In my final lecture in this series, I want to cover a potpourri of topics to finish out the richness of galaxy morphology and to highlight some thoughts of what future studies of galaxy morphology and secular evolution might be like. Examining environmental effects may seem a little off the topic of secular evolution, but such effects do not have to be rapid or violent. Environmental effects can also produce exotic and very rare morphologies that are worthy of discussion. A few such morphologies (double detached outer rings, intrinsic bar-ring misalignment) have already been discussed in Lecture 2 (see Section 5).
7.1. Environmental effects in clusters
Gravitational interactions, ranging from minor, distant or fast encounters, to violent collisions and major/minor mergers, certainly play a role in galaxy morphology. Up to 4% of bright galaxies are currently involved in a major interaction (Knapen & James 2009).
The cluster environment offers a number of processes that can modify or even transform galaxy morphology. The cluster environment also brings attention to the idea of environmental secular evolution (see Kormendy's lectures, this volume), that adds a new dimension to morphological studies. The main processes or effects are:
Gas stripping has long been thought to be one of the dominant mechanisms of transforming a spiral galaxy into an S0 galaxy. Spitzer & Baade (1951) suggested that repeated collisions with other galaxies can strip a galaxy clean of its ISM. An alternative idea is that, as a galaxy moves in a cluster, it interacts with the intracluster medium in the form of X-ray emitting gas and can be stripped clean of its own interstellar gas, causing star formation to cease and spiral arms to disappear (Gunn & Gott 1972). Van den Bergh (1998) has argued that perhaps even some of the stellar mass is lost during the transformation, which might explain why S0s tend to have a lower luminosity than spirals.
If this process occurs and takes time, then partially stripped spiral galaxies ought to be observed. These are the galaxies referred to as `anemic' by van den Bergh (1976), and morphologically, these seem to be a legitimate class of objects. Figure 64 shows an example of an anemic spiral identified by van den Bergh (1976). Anemic spirals are cluster spirals deficient in Hi gas and as a result have a lower amount of dust and star formation. NGC 4921 is a Coma cluster spiral that strongly resembles a normal SBb spiral like NGC 3992, but has much smoother spiral arms. The smoothness indicates gas deficiency probably caused by stripping due to the intracluster medium. According to the hydrogen index parameter given in the RC3 (de Vaucouleurs et al. 1991), NGC 4921 is Hi deficient compared to NGC 3992 by a factor of more than six. Bothun & Sullivan (1980) argue that the galactic anemia `look' can be caused by other factors besides Hi deficiency.
 |
Figure 64. Comparison between the `anemic' barred spiral NGC 4921 and the more normal barred spiral NGC 3992. The van den Bergh (1976) parallel sequence classification is also shown. |
Environmental effects on the star-forming disks of Virgo cluster galaxies were examined by Koopmann & Kenney (2004) using Hα images. They recognised seven environmental categories based on star formation rates and the Hα disk as compared to a field sample. Class N had normal star formation rates for the galaxy type. Class E had star formation enhanced by a factor of three or more over normal, while class A (anemic) had star formation a factor of three or more below normal. The remaining categories are the truncated star-forming disk morphologies, several examples of which are shown in Fig. 65. Class T/N is `truncated/normal', meaning the star-forming disk is confined to the inner regions, but in those regions the star formation rates are normal. Class T/N[s] is similar but the truncation of the disk is severe. Class T/A is similar except that the star-forming part of the disk is anemic. Finally, class T/C is a truncated star-forming disk where the star formation is confined to a compact central area. Koopmann & Kenney (2004) illustrate the Hα images of all of these categories, while Fig. 65 (from B13) illustrates the truncated categories in blue light.
 |
Figure 65. Koopmann & Kenney (2004) environmental categories of Virgo cluster galaxies having truncated star-forming disks, illustrated using blue light images (from B13). |
These optical categories translate into similar categories of Hi distributions. The VLA (Very Large Array) imaging of Virgo spirals in Atomic gas (VIVA) survey (Chung et al. 2009) shows the interesting truncated morphologies that are found near the centre of the cluster (Fig. 66). Cases like NGC 4064, NGC 4405, and NGC 4457 are among the most extreme truncations. The survey also found evidence of ongoing stripping and fallback, the possibility that some galaxies have already fallen through the core and reached farther distances from the cluster centre as stripped spirals, as well as signs of gravitational interactions.
 |
Figure 66. VIVA images of Hi disks of Virgo cluster galaxies (from Chung et al. 2009, reproduced with permission). |
Yagi et al. (2010) identified a dozen galaxies in Coma that show ionised gas clouds mostly outside their disks. They suggest that these galaxies are likely to be `new arrivals' to the core region of Coma in different phases of stripping. These authors identified three morphological gas stripping categories:
A summary illustration of these categories can be found in B13.
