The Globular Cluster System of NGC 1399. II. Kinematics of a Large Sample of Globular Clusters

We study the kinematics and dynamics of the globular cluster system of NGC 1399, the brightest elliptical galaxy near the center of the Fornax cluster of galaxies. The observational data consists of medium-resolution spectra, obtained at the Very Large Telescope with FORS2 and the Mask Exchange Unit (MXU). Our sample comprises 468 radial velocities in the magnitude range 20 < mR < 23. This is the largest sample of globular cluster velocities around any galaxy obtained so far. Typical velocity uncertainties are 50 km s-1, significantly improving on earlier samples. The radial range is 2′ < r < 9′, corresponding to 11 kpc to 50 kpc of galactocentric distance. The shape of the velocity distribution of the sample is compatible with being a Gaussian distribution. However, under moderate error selection, a slight asymmetry is visible between high and low radial velocities. We find bright clusters with radial velocities below 800 km s-1, while they are not found at the corresponding high-velocity side above 2000 km s-1. There is the possibility that unbound clusters and/or objects in the foreground contaminate the NGC 1399 cluster sample. Under strong error selection, practically no objects are found with velocities lower than 800 km s-1 or higher than 2000 km s-1. Since the extreme velocities influence the velocity dispersion considerably, uncertainty regarding the exact value of the dispersion remains. With the above velocity limits, we derive a projected velocity dispersion for the total sample of 274 ± 9 km s-1 which within the uncertainties remains constant over the entire radial range. Without any velocity restriction, it increases to 325 km s-1. Guided by the bimodal color distribution of clusters, we distinguish between red clusters (C-R > 1.6) and blue clusters (C-R < 1.6), and find velocity dispersions for these groups of 255 ± 13 and 291 ± 14 km s-1, respectively, again radially constant. Any possible rotation of either of these cluster populations is below the detection limit, with the exception of a weak signature of rotation for the blue clusters more distant than 6′. Spherical models point to a circular velocity of 419 ± 30 km s-1, assuming isotropy for the red clusters. This value is constant out to 40 kpc. The inferred dark halo potential can be well represented by a logarithmic potential. A halo of the NFW type also provides a good fit to the observations. The orbital structure of the clusters can only be weakly constrained. It is consistent with isotropy for the red clusters and a slight tangential bias for the blue clusters. Some mass profiles derived from X-ray analyses do not agree with a constant circular velocity within our radial range, irrespective of its exact value. Interpreting the extreme low radial velocities as space velocities of bound clusters near their pericentric distances would require an extension of the cluster system of at least 200 kpc. Implications for formation scenarios of the cluster system are briefly commented on.