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

4.1 Which Processes Dominated at Galaxy Formation?

The formation processes can be divided into two categories: dissipationless processes, such as the formation of structure in pure cold dark matter models, and dissipational processes, which include star formation, cooling and heating of gas, etc. In an entirely dissipationless scenario, most of the observables can be related directly to the initial conditions. In a dissipational scenario, the physics of star formation, galactic winds, etc., determine many of the observables.

The r^1/4-law profiles can be produced entirely by dissipationless collapse or mergers (Section 2.7). Simulations have shown that realistic profiles result from a wide variety of initial conditions (4, 232, 340, 357, 363, 364). Most of these simulations did not distinguish between dark matter and luminous matter, and are somewhat difficult to interpret directly. Simulations of mergers between disk galaxies with dark halos produce realistic ellipticals (15, 16, 17, 115, 134, 260).

The scenarios with dissipationless hierarchical formation do not produce the current luminosity function directly. The typical structures formed at the present epoch are not galaxies groups and clusters of galaxies. As substructure tends to be erased in these scenarios, dissipational processes must have played a role (e.g., 368). The cooling characteristics of the baryonic matter are most often used to explain the typical galactic masses (285, 313, 366, 368). Recent simulations including dissipation were successful at preserving galaxies during the cluster collapse (57). Dissipational scenarios in which galaxies form from shocked gas usually invoke the same hierarchical process to explain the typical galactic mass (21, 105).

As discussed in Section 3.1, the origin of the Faber-Jackson relation and the fundamental plane is not entirely clear. One of the main uncertainties is the dark matter content of the luminous parts. These problems are related to the formation of spiral galaxies with their flat rotation curves (10). The formation scenarios generally predict a significant dark matter content for spirals in the visible regions (14, 51, 297). By analogy, the same may hold for ellipticals.

The color gradients and the color-magnitude relation for ellipticals no doubt must be explained in terms of gaseous processes, but the exact mechanism is not known. Radial inflows, galactic scale winds, and the hot X-ray gas envelopes probably play a role (56, 123, 152, 205, 206).

The slow rotation of bright ellipticals can be explained by a dissipationless merger scenario (15, 17). The rapid rotation of faint elliptical galaxies and the correlations of v / with various parameters are more difficult to understand. There is a wide range of parameters in a merger (e.g., relative masses, bulge/disk ratios, gas fractions, orbital parameters, etc.), so that scatter in the properties of the merger remnants can be expected. Whether systematic changes of v / can be produced remains to be seen.

The analysis of the intrinsic shapes of elliptical galaxies may provide equally interesting information. The predictions from dissipationless simulations are usually more triaxial than permitted by the data (124). Both the data sample and the analysis of the simulations should be improved before we can draw firm conclusions. However, if this trend is confirmed, then we may need the inclusion of gas dynamics to explain the discrepancy. Gas disks may be able to deform the galaxies, as the gas tends to settle in an elliptical disk, which is elongated perpendicular to the main body of the galaxy.

We conclude that both gaseous and collisionless processes played an important role during galaxy formation, and determined many of the observed properties. Significant progress has been made toward modeling of dissipationless processes. The challenge is to achieve a similar level of sophistication for gaseous processes, as these are crucial for our understanding of many of the fundamental parameters.