2013ApJ...764..184M

New kinematic data and modeling efforts in the past few years have substantially expanded and revised dynamical measurements of black hole masses (M_•) at the centers of nearby galaxies. Here we compile an updated sample of 72 black holes and their host galaxies, and present revised scaling relations between M_• and stellar velocity dispersion (σ), V-band luminosity (L), and bulge stellar mass (M_bulge), for different galaxy subsamples. Our best-fitting power-law relations for the full galaxy sample are log₁₀(M _•) = 8.32 + 5.64log₁₀(σ/200 km/s), log₁₀(M_•) = 9.23 + 1.11log₁₀(L/10¹¹ L_☉), and log₁₀(M_•) = 8.46 + 1.05log₁₀(M_bulge/10¹¹ M_☉). A log-quadratic fit to the M_•-σ relation with an additional term of β₂[log₁₀(σ/200 km/s)]² gives β₂= 1.68±1.82 and does not decrease the intrinsic scatter in M_•. Including 92 additional upper limits on M_• does not change the slope of the M_•-σ relation. When the early- and late-type galaxies are fit separately, we obtain similar slopes of 5.20 and 5.06 for the M_•-σ relation but significantly different intercepts–M_• in early-type galaxies are about two times higher than in late types at a given sigma. Within early-type galaxies, our fits to M_•(σ) give M_• that is about two times higher in galaxies with central core profiles than those with central power-law profiles. Our M_•-L and M_•-M_bulge relations for early-type galaxies are similar to those from earlier compilations, and core and power-law galaxies yield similar L- and M_bulge-based predictions for M_•. When the conventional quadrature method is used to determine the intrinsic scatter in M_•, our data set shows weak evidence for increased scatter at M_bulge< 10¹¹ M_☉ or L_V< 10^10.3 L_☉, while the scatter stays constant for 10¹¹ < M_bulge< 10^12.3 M_☉ and 10^10.3 < L_V< 10^11.5 L_☉. A Bayesian analysis indicates that a larger sample of M_• measurements would be needed to detect any statistically significant trend in the scatter with galaxy properties.