7.2. Gravitational encounter phenomena
Among gravitational encounter morphologies, ring phenomena have captured a great deal of attention. Here I discuss what I will refer to as `catastrophic rings', or rings that have formed because of a collision between two galaxies. The main classes of catastrophic rings: accretion, polar, and collisional, must first be distinguished from the background of normal rings, usually referred to as `resonance rings' which were discussed in Lecture 2 (see Section 5). Figure 67 shows an example of each ring type.
 |
Figure 67. Examples of different galactic ring categories. |
`Resonance rings' constitute the vast majority of all rings observed. They are interpreted as features which form by gas accumulation at resonances, owing to continuous action of gravity torques from a bar (Buta & Combes 1996). The main resonances theoretically linked to rings are (Schwarz 1981, 1984a, b):
These theoretical links found strong support in statistics of intrinsic ring shapes and orientations, as well as in morphology (Lecture 2, see Section 5). However, there have been alternative interpretations suggested over the years. Regan & Teuben (2004) noted that the concept of a `resonance radius' (i.e., a single radius where a resonance occurs) is not really valid for strong bars because the epicyclic approximation breaks down. In the inner regions of a strongly barred galaxy, one instead has a broad `ILR region'. Because rings in models seem to be connected more to orbit transitions (i.e., orbit family changes, which are usually accompanied by orientation changes), Regan & Teuben (2004) preferred the term `orbit transition rings' for inner and nuclear rings at least, but still leaving outer rings and pseudorings in the `resonance' category.
A second alternative interpretation of rings is that they are connected to `invariant manifolds', or collections of orbits which emanate from the L1 and L2 Lagrangian points of a bar potential (Athanassoula et al. 2009a, b). This interpretation can successfully predict broad aspects of the morphologies of inner and outer rings in barred galaxies. Please see the lectures by Athanassoula (this volume) for more details.
The PDPS provides additional interpretations of normal rings. As described in Lecture 3 (Section 6), the nearby barred galaxy NGC 3351 has a bright inner ring whose morphology fits well into the resonance idea, yet the PDPS method suggests that the ring is a spiral mode with a different pattern speed from the bar. ZB12 interpret the ring as a spiral located at its own inner 4:1 resonance (not the bar's 4:1 resonance, because the bar in NGC 3351 is `superfast').
Accretion rings are thought to be made of material from an accreted satellite (Schweizer et al. 1987). Evidence for this is sometimes counter-rotation, where the material in the ring rotates in the opposite sense from the material in the main galaxy (Schweizer et al. 1989; de Zeeuw et al. 2002). The host galaxy can be an elliptical or disk galaxy. Examples of the former are Hoag's object (Schweizer et al. 1987) and IC 2006 (Schweizer et al. 1989), while an example of the latter is NGC 7742 (de Zeeuw et al. 2002). In the case of a disk galaxy like NGC 7742, the ring material mostly lies in the same plane as the receiving disk. It is also possible, as shown by the interesting case of NGC 1211 (Lecture 2, see Section 5), that a `dead' resonance ring galaxy could accrete a satellite in its outer regions and have a blue second outer ring.
Polar rings are also accretion features, except that the accreting object is usually a disk system, like an S0 (Schweizer et al. 1983). The most stable configuration is an accretion angle close to 90o. This limits the ability of differential precession to cause the ring to settle into the main disk (Schweizer et al. 1983). S0s are preferred probably because the disk is generally clean of interstellar gas and dust, allowing the high-inclination orbital material to pass unimpeded through the plane.
The most easily recognisable polar ring galaxies are those where the two disks are seen nearly edge-on (Whitmore et al. 1990). Cases where only one disk is edge-on are less obvious unless the main disk is a spiral rather than an S0. Cases like NGC 660 (Whitmore et al. 1990) and ESO 235-58 (Buta & Crocker 1993b) both have been classified as barred spirals, but both show an aligned dust lane along their apparent `bar' that demonstrates convincingly that their bars are actually edge-on disks. Thus, these galaxies show what the accreted high-inclination ring material looks like in a more face-on view. Cases where neither disk is edge-on are also possible but even harder to recognise. Some possibilities include NGC 1808, NGC 4772, and NGC 6870.
