4.4. Assorted Issues
The Mg2- relation (Fig. 13) is tighter than other population-to-structural correlations such as Mg2-MB or <Fe>-. (Bender, Burstein, & Faber 1993) There is some powerful connection between velocity dispersion and Mg abundance as traced by the Mg b feature, the exact nature of which eludes us at the moment. One possibility, suggested by Faber et al. (1992), is that cloud-cloud collision velocity modulates the IMF to favor more massive stars in high- environments. This suggestion is in harmony with the Mg/Fe abundance trend. Another possibility is that larger local escape velocities resist supernova winds more effectively, holding onto heavy-element contaminants better. Star formation is finally truncated when supernova winds are able to blow the remaining gas out of the galaxy. This is the now-standard picture of chemical evolution in elliptical galaxies (see, e.g., Arimoto & Yoshii 1987; Matteucci & Tornambé 1987). This picture can also be made harmonious with the Mg/Fe trend if there is an additional mechanism for varying Mg/Fe as a function of galaxy size, but the fact that Fe abundance is virtually independent of galaxy size causes some trouble.
Discontinuities in kinematic profiles are coincident with discontinuities in line strength profiles. Bender & Surma (1993) discovered in the course of studying peculiar kinematics in elliptical galaxies that many have counterrotating cores or other kinematic discontinuities. In every case that they studied, a discontinuity appeared in the Mg2 line strength profile at the same location as the kinematical discontinuity. At the very least, this implies that whatever formation mechanism produced the distinct core also influenced the local chemistry. Muted echoes of formation exist still in both the stellar kinematics and the chemical signature in the stars.
Study of the globular cluster systems around elliptical galaxies yields insight into the formation of halos in general and elliptical galaxies in particular. Some elliptical galaxies host a very large number of globular clusters per unit luminosity (e.g., M87), while others have about as many as are seen in spiral galaxies. Questions remain about how the globular clusters are created and destroyed to explain the wide variation in number and whether merging events are important or not (van den Bergh 1995; Zepf & Ashman 1993). Most abundance studies concentrate on the integrated colors of the clusters (Ostrov, Forte, & Geisler 1998; Lee, Kim, & Geisler 1998; Ajhar, Blakeslee, & Tonry 1994) because of their faintness, but some spectroscopic studies are beginning to appear (see, e.g., Cohen, Blakeslee, & Ryzhov 1998). These efforts show a variety of interesting results. For example, some globular cluster systems are metal poor, some metal rich, and some bimodal or multimodal. And when both metal-poor (blue) and metal-rich (red) populations coexist, the red population tends to be more centrally concentrated than the blue.