2.4. Toward a Physical Morphology of Galaxies

Seen in the large, the parameter correlations discussed in the previous section lead to important successes, or at least to encouraging signs of progress. However, the implications of classical morphology for more detailed studies of internal processes in galaxies are less clear. As we come to understand galaxies better we find more and more morphological features that are physically meaningful enough to deserve a place in the classification. Some of these, such as inner and outer rings, are already included. Others are not; these include lenses (section 5.3), ovals (section 2.5.1) and the distinction between various kinds of spiral structure (section 2.5.3 and Fig. 2). The most important feature in galaxies is not even visible: halos are not included in any classification. Already the very incomplete Revised Morphological Types require about 100 classification cells (de Vaucouleurs 1963). Many people are already reluctant to use morphological types with the relatively modest complication of, say: (R)SB(rs)b. The large number of features demanding recognition would lead to an even more cumbersome notation, without necessarily making sense of the bewildering variety of observed forms. This situation motivates attempts to find a morphological approach which is more physically interpretable. In this section I describe one possible approach for constructing a physical morphology of galaxies, i.e., a classification system which relates more directly to processes of galaxy formation and dynamics (Kormendy 1979a,b, and references therein).

The essential feature of this approach is that galaxies are considered to be composed of a small number of building blocks, the distinct components of the mass distribution (Table 1). The aim of the classification is that each component consist of a collection of stars and gas whose structure, dynamics and origin are profitably thought of as distinct from those of the rest of the galaxy. For example, elliptical galaxies are known to have formed with much less dissipation than disks, resulting in different structure (section 3) and dynamics (section 4). Thus we can think of the terms elliptical (E) and disk (S) as being summaries of dynamical states and formation mechanisms, and not just descriptions of form. The aim of the component approach is as much as possible to replace descriptions of form with summaries of dynamical processes. The main problem with this procedure is immediately clear. We need to understand galaxies very well - certainly better than we do now - to set up the envisaged classification. Any preliminary attempts which are made now (Table 1) are much more open to future revision than are the classification systems of section 2.2.

The possible ways in which components can combine do not exhaust the variety of galaxy structure. All components can have properties which vary from galaxy to galaxy. For example, disks in late-type galaxies contain more young stars and gas than disks in early-type galaxies. Also, components can have secondary behavior, such as spiral density waves. We will see in section 5 that secondary components such as rings can form from the primary components (i.e., halos, bulges, disks and bars) through interactions. To some extent the distinction between new components produced by interactions and old ones modified by interactions is arbitrary. Generally, a new component is identified when the rearrangement of material is believed to be permanent, but not when the phenomenon is a temporary perturbation of existing structure. The maximum observed amplitude is also considered. Too detailed a classification is not useful, because observations of galaxies are limited in scope. (Thus no attempt is made to identify "components" at the level of detail usually used in Galactic structure studies.) Components in Table 1 are listed in a definite order; the formation of successive components in the list is progressively more and more derivative. Of course, the ordering is not rigorous in detail. However, there is no doubt that rings are less fundamental than, say, halos or bulges. Also, secondary components such as rings are arguably less fundamental than some kinds of features, such as spiral structure, in primary components. In general, the notation of Table 1 corresponds as closely as possible to that of Revised Morphological Types. Some of this notation refers to combinations of components, e.g., SB(r). Other parts refer to properties of components, e.g., the sequence Sa-Sb-Sc. Eventually the two different meanings of symbols should perhaps be distinguished. However, a more definitive physical morphology requires a better understanding of galaxies. The only present revision of Revised Morphological Types is to distinguish between lenses (l) and inner rings (r), see section 5. Thus the component approach is presently not a large revision of classical morphological types; it consists only of a change in their interpretation.

Justification of the component approach requires answers to the following questions (Kormendy 1979b). (1) Do galaxies have features in their brightness distributions which are clearly distinct and which occur in different proportions in different objects? (2) Are the origins and dynamics of these features different? (3) Does this approach help us to understand galaxies better, and in particular, is it a useful vehicle for deriving new results?

Progress toward answering these questions is discussed below and in Kormendy (1979b). Question (1) is relatively easy: the features listed in Table 1 are fairly easily distinguished photometrically; e.g., their brightness distributions generally have different functional forms. Of course, different components overlap. This limits the detail with which we can study component properties. Photometric properties of the basic components are the subject of section 3.

Sections 4 and 5 discuss kinematics and dynamics, and their implications for theories of formation of the components. We will see that the features studied to date are dynamically distinct, again within limits. The data show that transitions between the components in Table 1 are mostly continuous. This is particularly true of the sequence ellipticals - bulges - disks - barred disks. For example, the bulge of NGC 4594 is similar to an elliptical, while that of NGC 3115, which is E6 in its outer parts (Strom et al. 1977), is already rather like a thick disk (E7-8, Burstein 1979d). Also, in any given galaxy there are stars at all photometric and kinematic stages of transition between components such as bulges and disks, or disks and bars. That is, the phase space of a galaxy does not consist of isolated swarms of stars. Therefore, galaxies cannot conceptually be separated into components in arbitrary detail. Furthermore, even distinct components are not dynamical independent, and cannot be studied in isolation in arbitrary detail. As with most techniques, there are fundamental limits beyond which the component approach is no longer useful. "The situation is analogous to that in Hubble classification, where there is also a continuum of form, but where bins in the classification scheme still in isolating different integral properties of galaxies" (Kormendy 1979b).

Examples of how the component approach is useful for deriving new results are scattered throughout this paper, including section 3.4.3 on photometric decomposition, section 5 on bars, and section 2.5, below. One basic theme is that it is convenient to break up galaxy structure problems into relatively simple, physically homogeneous parts. More important, we are naturally led to think of interactions between the components. These turn out to be important because collective interactions have much larger amplitudes and shorter time scales than two-body interactions. This leads to studies of secular evolution in dynamics, discussed in section 5 and in Binney (1982a). As usual, morphology is efficient in suggesting the presence of previously unrecognized phenomena, but it must be followed up with detailed quantitative work. Within limits, then, the component approach is justified and useful in providing a framework for physical studies of galaxies.