Annu. Rev. Astron. Astrophys. 1991. 29: 239-274 Copyright © 1991 by Annual Reviews. All rights reserved

3.5.2 DARK MATTER Strong evidence for the presence of dark halos around spiral galaxies is provided by their HI rotation curves, but the evidence for dark matter around ellipticals is still rather poor. This topic was reviewed recently by Kent (186).

The velocity fields of emission-line gas disks have been used to determine M / L-profiles in the central regions of elliptical galaxies (e.g., 31, 55). Combination with stellar velocity dispersion profiles constrains the anisotropy in the velocity distribution, but the intrinsic shape of the galaxy has a significant effect on the modeling. This was demonstrated clearly for the E3 galaxy NGC 5077, where spherical models with the gas on circular orbits give a central decline in M / L, but triaxial models with constant M / L are able to fit the data equally well (31). There is evidence, however, that the central gas rotation curves are suspect: the emission-line gas in many bulges often seems to rotate much slower than expected from the stellar velocity dispersions (117, 196, 311). For NGC 4594, the gas velocities are smaller than the mean stellar streaming velocities (196). This may indicate that the gas is flowing in towards the center.

Few ellipticals contain substantial amounts of HI out to large radii, e.g., NGC 1052 (352), NGC 2974 (187), NGC 4278 (208, 283), NGC 5128 (353), NGC 5666 (204), and IC 2006 (311). In NGC 1052 no radial change in M / L is found, despite that fact that HI is observed as far out as 10r_e. A strong outwards increase of M / L was found for IC 2006, from 5h₅₀ in the inner parts to 16h₅₀ at 6.5 r_e (311). However, interpretation of these observations is complex, as ellipticals as a class are triaxial, so that even if the gas has settled, it generally is not on circular orbits (244, 342). This was generally ignored in the past, but it influences the inferred M / L behavior (31, 208).

The velocities of individual planetary nebulae and globular clusters can be measured to about 4r_e, but they do not give strong constraints on M / L, because of the relatively small number of measurements (119, 158, 159, 256). Hot X-ray gas may well provide one of the best constraints on M / L at large radii. However, few temperature profiles are available, and as a result the mass determinations are still uncertain [e.g., 120, 300, 337, and the review by Fabbiano (111)]. The new data from ROSAT may help to improve this situation.

A completely different constraint on the dark halos comes from cluster dynamics. Merritt has argued that no more than 15% of the mass in clusters can be attached to individual galaxies (241). This implies that the total M / L for a galaxy is 0.15 times the M / L of the total cluster (typically 200h₅₀), resulting in M / L 30h₅₀. This would imply that the halos contain less than three times the mass of the luminous parts. This rather low factor may result from stripping during the formation of the cluster; it may very well be that the progenitors of ellipticals had heavier halos. Furthermore, Merritt's result shows that in order to answer the question of whether elliptical galaxies have dark halos one has to demonstrate that any dark matter found is in fact related to the galaxy, and has a similar velocity dispersion.