Context. Some galaxy clusters exhibit shallow or even cored dark matter density profiles in their central regions rather than the predicted steep or cuspy profiles, conflicting with the standard understanding of dark matter. NGC 3311 is the central cD galaxy of the Hydra I cluster (Abell 1060). Aims. We use globular clusters around NGC 3311, combined with kinematical data of the galaxy itself, to investigate the dark matter distribution in the central region of Hydra I. Methods. Radial velocities of 118 bright globular clusters, based on VLT/VIMOS mask spectroscopy, are used to calculate velocity dispersions which are well defined out to 100 kpc. NGC 3311 is the most distant galaxy for which this kind of study has been performed. We also determine the velocity dispersions of the stellar component from long-slit spectroscopy of NGC 3311 acquired with VLT/FORS1 out to 20 kpc. We present a new photometric model for NGC 3311, based on deep VLT/FORS1 images in the V-band. We search for a dark halo that, in the context of a spherical Jeans model, can reproduce the kinematical data. We also compare the radial velocity distributions of globular clusters and planetary nebulae. Results. The projected stellar velocity dispersion rises from a central low value of about 185 km s-1 to 350 km s-1 at a radius of 20 kpc. The globular cluster dispersion rises as well from 500 km s-1 at 10 kpc to about 800 km s-1 at 100 kpc, comparable to the velocity dispersion of the cluster galaxies. A dark matter halo with a core (Burkert halo) closely reproduces the velocity dispersions of stars and globular clusters simultaneously under isotropy. The central stellar velocity dispersions predicted by cosmological NFW halos do not agree well with those observed, while the globular clusters allow a wide range of halo parameters. A suspected radial anisotropy of the stellar population found in merger simulations aggravates the disagreement with observations. A slight tangential anisotropy would enable the data to be more accurately reproduced. However, we find discrepancies with previous kinematical data that we cannot resolve, which may indicate a more complicated velocity pattern. Conclusions. Although one cannot conclusively demonstrate that the dark matter halo of NGC 3311 has a core rather than a cusp, a core seems to be most consistent with the present data. A more complete coverage of the velocity field and a more thorough analysis of the anisotropy is required to reach firm conclusions.
Available at: http://works.bepress.com/aaron_romanowsky/39/