The de Vaucouleurs classification volume recognizes three principal aspects of galaxy morphology, but clearly there are many more dimensions than three. Stage, family, and variety are the dimensions most clearly highlighted in blue light images and have a wide scope. Other dimensions may be considered and for some there is explicit notation in use.
6.1. Outer Rings and Pseudorings
Published de Vaucouleurs types include an extra dimension known as the outer ring/pseudoring classification. Several examples of outer rings and pseudorings are shown in Figure 13. An outer ring is a large, often diffuse structure, typically seen in barred early-type galaxies (stages S0+ to Sa) at a radius approximately twice that of the bar. Closed outer rings are recognized with the type symbol (R) preceding the main part of the classification. For example, an SB(r)0+ galaxy having an outer ring has a full classification of (R)SB(r)0+. Interestingly, rare cases of double outer ring galaxies, type (RR), are known, where two detached outer rings are seen; an example is NGC 2273 shown in the upper left frame of Figure 13.
Figure 13. Examples of outer rings (R) and outer pseudorings (R') in barred and nonbarred galaxies. Also shown is a rare example with two largely detached outer rings (RR). The galaxies are (left to right): Row 1 - NGC 7217, IC 1993, and NGC 2273; Row 2 - NGC 3945, NGC 1358, and NGC 1371. All images are from the dVA and are B-band except for NGC 2273, which is r-band.
In later-type galaxies, a large outer ring-like feature is often seen made of outer spiral arms whose variable pitch angle causes them to close together. These features are classified as outer pseudorings, symbolized by (R') preceding the main type symbols [e.g., as in (R')SB(r)ab]. Outer pseudorings are mainly observed in Sa to Sbc galaxies, and are only rarely seen in the very late stages Sc-Sm.
Among bright nearby galaxies, outer rings and pseudorings are found at about the 10% level (Buta & Combes 1996). Typically, outer rings are fainter than 24 mag arcsec-2 in blue light. With such low surface brightnesses, the rings can be easily lost to Galactic extinction. The division between outer rings and pseudorings is also not sharp. Some outer pseudorings are only barely distinguishable from outer rings. Continuity applies to these features as it does for inner rings although there is no symbol other than "S" for an outer spiral pattern which does not close into an outer pseudoring.
Although closed outer rings (R) are equally well-recognized in the RSA and the Carnegie and Hubble Atlases, outer pseudorings are a unique feature of the de Vaucouleurs revision. The value of recognizing these features is that many show morphologies consistent with the theoretical expectations of the outer Lindblad resonance (OLR, Schwarz 1981), one of the major low-order resonances that can play a role in disk evolution. Resonance rings are discussed further in section 10.1, but Figure 14 shows schematics of the morphologies generally linked to this resonance. The schematics are designed to highlight the subtle but well-defined aspects of these features, while Figure 15 shows images of several examples of each morphology, including the "models" used for the schematic. Outer rings of type R1 are closed rings that are slightly dimpled towards the bar axis, a shape which connects directly to one of the main periodic orbit families near the OLR as shown in Schwarz (1981) and in the dVA. Outer pseudorings of type R'1 are similar to type R1 but are made of two spiral arms that wind approximately 180° with respect to the ends of the bar. Even these will usually show a dimpled shape. Outer pseudorings of type R'2 are different from this in that two spiral arms wind 270° with respect to the ends of the bar, such that the arms are doubled in the quadrants immediately trailing the bar.
Figure 14. Schematic representations of outer Lindblad resonance (OLR) morphologies (Buta & Combes 1996).
Figure 15. Examples of OLR subclasses of outer rings and pseudorings. The galaxies are (left to right): Row 1 - NGC 1326, NGC 2665, ESO 509-98, and ESO 325-28; Row 2 - NGC 3081, UGC 12646, NGC 1079, and NGC 210; Row 3 - NGC 5945, 1350, 7098, and 2935. All images are B-band from the dVA.
