Key points are not available for this paper at this time.
The self-consistent dynamical models of spherically symmetric galaxies presently constructed by means of Schwarzschild's (1979) linear programming method have mass-to-light ratios that are independent of radius, and surface brightness profiles obeying de Vaucouleurs (1948) or King (1972) laws. Although some of these models have dispersion profiles which one would not expect to see in a real galaxy, it appears impossible to determine, in the absence of very restrictive ad hoc assumptions about the dispersion profile, the mass-to-light ratio on the basis of a dispersion measurement at a single radius. Accurate determination of the mass-to-light ratio value requires a measurement of the rms dispersion in a centered aperture that contains a large fraction of the total galaxy luminosity, or a series of dispersion measurements at well separated radii spread out across the galaxy.
Richstone et al. (1984) studied this question.