2008ApJ...678..187V

We present several different statistical methods to determine the transverse velocity vector of M31. The underlying assumptions are that the M31 satellites on average follow the motion of M31 through space and that the galaxies in the outer parts of the Local Group on average follow the motion of the Local Group barycenter through space. We apply the methods to the line-of-sight velocities of 17 M31 satellites, to the proper motions of the two satellites M33 and IC 10, and to the line-of-sight velocities of five galaxies near the Local Group turnaround radius, respectively. This yields four independent but mutually consistent determinations of the heliocentric M31 transverse velocities in the west and north directions, with weighted averages <v_W≥-78±41 km/s and <v_N≥-38±34 km/s. The uncertainties correspond to proper motions of ∼10 µas/yr, which will be difficult to verify observationally within the next decade. The galactocentric tangential velocity of M31 is 42 km/s, with 1 σ confidence interval V_tan≤56 km/s. The implied M31-Milky Way orbit is bound if the total Local Group mass M exceeds 1.72^+0.26_–0.25x10¹² M_☉. If the orbit is indeed bound, then the timing argument combined with the known age of the universe implies that M=5.58^+0.85_–0.72x10¹² M_☉. This is on the high end of the allowed mass range suggested by cosmologically motivated models for the individual structure and dynamics of M31 and the Milky Way, respectively. It is therefore possible that the timing mass is an overestimate of the true mass, especially if one takes into account recent results from the Millenium Simulation that show that there is also a theoretical uncertainty of 41% (Gaussian dispersion) in timing mass estimates. The M31 transverse velocity implies that M33 is in a tightly bound orbit around M31. This may have led to some tidal deformation of M33. It will be worthwhile to search for observational evidence of this.