The shapes R1, R'1, and R'2 were predicted by Schwarz (1981) based on "sticky-particle" numerical simulations. Not predicted by those simulations (but later shown in extensions of those simulations by Byrd et al. 1994 and Rautiainen & Salo 2000) is an interesting combined ring morphology called R1R'2, which consists of a closed R1 ring and an R'2 pseudoring. This combination is especially important because it demonstrates not only a continuity of morphologies among outer rings and pseudorings different from the continuity between outer rings and pseudorings in general, but also it is a morphology that can be linked directly to the dynamics of barred galaxies.
Note that the classification shown in Figures 14 and 15 does not depend on whether the rings are in fact linked to the OLR. The illustrated morphologies are abundant and easily recognized regardless of how they are interpreted. (Section 10.1 discusses other interpretations that have been proposed.) Although the Schwarz models guided the search for these morphologies, Rautiainen, Salo, & Buta (2004) and Treuthardt et al. (2008) showed that some outer pseudorings classified as R'1 are more likely related to the outer 4:1 resonance and not the OLR. These cases are generally recognizable by the presence of secondary spiral arcs in a four-armed pattern in the area of the bar (NGC 1433 in Figure 1 and ESO 566-24 in Figure 19 are examples).
The OLR subclassifications are used in the same manner as the plain outer ring and pseudoring classifications. For example, NGC 3081 has the full type (R1R'2)SAB(r)0/a.
6.2. Inner and Outer Lenses
The value of recognizing lenses as significant morphological components was first emphasized by Kormendy (1979), who suggested a dynamical link between inner lenses, which are often filled by a bar in one dimension, and dissolved or dissolving bars. Kormendy noted that lenses can be of the inner or outer type, in a manner analogous to inner and outer rings. He suggested the notation (l) for inner lenses and (L) for outer lenses to be used in the same position of the classification as inner and outer rings. For example, the galaxy NGC 1543 is type (R)SB(l)0/a while galaxy NGC 2983 is type (L)SB(s)0+. Figure 16 demonstrates the continuity between rings and lenses, which is evident not only among barred galaxies but among nonbarred ones as well. This continuity is recognized by the type symbol (rl), also used by Kormendy (1979). This type refers to a low contrast inner ring at the edge of a clear lens. Even underline classifications (rl) and (rl) may be recognized. A rare classification, (r'l), refers to an inner pseudoring/lens, a type of feature that is seen in NGC 4314 and recognized as such in the dVA.
Figure 16. Examples showing the continuity of inner rings (r) and lenses (l), for barred and nonbarred galaxies. The galaxies are (left to right): Row 1 - NGC 7187, 1553, and 4909; Row 2 - NGC 1326, 2859, and 1291; Row 3 - ESO 426-2, NGC 1211, NGC 1543. All images are B or g-band from the dVA.
Similarly, Figure 17 shows a continuity between outer rings and outer lenses through the type classification (RL), referring to an outer lens with a weak ring-like enhancement. Underline types RL and RL may also be recognized. The origin of outer lenses could be in highly evolved outer rings.
6.3. Nuclear Rings and Bars
The central regions of barred galaxies often include distinct morphological features in the form of small rings and secondary bars. The rings, known as nuclear rings because of their proximity to the nucleus well inside the ends of the primary bar, are sites of some of the most spectacular starbursts known in normal galaxies. The rings are typically 1.5 kpc in linear diameter and intrinsically circular in shape. Figure 18 (top row) shows three examples: NGC 1097, 3351, and 4314. These images highlight the small bulges that seem characteristic of nuclear-ringed barred galaxies. The three galaxies illustrated have types ranging from Sa to Sb, but based on the bulge size the types would be considerably later. For example, NGC 3351 has the bulge of an Sd galaxy.
Figure 18. Examples of nuclear rings and secondary (nuclear) bars. The galaxies are (left to right): Row 1 - NGC 1097, 3351, and 4314; Row 2 - NGC 1543, 5850, and 1291. All images are B-band from the dVA.
Comerón et al. (2010) carried out an extensive statistical study of nuclear ring radii, and identified a subclass known as "ultra-compact" nuclear rings (UCNRs). Such rings were recognized mainly in Hubble Space Telescope images and are defined to be less than 200pc in diameter. (See Figure 26 for an example in NGC 3177.) Comerón et al. showed that UCNRs are the low size tail of the global nuclear ring population. This study also showed that bar strength impacts the sizes of nuclear rings, with stronger bars generally hosting smaller nuclear rings than weaker bars.