Collisional ring galaxies (also simply called `ring galaxies') are cases where a small galaxy has crashed down the rotation axis of a larger disk galaxy, triggering a radially expanding density wave (Lynds & Toomre 1976). Multiple rings are possible. Different morphologies represent different encounter parameters and different time frames. To be classified as collisional, there must be a viable intruder (Madore et al. 2009). Struck (2010) provides a useful overview of ring galaxy theoretical studies as compared with the limited observational material available.
The three catastrophic ring types are all much rarer than resonance rings. For example, according to Madore et al. (2009), only one in a thousand galaxies is a ring galaxy.
7.3. Interaction and merger morphologies
There are several categories of interaction and merger-driven morphologies. A dustlane elliptical is an E or E-like galaxy showing lanes of obscuring dust (Fig. 68). Minor axis, major axis, and misaligned lanes are found. Whether these should be classified as `ellipticals' or not was controversial, as de Vaucouleurs had once quipped: `If an elliptical shows dust, then it's not an elliptical'. Bertola (1987) established dustlane Es as a class of interacting galaxies where a small gas-rich companion undergoes a minor merger with a more massive E galaxy (Oosterloo et al. 2002). The current general view of these objects is to call them `dustlane early-type galaxies' (or dustlane ETGs; Kaviraj et al. 2011).
 |
Figure 68. NGC 5353 (right) is an example of a minor axis dustlane ETG. Its companion, NGC 5354, also has an inner dustlane. |
A shell galaxy is a normal elliptical or S0 galaxy showing faint ripples or ellipsoidal, sharp-edged features in its outer regions. Well-known examples are NGC 1344 and NGC 3923 (Fig. 69). Shells are likely to be 3D in geometry, not parts of a disk. Schweizer & Seitzer (1988) considered the term `shell' as imposing an unjustified interpretation on the shapes of the features, and suggested the term `ripples' instead. The extent of shells is huge, > 100 kpc in some cases. Typically, the features are interleaved in radius, although some may have an `all around' pattern (Prieur 1988). The best theory suggests that shells/ripples are remnants of a minor merger between a massive E galaxy and a small, cold, disk-shaped galaxy (e.g., Quinn 1984). The shells/ripples are thought to represent the maximum excursions of the disrupted orbits of the small companion's stars.
 |
Figure 69. Two `shell' elliptical galaxies (copyright David Malin, reproduced with permission). |
Tidal tails and bridges are common features of closely interacting but not necessarily merging disk galaxies. Tidal tails are a consequence of the tidal field due to the interaction and the shearing off of stars from the rotating disks. In M 51-type interacting pairs, a small companion is seen near the end of a bright spiral arm, as in M 51.
Figure 70 shows three examples of ongoing and advanced mergers. In these cases, the identities of the two separate galaxies are becoming less distinct, although the two separate nuclei may still be seen. The most extreme of such interactions may lead to ultra-luminous infrared galaxies (ULIRGs) (Fig. 70, lower frames). These objects have LIR > 1012 L⊙ ≈ Lbol of QSOs. As shown by Surace et al. (1988), high-resolution HST images of ULIRGs reveal clear evidence of likely strong interactions. These authors suggest that warm ULIRGs are transition cases to optical QSOs.
 |
Figure 70. Three cases of advanced and ongoing mergers (top row, from B13) and three ULIRGs (bottom row, from Surace et al. 1998, reproduced with permission). |
7.4. The morphology of active galaxies
The global morphology of active galaxies has always been of interest as people seek to understand the triggering mechanism of active galactic nuclei (AGN). Figure 71 shows three examples of active galaxies that highlight some of the features that have been implicated: rings, bars, and interactions. Hunt & Malkan (1999) found that the frequencies of outer rings and inner/outer ring combinations are three to four times higher in Seyferts than in normal spirals. Knapen et al. (2000) also showed that bars are more frequent in active galaxies than in non-active galaxies.
 |
Figure 71. Three active galaxies (B13). |
With HST resolution, the details of the host galaxies of quasars have been imaged. Bahcall et al. (1997) obtained HST images of quasars that show a variety of host morphologies, including Es, interacting pairs, systems with obvious tidal disturbances; and normal-looking spirals. These authors concluded that interactions may trigger the quasar phenomenon.
