A Keck/DEIMOS spectroscopic survey of faint Galactic satellites: searching for the least massive dwarf galaxies

We present the results of a spectroscopic survey of the recently discovered faint Milky Way satellites Bootes, Ursa Major I, Ursa Major II and Willman 1 (Wil1). Using the DEep Imaging Multi-Object Spectrograph mounted on the Keck II telescope, we have obtained samples that contain from similar to 15 to similar to 85 probable members of these satellites for which we derive radial velocities precise to a few km s(-1) down to i similar to 21-22. About half of these stars are observed with a high enough signal-to-noise ratio to estimate their metallicity to within +/- 0.2 dex. The characteristics of all the observed stars are made available, along with those of the Canes Venatici I dwarf galaxy that have been analysed in a companion paper. From this data set, we show that Ursa Major II is the only object that does not show a clear radial velocity peak. However, the measured systemic radial velocity (v(r) = 115 +/- 5 km s(-1)) is in good agreement with simulations in which this object is the progenitor of the recently discovered Orphan Stream. The three other satellites show velocity dispersions that make them highly dark matter dominated systems (under the usual assumptions of symmetry and virial equilibrium). In particular, we show that despite its small size and faintness, the Wil1 object is not a globular cluster given its metallicity scatter over -2.0 less than or similar to [Fe/H] less than or similar to -1.0 and is therefore almost certainly a dwarf galaxy or dwarf galaxy remnant. We measure a radial velocity dispersion of only 4.3(-1.3)(+2.3) km s(-1) around a systemic velocity of -12.3 +/- 2.3 km s(-1) which implies a mass-to-light ratio of similar to 700 and a total mass of similar to 5 x 10(5) M-circle dot for this satellite, making it the least massive satellite galaxy known to date. Such a low mass could mean that the 10(7) M-circle dot limit that had until now never been crossed for Milky Way and Andromeda satellite galaxies may only be an observational limit and that fainter, less massive systems exist within the Local Group. However, more modelling and an extended search for potential extratidal stars are required to rule out the possibility that these systems have not been significantly heated by tidal interaction.