The 3D shapes of elliptical galaxies are still not known; knowing them would tell us much about the processes operating during the epoch of galaxy formation. The obstacle to our undertanding is the bewildering array of orbits on which the stars in the galaxy can move. The picture at right shows typical volumes filled by orbits in integrable triaxial potentials (Statler 1987). These orbits can be populated and added together in many different ways to produce an elliptical system. Otherwise identical systems can therefore present entirely different dynamical faces, depending on which orbits are populated. Disentangling the effects of orbit populations from the influence of the shape of the gravitational potential -- especially when we can observe only the line-of-sight component of the motion -- is a challenge.

Theoretical models for the average flow of the stellar "fluid" through the galaxy suggest that the amount of triaxiality (or conversely axisymmetry) of ellipticals can be distinguished by the degree of symmetry of their projected velocity fields. The mean velocity field in effect shows the locations of critical points associated with orbit family boundaries, which reveal the shape of the potential. The figure at left shows the internal mean flow patterns of the "stellar fluid" for three model galaxies differing only in their triaxialities (Statler 1994a).

We use Bayesian statistical methods to estimate the true shapes of elliptical galaxies from observations of their projected light distributions and stellar motions (Statler 1994c). The figure below shows the intrinsic shape distribution inferred from the Davies & Birkinshaw (1988) sample of 13 galaxies (Bak & Statler 2000). Panel (a) shows that there is a tendency for ellipticals to be either nearly oblate (T=0) or nearly prolate (T=1), but that there is still a population of triaxial objects. Panel (b) shows the result obtained by omitting the stellar kinematic data and working with photometry alone; this shows that kinematics are essential. Further work will better define the true shape distribution, and shed light on the possible existence of two or more sub-families of ellipticals, formed by different processes.

Intrinsic Shape Distribution

What's that? You don't think this could possibly work?

The test at right shows that we are able to recover the shape distribution of six simulated merger remnants, from only one view of each simulated galaxy. Systematic errors in the shape parameters are <0.1. For details, see Statler, Lambright, & Bak (2001).

Related Links

Tom Statler's Publications and Preprints
RETURN to the Astrophysics Group's Home Page
Last updated 31 July 01.