Comerón et al. (2010) were also able to derive a reliable estimate of the relative frequency of nuclear rings as 20% ± 2% over the type range S0- to Sd, confirming with smaller error bars the previous result of Knapen (2005). Assuming that nuclear rings are a normal part of galaxy evolution, these authors argue that the rings may survive for 2-3 Gyr. Interestingly, it was also found that 19% ± 4% of nuclear rings occur in nonbarred galaxies, implying either that the rings may have formed when a bar was stronger (evidence of bar evolution) or that ovals or other mechanisms can lead to their formation. Mazzuca et al. (2009; see also Knapen 2010) connect some of the properties of nuclear rings to the rate at which the rotation curve rises in the inner regions.
The most extreme nuclear ring known is found in the SBa galaxy ESO 565-11 (see also section 7). At 3.5kpc in diameter, not only is it one of the largest known nuclear rings, but also the ring has an extreme elongated shape compared to more typical nuclear rings.
Nuclear bars lie in the same radial zone as nuclear rings and sometimes lie inside a nuclear ring. Three examples are shown in the second row of Figure 18. These average about one-tenth the size of a primary bar. There is no preferred angle between the axis of the nuclear bar and the primary bar, suggesting that the two features have different pattern speeds (Buta & Combes 1996; dVA).
Neither nuclear rings nor nuclear bars were recognized in the original Hubble-Sandage-de Vaucouleurs classifications, presumably in part because the use of small-scale photographic plates for extensive galaxy classification limited the detectability of the features in the (typically) overexposed centers. Modern multi-band digital imaging greatly facilitates the detection of the small rings and bars, allowing their inclusion in the classification. Buta & Combes (1996) and Buta et al. (2010a) suggested the notation nr for nuclear rings and nb 3 for nuclear bars, respectively, to be used as part of the variety classification as in, for example, SB(r,nr)b for NGC 3351, or SAB(l,nb)0/a for NGC 1291. Continuity may exist for these features like other rings and primary bars. [For example, nuclear lenses (nl) may also be recognized.] In blue light images, the appearance of the central region of a barred galaxy can be strongly affected by dust. For example, NGC 1365 shows a nuclear spiral in blue light, while in the infrared, the morphology is that of a nuclear ring (Buta et al. 2010a). The morphologies of some galaxies have a full complement of classifiable features. For example, accounting for all the rings and bars seen in NGC 3081, the classification is (R1R'2)SAB(r,nr,nb)0/a.
Lisker et al. (2006) use the terminology "S2B" for double-barred galaxies, a reasonable alternative approach to classifying these objects. Lisker et al. successfully identified nuclear bars in galaxies at redshifts z = 0.10-0.15 (from HST ACS observations), the most distant ones recognized thus far.
6.4. Spiral Arm Morphologies
A classification such as "Sb" tells one that a galaxy is a spiral of moderate pitch angle and degree of resolution of the arms, and that a significant bulge may be present. The type does not directly tell: (1) the multiplicity of the spiral pattern; (2) the character of the arms (massive, filamentary, grand design, or flocculent); or (3) the sense of winding of the arms (leading or trailing the direction of rotation). These are nevertheless additional dimensions to galaxy morphology.
The multiplicity of the spiral pattern refers to the actual number of spiral arms, usually denoted by the integer m. Examples of spirals having m = 1 to 5 are illustrated in Figure 19. The multiplicity is not necessarily straightforward to determine and may be a function of radius. For example, a spiral may be two-armed in the inner regions and multi-armed in the outer regions. Spirals of low m are usually grand design, a term referring to a well-defined global (meaning galaxy-wide) pattern of strong arms. The typical grand design spiral has two main arms, as in NGC 5364 (lower left frame of Figure 19). In contrast, a flocculent spiral has piecewise continuous arms but no coherent global pattern (Elmegreen 1981). NGC 5055 is an example shown in the middle left frame of Figure 19. This category is relevant mainly to optical wavebands. In the infrared, an optically flocculent spiral like NGC 5055 reveals a more coherent global grand design spiral (Thornley 1996; see also Figure 44), indicating that dust is partly responsible for the flocculent appearance.
