The discovery of X ray halos around bright cluster galaxies led to its use as a tracer of the gravitational potential typically within 5-10 effective radii. Satellites such as ASCA and ROSAT allowed global mass measurements which confirmed the dominant role of the dark matter at these scales, and suggested a relation between the temperature of the hot gas and the central velocity dispersion of the stellar component (see [6, 7]). Such a relation led Loewenstein & White  to globally constrain the M / L of a sample of about 30 galaxies in clusters, implying a relatively low fraction of dark matter within 1 Re of 20%, going up to 40-85% within 6 Re. These measurements are however conditioned by the assumption of hydrostatic equilibrium, and could be severely affected by (gravitational) perturbations in the environment of the galaxies. In order to minimise the effect of the surrounding intra-cluster medium, as well as the additional contamination from point-like sources, O'Sullivan & Ponman  studied the isolated elliptical galaxy NGC 4555 with the ACIS/Chandra instrument (Fig. 1), and found a large M / LB of about 57 at 50 kpc implying a very significant fraction of dark matter.
Detailed mass profiles were subsequently derived by Fukuzawa et al.  for a sample of 53 early-type galaxies via data retrieved from the Chandra archive. They first found relatively good agreement with mass profiles within 1 Re derived either from stellar population studies or dynamical modelling. Fukuzawa et al.  then emphasised the apparent dichotomy in the temperature profiles between X-ray luminous and dim galaxies, with the former exhibiting a temperature increasing with radius, while the latter show declining or flat profiles. The derived mass-to-light ratios seemed to be constant within the central 1 Re, but clearly rise up outwards (but see e.g., Pellegrini et al. ). Similar results were obtained by Humphrey et al.  for 7 nearby ellipticals, with the central M / LK being consistent within 1 Re with a Kroupa initial mass function (IMF; Fig. 2).