7.5. The morphology of brightest cluster members
These are the extremely luminous galaxies often found in the centres of rich galaxy clusters. Two of these were shown in Lecture 1 (see Section 3, Fig. 10). Figure 72 shows several more examples. Schombert (1986, 1987, 1988) classified the brightest cluster members according to luminosity profile shape:
The properties of the brightest cluster members fit well with merger simulations, including accretion and cannibalism (Schombert 1988).
 |
Figure 72. Four brightest cluster members as seen in optical images (B13). Classifications are from Schombert (1986). |
Edge-on views of many disk galaxies reveal a clear warping of the outer disk light. An excellent example is found in ESO 510-13 (Fig. 73). B13 shows three additional examples: NGC 4762, NGC 4452, and UGC 3697. These are all extreme cases detectable in optical light. Many more cases are found when the Hi layer is considered. The most promising interpretation of warps is that they are connected to the properties of dark matter haloes. For example, the stellar disk may be slightly misaligned with the equatorial symmetry axis of the dark halo. In this circumstance, the inner, more tightly bound part of the disk remains perfectly flat, while the outer parts of the disk are bent towards the equatorial plane of the halo, leading to warping of the disk at large radii (Binney & Tremaine 2008).
 |
Figure 73. The warped disk galaxy ESO 510-13 (Hubble Heritage). |
Do warps secularly evolve? Debattista & Sellwood (1999) argued that a warp would have a `winding problem' if the effects of a misaligned halo are not taken into account. Dynamical friction with a misaligned halo can drive a long-lived warp. Still, warps so driven are expected to be transient, because the friction force decays over time.
7.7. Non-barred ringed galaxies
This is an important but mostly neglected class of galaxies. While the majority of ringed galaxies are barred, a non-negligible fraction of non-barred ringed galaxies exists. Unlike barred galaxy rings, which are fairly homogeneous in metric and morphological properties, non-barred galaxy rings have a larger dispersion in properties that seems to point to a variety of mechanisms of ring formation.
NGC 7217 is a well-studied nearby example showing a spectacular blue outer ring in colour index maps (Buta et al. 1995b; Fig. 74, top row). At low light levels, the isophotes in NGC 7217 become almost exactly round, suggesting an extreme bulge-dominated system like the Sombrero Galaxy, NGC 4594. The strong bulge could therefore be the reason NGC 7217 is non-barred. Buta et al. (1995b) suggested that the strong outer ring formed in response to a subtle broad oval in the mass distribution.
 |
Figure 74. Examples of non-barred ringed galaxies. |
Other examples appear to be cases of former barred galaxies that in the course of secular evolution lost their primary bars. A good possible example of this is NGC 7702, a double-ringed late S0 that looks very much like an early-type barred galaxy with inner and outer rings and a nuclear bar (Buta 1991; Fig. 74, bottom row). However, no clear primary bar crosses the bright inner ring, although this ring is undoubtedly oval in intrinsic shape and therefore bar-like. Bar dissolution is a real possibility that may be tied to the gradual build-up of a strong central mass concentration (e.g., Hasan & Norman 1990).
Other non-barred ringed galaxies could involve galaxies that accreted a small companion, as described in Lecture 3 (see Section 6). Hoag's Object, IC 2006, and NGC 7742 would all be described as non-barred, and it is clear that a mixture of such objects, ex-barred galaxies, and possibly tidally driven rings like in NGC 4622 could account for the larger dispersion in the properties of non-barred galaxy rings.
This is a very rare, possibly interaction- or minor merger-driven morphology where a disk-shaped galaxy has two non-overlapping spiral patterns that wind outward in opposite senses. While the bulk of spirals have been demonstrated to be trailing the direction of rotation (de Vaucouleurs 1958), and the Toomre (1981) swing amplification mechanism (as well as the Lynden-Bell & Kalnajs 1972 mechanism) seems to explain why (Binney & Tremaine 2008), counter-winding spirals appear to present genuine examples of leading spiral arms.
Determining the sense of winding of spiral arms has typically depended on two things: knowing which half of the galaxy is receding from us, and which side of the galaxy is nearer to us. From the schematic in Fig. 75, right, it can be seen that on the near side, the bulge is viewed through the dust layer, while on the far side, the dust layer is viewed through the bulge. This leads to a reddening and extinction asymmetry that can be seen in the colour index map of NGC 7331 in Fig. 48 of Lecture 3 (see Section 6). Once the near-side is established, the velocity field tells us which way the galaxy must be turning. With these two pieces of information together, the sense of winding of the spiral arms can be reliably judged. Of course, the less inclined a galaxy is, the less evident is the nearside extinction and reddening effect.