Figure 19. Examples showing spiral arm character differences in the form of arm multiplicity, grand design and flocculent spirals, counter-winding spirals, and an anemic spiral. The galaxies are (left to right): Row 1 - NGC 4725, 1566, 5054, ESO 566-24, and NGC 613; Row 2 - NGC 5364, 5055, 4622, 3124, and 4921. All images are B-band from the dVA except NGC 5055, which is SDSS g-band, and NGC 4921, which is from Hubble Heritage.
The terms "massive" and "filamentary" arms are due to Reynolds (1927) and are discussed by Sandage (1961, 1975). Massive arms are broad, diffuse, and of relatively low contrast, as in M33, while filamentary arms are relatively thin in comparison, and lined by knots or filaments, as in NGC 5457 (M101). De Vaucouleurs (1956) originally used these distinctions as part of his classification, but later dropped the references to spiral arm character probably because of the complexity it added to his types.
Elmegreen & Elmegreen (1987) used a different approach to spiral arm character by recognizing a series of spiral arm classes based on arm continuity and length (but not necessarily contrast). Ten classes ranging from flocculent (ACs 1-4) to grand design (5-12; numbers 10 and 11 were later dropped). Examples of each are illustrated in Figure 20 (see Elmegreen & Elmegreen 1987 for a description of each class). Thus, spiral character is a well-developed additional dimension to galaxy classification. A simpler approach advocated by Elmegreen & Elmegreen is "G" for grand-design, "F" for flocculent, and "M" for multiple-armed. The arms of grand design spirals are in general thought to be density waves and may in fact represent quasi-steady wave modes (e.g., Bertin et al. 1989; Zhang 1996, 1998, 1999), although there is also some evidence that spirals may be transient (see review by Sellwood 2010). Flocculent spirals may be sheared self-propogating star formation regions (Seiden & Gerola 1982).
Figure 20. Examples showing the spiral arm classes of Elmegeen & Elmegreen (1987). The galaxies are (left to right): Row 1 - NGC 45, 7793, 5055, 2403, and 1084. Row 2: NGC 6300, 2442, 3504, 5364, and 1365. All images are B-band from the dVA, except NGC 5055 which is SDSS g-band.
Figure 19 also shows two examples of a new class of spirals, called counter-winding spirals. In these cases, an inner spiral pattern winds outward in the opposite sense to an outer spiral pattern. In the case of NGC 4622 (row 2, middle), the inner pattern has only a single arm and the outer pattern has two arms, while in NGC 3124 (row 2, middle right), the inner pattern is two-armed while the outer pattern is at least four-armed. The two cases are very different because NGC 4622 is essentially nonbarred while in NGC 3124, the inner spiral is classified as a bar. The presence of oppositely winding spiral patterns in the same galaxy means that one set of arms is trailing (opening opposite the direction of rotation) while the other set is leading (opening into the direction of rotation). In general, studies of the dust distribution as well as the rotation of spirals has shown that trailing arms are the rule (de Vaucouleurs 1958). Surprisingly, straightforward analysis of a velocity field and the dust pattern in NGC 4622 led Buta, Byrd, & Freeman (2003) to conclude that the strong outer two-armed pattern in this galaxy is leading, while the inner single arm is trailing. This led to the characterization of NGC 4622 as a "backwards spiral galaxy," apparently rotating the wrong way. An additional nonbarred counter-winding spiral has been identified in ESO 297-27 by Grouchy et al. (2008). In this case, the same kind of analysis showed that an inner single arm leads while a 3-armed outer pattern trails. No comparable analysis has yet been made for NGC 3124.
Vaisanen et al. (2008) have shown that a two-armed (but not counter-winding) spiral in the strongly interacting galaxy IRAS 18293-3413 is leading. Even with this, the number of known leading spirals is very small (dVA). Leading spirals are not expected to be as long-lived as trailing spirals since they do not transfer angular momentum outwards and this is needed for the long-term maintenance of a spiral wave (Lynden-Bell & Kalnajs 1972).
