The globular cluster NGC 2419: a crucible for theories of gravity

Abstract

We present the analysis of a kinematic data set of stars in the globular cluster NGC 2419, taken with Keck/DEIMOS. Combined with a reanalysis of deep HST and Subaru imaging data, which provide an accurate luminosity profile of the cluster, we investigate the validity of a large set of dynamical models of the system, which are checked for stability via N-body simulations. We find that isotropic models in either Newtonian or Modified Newtonian Dynamics (MOND) are ruled out with extremely high confidence. However, a simple Michie model in Newtonian gravity with anisotropic velocity dispersion provides an excellent representation of the luminosity profile and kinematics. In contrast, with MOND we find that Michie models that reproduce the luminosity profile either over-predict the velocity dispersion on the outskirts of the cluster if the mass to light ratio is kept at astrophysically-motivated values, or else they under-predict the central velocity dispersion if the mass to light ratio is taken to be very small. We find that the best Michie model in MOND is a factor of 10000 less likely than the Newtonian model that best fits the system. A likelihood ratio of 350 is found when we investigate more general models by solving the Jeans equation with a Markov-Chain Monte Carlo scheme. We verified with N-body simulations that these results are not significantly different when the MOND external field effect is accounted for. If the assumptions that the cluster is in dynamical equilibrium, spherical, not on a peculiar orbit, and possesses a single dynamical tracer population of constant M/L are correct, we conclude that the present observations provide a very severe challenge for MOND. [abridged]