header graphic



[*] The velocity dispersion anisotropy of elliptical galaxies
van der Marel R.P.
in `Structure, dynamics, and chemical evolution of elliptical galaxies', Danziger I.J., Zeilinger W.W., Kjar K., eds., ESO, Garching, Germany, p. 79-86, 1993
© 1993. European Southern Observatory. All Rights Reserved.

It has been known for some time that elliptical galaxies are not flattened by rotation but by the anisotropy of their velocity dispersion tensor. However, it is not known what the sense of this anisotropy is, ``radial'' or ``tangential''. For flattened galaxies a constraint on the velocity dispersion anisotropy can be obtained by comparing the predictions of dynamical models to kinematical observations on both the major and the minor axis. Models in which the distribution function depends on only the two classical integrals of motion, f=f(E,L_z), generally predict too much motion on the major axis relative to the minor axis (van der Marel 1991). This implies that real elliptical galaxies must have more radial motion than these models, i.e., must have sigma_r > sigma_theta. A new and promising method to further constrain the velocity dispersion anisotropy of elliptical galaxies relies on measuring symmetric deviations of observed line profiles from Gaussians. This is now possible. Line profiles are discussed for a simple set of models with different velocity dispersion anisotropy. Fitting Gaussians to such line profiles can introduce systematic errors in the estimates of rotation velocities and velocity dispersions. A useful method to quantify deviations of observed line profiles from Gaussians relies on expanding the line profile in a Gauss-Hermite series, as was recently argued by van der Marel & Franx (1992) and independently by Gerhard (this volume; 1992).

Arrow Return to my bibliography.              Home Return to my home page.

Last modified December 8, 1998.
Roeland van der Marel, marel@stsci.edu.
Copyright Notice.