2.5.3. Varieties of Spiral Structure

The term "spiral galaxy" evokes in our minds a picture of objects like M81 and NGC 5364 (Fig. 2) which have two symmetric spiral arms of young stars and gas extending continuously from near the center to the edge of the visible disk. Such coherent two-armed spirals are used to illustrate the basic Hubble types in Figure 1. This picture of very regular structure has been emphasized since the earliest days of galaxy studies, because symmetry and beauty attract attention and demand explanation. Global spiral structure is now widely interpreted as a density wave whose shape rotates almost rigidly, while the disk material rotates differentially and flows through it. Extensive theoretical work (see Lin 1967, 1970; Contopoulos 1973; Toomre 1977a, 1981; Lin and Bertin 1981 for reviews) has produced, if not a complete theory of self-consistent density waves, at least a detailed discussion of the physical processes involved. Morphology played an important role in this success story: the theory was directly motivated by the emphasis (Oort 1962) on a compelling regularity observed in galaxy structure.

In contrast to classical density-wave studies, the emphasis in this section is on the variety of spiral structure forms. Global, two-armed or multi-armed spiral structure is not as prevalent as is generally believed. Also, in its most regular form it is confined to very luminous galaxies; this forms the basis of luminosity classification. Some of the varieties of spiral structure are illustrated in Figure 2, and are discussed in Kormendy and Norman (1979), and Elmegreen and Elmegreen (1982).

One distinctive kind of regular spiral structure is represented in Fig. 2 by the prototype, NGC 2841. The disk structure consists of a large number of short spiral filaments which do not join up to make a coherent pattern. Such structure is described and extensively illustrated in the Hubble Atlas. It is discussed further in Kormendy (1977d) and in Kormendy and Norman (1979), where it is called "NGC 2841-type" or "filamentary-armed" structure. Following Elmegreen (1981) and Elmegreen and Elmegreen (1982), I will here call it "flocculent" spiral structure. (My use of the term is slightly more restrictive than Elmegreen's, with more emphasis on lack of coherence than on patchiness.) Flocculent spirals form a distinct strain throughout the Hubble sequence, from Sa (NGG 2775, 3898) through Sb (NGC 2841, 3521, 5055, 7217) to Sc (NGC 4571, 5962, 7793); see the Hubble Atlas, Elmegreen (1981) and Fig. 2 for illustrations. Their structure is by no means irregular, but the regularity of the many spiral filaments is very different from that of global-pattern galaxies. The morphology suggests that two distinct kinds of physical processes are producing the different characteristic forms.

There is additional variety in observed spiral forms. I have already mentioned the distinction between spiral arms dominated by young stars and gas, and the smoother, usually low-contrast arms of van den Bergh's anemic spirals. Also, many spirals are just much more irregular than the folklore would suggest. If we only looked at flocculent, anemic or irregular spirals we would never be motivated to invent a density-wave theory to explain an unaccountable persistence of spiral structure. In fact, there has recently been a renewed emphasis on the probability that much spiral structure is transient, certainly in the more irregular and flocculent galaxies, and probably even in beautiful global spirals (Toomre 1977a, 1981). Galaxies which do not motivate a density-wave theory are not rare, in non-barred spirals they are the majority.

Whenever morphological features divide themselves into several natural and distinct groups, it is safe to assume that there is physics to be learned in studying these differences. This appears to be the case with global and flocculent spirals. Kormendy and Norman (1979) have examined the correlation of spiral structure types with rotation curve properties, motivated by the fact that the amount of differential rotation affects the difficulty with which the structure is maintained. In particular, theoretical work shows that it is relatively difficult to maintain global spiral structure in galaxies which contain an inner Lindblad resonance ⁽⁵⁾ (ILR). Kormendy and Norman found that global patterns occur in galaxies which plausibly contain an ILR only in the presence of a bar, an oval distortion or a companion. In the absence of these features a few galaxies which lacked ILRs had global spiral structure, but galaxies in which ILR could not be avoided lacked global spiral arms. Spiral structure in these galaxies was flocculent, or just irregular. This observation was interpreted as evidence in favor of the suggestion that bars, ovals and companions are efficient driving mechanisms for density waves (Toomre 1969). Additional morphological evidence for this suggestion is provided by Kormendy's (1979a) observation that spiral structure in barred galaxies is virtually always global, and is distorted to resemble oval rings (section 2.5.1). Elmegreen and Elmegreen (1982) and Elmegreen, Elmegreen and Dressler (1982) generally confirm this result, although it appears likely that bars do not invariably produce global spirals. They further show that the frequency of global spiral structure is also enhanced in galaxies with companions, and even more prevalent in galaxies with both bars and companions. These observations do not constitute proof, but do provide hints and supporting evidence for theory. The idea that bars can drive spiral structure is now well established (Toomre 1969, 1977a; Goldreich and Tremaine 1979; see also references in section 5.4). Driving of density waves by companions was difficult to understand in 1977, but is now believed also to be an efficient process (Toomre 1981).

Additional work on flocculent spirals would be valuable. One recent development is the availability of IV-N photographs in a 7300 - 9200 Å bandpass which is insensitive to young stars and gas. These photographs (Elmegreen 1981) show that the old stellar disks underlying global-pattern spirals have more regular spiral structure than the young stars and gas. Patches and interarm bridges are still visible, but they are reduced in contrast. On the other hand, flocculent spirals show less spiral structure in the infrared: the filaments are subdued and the galaxies look more like S0s than they do in the blue. This is consistent with the assumption that global-pattern spirals contain density waves with important contributions from the old disk, while flocculent spirals contain filaments of predominantly young material, with less participation by old stars. The arm segments in flocculent spirals may not be self-gravitating. It is possible that flocculent structure is produced by stochastic self-propagating star formation (Gerola and Seiden 1978; Seiden, Schulman and Gerola 1979; Seiden and Gerola 1979). This process may also be important in density-wave spirals (Elmegreen 1979), producing some of the irregularities, just as weak waves may exist in some relatively non-global spirals. (Sheared Jeans instabilities and transient events probably also contribute to the complication of spiral structure.) Since some spirals contain waves capable of compressing the interstellar medium and others do not, I suspect that there is considerable variety in star formation properties in different kinds of spirals. In particular, I suspect that flocculent spirals produce fewer massive stars than density-wave galaxies (Kormendy 1977d). No detailed comparison has been made of star formation in flocculent and grand-design spirals.

⁵ Near the center of the disk, stars revolve faster than the spiral wave. Their orbits are also non-circular. ILR occurs at that radius where the stars drift forward from one arm exactly to the other during one complete radial oscillation. That is, at ILR the orbits appear closed as seen by the spiral pattern. As a result, every time a star is at a given phase with respect to the pattern it is at a fixed phase in its radial oscillation. The pattern therefore repeatedly perturbs the orbit in the same way, producing large secular changes. This is a plausibility argument for the important theoretical result that spiral density waves are damped at ILR (see Toomre 1977a, p. 450). Back.