An interesting example of leading "armlets" was described by Knapen et al. (1995a), who used K-band imaging of the center of the grand design spiral M100 to reveal details of its nuclear bar and well-known nuclear ring/spiral. The nuclear bar has a leading twist that connects it to two bright K-band "knots" of star formation. This morphology was interpreted in terms of the expectations of gas orbits in the vicinity of an inner inner Lindblad resonance (IILR; Knapen et al. 1995b).
The final galaxy in Figure 19 is NGC 4921, an example of an anemic spiral. This is a type of spiral that is deficient in neutral atomic hydrogen gas, and as a consequence it has a lower amount of dust and star formation activity. The arms of NGC 4921 resemble those of an Sb or Sbc galaxy in pitch angle and extensiveness, but are as smooth as those typically seen in Sa galaxies. Anemic spirals were first recognized as galaxies with "fuzzy" arms (see van den Bergh 1998) where star formation has been suppressed due to ram-pressure stripping in the cluster environment. In the case of NGC 4921, the environment is the Coma Cluster. The idea is that such galaxies will eventually turn into S0 galaxies (van den Bergh 2009a). Anemic spiral galaxies are further discussed in section 10.2.
Seigar et al. (2008) have demonstrated the existence of a correlation between spiral arm pitch angles and supermassive central black hole masses. The sense of the correlation is such that black hole mass is highest for the most tightly wound spirals and lowest for the most open spirals. The correlation is expected because black hole mass is tightly correlated to bulge mass and central mass concentration, and spiral arm pitch angle is tied to shear in galactic disks, which itself depends on mass concentration (Seigar et al. 2005).
6.5. Luminosity Effects
Luminosity effects are evident in the morphology of galaxies through surface brightness differences between giants and dwarfs, and through the sophistication of structure such as spiral arms. van den Bergh (1998) describes his classification system which takes luminosity effects into account using a set of luminosity classes that are analogous to those used for stars. The largest, most massive spirals have long and well-developed arms, while less massive spirals have less well-defined arms.
The nomenclature for the classes parallels that for stars: I (supergiant galaxies), II (bright giant galaxies), III (giant galaxies), IV (subgiant galaxies) and V (dwarf galaxies). Intermediate cases I-II, II-III, III-IV, and IV-V, are also recognized.
Figure 21 shows galaxies which van den Bergh (1998) considers primary luminosity standards of his classification system. The original van den Bergh standards for these classes were based on the small scale paper prints of the Palomar Sky Survey. Sandage and Tammann (1981) adopted the precepts of the van den Bergh classes but revised the standards based on large-scale plates. In general, luminosity class I galaxies have the longest, most well-developed arms, luminosity class III galaxies have short, patchy arms extending from the main body, while luminosity class V galaxies have very low surface brightness and only a hint of spiral structure. The classes are separated by type in Figure 21 because among Sb galaxies, few are of luminosity class III or fainter, while among Sc and later type galaxies, the full range of luminosity classes is found. van den Bergh does not use types like Sd or Sm for conventional de Vaucouleurs late-types, but instead uses S- and S+ to denote "early" (smoother) and "late" (more patchy) subgiant spirals. Similarly, van den Bergh uses Sb- and Sb+ to denote "early" and "late" Sb spirals, respectively. (Some of these would be classified as Sab and Sbc by de Vaucouleurs.) According to the standards listed by van den Bergh (1998), an Sb I galaxy is 2-3 mag more luminous than an Sb III galaxy, while an Sc I galaxy is more than 4 mag more luminous than an S V galaxy.
Figure 21. Examples showing van den Bergh luminosity classes. The galaxies are (left to right): Row 1 - NGC 2841 (dVA B), 3675 (SDSS g), and 4064 (SDSS g). Row 2: NGC 5371 (dVA B), 5055 (SDSS g), and 4586 (SDSS g). Row 3: NGC 4321, 3184, 2403, 247, and 45 (all dVA B). The classifications are in the van den Bergh system (see van den Bergh 1998).
3 In a study of galactic nuclei, van den Bergh (1995) proposed the notation "NB" for nuclear bars, although what he refers to are not the same as the features described here. Back.