to treat galaxies as composite systems -- not integrated masses -- in observational and theoretical work.
