3.1 The Morphology of Galaxies in Clusters
The fact that nebulae in clusters were different than nebulae in the field was known even before it was understood that these nebulae were galaxies like our own Milky Way. For example, Curtis (1918) observed 304 small diffuse nebulae in the Coma cluster, but his belief that "all the many thousands of nebulae not definitely to be classed as diffuse or planetary are true spirals, and that the very minute spiral nebulae appear as textureless disks or ovals solely because of their small size", kept him from realizing the importance of his observations. By the early 1930's the difference between field and cluster galaxies was clearly understood (Hubble and Humason 1931).
Oemler (1974) quantified this trend by defining three types of clusters with various ratios of elliptical, S0, and spiral galaxies. These are "spiral rich" (E/S0/S = 1/2/3; 17/33/50%), "spiral poor" (E/S0/S = 1/2/1; 25/50/25%), and "cD" clusters (E/S0/S = 3/4/2; 33/44/22%). For comparison, the percentages in the field are E/S0/(S+I) = 10/10/80% (Sandage and Tammann 1981). While these figures provide a basic framework for categorizing different types of clusters, the luminosity functions for different morphological types are not the same (Binggeli, Sandage, and Tammann 1988), so the actual percentages will change as a function of limiting magnitude. For example, Tully (1988b) finds percentages of E/S0/(Sp+I) = 3/7/90% in the lowest density regions from the Nearby Galaxies Catalog (Tully 1988a). This sample contains a large number of faint Sd and Im galaxies that would be completely missed in the more distant cluster samples.
Melnick and Sargent (1977) demonstrated the existence of population gradients as a function of distance from the center of the cluster, with elliptical galaxies occurring mainly near the centers and spirals predominating in the outer regions. These gradients make the determination of global population fractions dependent on how large a region is included in the calculation. The outer edge of a cluster is difficult to define and often merges into the field or the outer part of another cluster, so the global population is of limited usefulness.
The classic work in this field is unquestionably that of Dressler (1980). He determined positions, morphological types, magnitudes, and ellipticities for over 6000 galaxies in 55 clusters using excellent plate material. He concluded that the fundamental correlation was between morphological type and local projected galaxy density. Figure 1 shows Dressler's basic result, that the fractions of elliptical and S0 galaxies increase as a function of local galaxy density, while the fraction of spirals decrease. As Dressler states "The gradients are much more striking when the density is employed as the independent parameter, which indicates that the local projected density enhancements represent real physical associations and that populations are largely a function of local rather than global conditions". Dressler also examines the morphology-density relation for samples of low-concentration, high-concentration, and X-ray emitting clusters, and concludes that the relations for the three samples are very similar. Although the elliptical fractions look the same in all three samples, there is actually a fairly clear trend for the fractions of spirals to be lower (at the same local galaxy density) in high-concentration and x-ray emitting clusters than in low-concentration clusters. This suggests that the global properties of a cluster may also be important in determining the morphology of galaxies in clusters. We shall return to this point later, since Salvadore-Solé, Sanromà, and Jordana (1989), Sanromà and Salvadore-Solé (1990), and Whitmore and Gilmore (1990) have all suggested that correlations with global properties may actually be the fundamental correlation.
|
Figure 1. The fraction of E, S0, and S+I galaxies as a function of the log of the local projected galaxy density, in galaxies Mpc-2 for 55 clusters. An estimated scale of true space density in galaxies Mpc-3 is also included. The upper histogram shows the number distribution. (from Dressler 1980) |
Postman and Geller (1984) extended Dressler's morphology-density relation to less dense environments by analyzing the CfA Redshift Survey. They find that the fractions of elliptical, S0, and spiral galaxies within lower density regions in the field join smoothly onto the low density regions in the outer parts of Dressler's clusters. Sodré et al. (1989) have recently shown that the velocity dispersions for the system of spirals in clusters are generally larger than the velocity dispersions for the system of ellipticals. They interpret this as evidence that late-type galaxies are still in the process of falling into the virialized core of the clusters. This would explain the continuity of the population fractions from the field to the outer regions of clusters.
Giovanelli, Haynes, and Chincarini (1986) and Tully (1988b) both find that early and late spiral galaxies show different trends in the morphology-density diagram. The fraction of Sa's is approximately constant as a function of local galaxy density, while the Sc, Sd, and Im galaxies show a strong preference for the low density regions of the clusters.
Binggeli (1987) has provided tentative evidence that the shape of the luminosity function for each Hubble type is independent of the environment. Only the relative frequency of the Hubble types appear to change. If confirmed, this would rule out various scenarios for explaining the morphology-density relation by changing one type of galaxy into another (e.g., spirals into S0s via ram-pressure stripping).
Several mechanisms have been suggested as the cause of the morphology-density relationship, including both local and global mechanisms (i.e., initial conditions, tidal stripping, merging, tidal shear, ram-pressure stripping, gas evaporation, and truncated star formation). At present, none of these explanations are compelling. However, recent cosmological simulations (e.g., Frenk et al. 1985; Zurek, Quinn, and Salmon 1988; and Carlberg and Couchman 1988) offer hope that we may soon be able to pin down the relevant mechanisms, especially when it becomes possible to add dissipation in a physically meaningful way.