 |
Figure 75. (Top row): V-band image of NGC 4622 with m = 1 and 2 Fourier decompositions (Buta et al. 2003). (Bottom frames): V - I colour index map (left), Fabry-Perot velocity field (upper, middle), colour image (lower, middle), all from Buta et al. (2003). Also shown at right is a schematic illustrating how tilt leads to an extinction and reddening asymmetry across the galaxy minor axis. The arrows on the colour image indicate the implied sense of rotation. |
The first counter-winding spiral identified, and the one that is best studied, is the nearly face-on SA(r)a spiral NGC 4622 located in the Centaurus cluster (Byrd et al. 1989). The galaxy has two high-contrast outer arms and a single lower-contrast inner arm, with no overlap between the patterns (Fig. 75, top row). Instead, the single inner arm and the two outer arms blend at the position of a conspicuous, offset inner ring. Buta et al. (2003) obtained HST imaging and a ground-based Hα velocity field, and used the dust distribution to judge the near-side and the velocity field to judge the sense of rotation (Fig. 75, bottom frames). A V - I HST colour index map showed thin curved dust lanes all lying on one side of the line of nodes, which can be identified as the near side. With the velocity field telling us that the north side is receding relative to the nucleus, this implied a clockwise rotation of the disk. This gave the surprising and unexpected result that the two strong outer arms, not the weaker inner arm, have the leading sense, a result that is difficult to accept. In the centre a small edge-on dust lane is found that suggests the galaxy has suffered a recent minor merger that could be responsible for the peculiar morphology.
A second example of a counter-winding spiral was identified by Buta (1995). This was the Sb-Sbc spiral ESO 297-27. Although not in a cluster environment, ESO 297-27 has similarities to NGC 4622. In this case, the inner arm leads and the outer arms trail (Grouchy et al. 2008). Other galaxies with leading spiral structure:
What causes counter-winding spiral structure? The phenomenon is very rare, so perhaps an unusual circumstance is at work. If the two outer arms of NGC 4622 are really leading, they would be difficult to explain in current theories of spiral structure. Swing amplification depends on the swing of a leading density wave. Also, the comprehensive study of Lynden-Bell & Kalnajs (1972) showed that only a trailing spiral pattern can transfer angular momentum outwards, which allows the wave itself to be maintained.
7.9. Giant low-brightness galaxies and stellar streams
Giant low surface brightness galaxies were first identified by Bothun et al. (1987). They are galaxies having a relatively normal bulge and an extremely low surface brightness, very large disk. The first example found was Malin 1 (Bothun et al. 1987). Figure 76 shows other examples: Malin 2, UGC 6614, UGC 1230, and in addition a strange possibly unrelated case called SGC 2311.8-4353, which was first studied in detail by Buta (1987). SGC 2311.8-4353 is a very low surface brightness companion to the giant spiral NGC 7531, and is notable for having an isophotal diameter 2/3 the size of NGC 7531. If at the same distance as NGC 7531, SGC 2311.8-4353 would be a dwarf in luminosity but not in size. That is, it would be a large low surface brightness galaxy, but not a giant. In a recent study of tidal streams in late-type spirals, Martínez-Delgado et al. (2010) have interpreted SGC 2311.8-4353 as a companion in the act of tidal disruption by NGC 7531. These authors show how commonly `normal' late-type galaxies can show extremely low surface brightness tidal features that are analogous to the streams found around the Milky Way.
 |
Figure 76. Giant low surface brightness galaxies, from B13 and references therein. The number above each box is the length of the top side at the distance of the galaxy. |
7.10. Magellanic barred spiral galaxies
These were identified as a distinct morphological class by de Vaucouleurs & Freeman (1972; Fig. 77; see also Lundmark 1927 and Lecture 1, Section 3). The class is characterised by a bar with no bulge, a single main spiral arm, shorter spiral features, and an offset of the centre of the bar from the centre of outer isophotes. De Vaucouleurs & Freeman noted that SBm galaxies often come in pairs (e.g., the large Magellanic cloud [LMC] and small Magellanic cloud [SMC], NGC 4618 and NGC 4625, NGC 2537 and NGC 2537A), suggesting that the morphology may be interaction-driven. Although the bar is usually offset in these galaxies, and in the LMC and SMC the bar is also kinematically offset from the rotation centre, prominent cases like NGC 4027 were found to have a rotation centre coincident with the centre of the bar (Pence et al. 1988). Too few cases have been observed to definitively prove that an offset rotation centre is characteristic of these galaxies.
 |
Figure 77. The LMC, the prototype of classification SB(s)m. Reproduced with permission from de Vaucouleurs & Freeman (1972). |
Compact galaxies (Fig. 78) have been known since Zwicky & Zwicky (1971) published a major catalogue. The departure of three compact elliptical (cE) galaxies, M 32, NGC 4486B, and NGC 5846AV-band surface brightness versus absolute V-band magnitude for E galaxies led Faber (1973) to suggest that these galaxies could be the stripped cores of formerly larger elliptical galaxies. However, it is now known that cE galaxies are simply the lower-luminosity tail of normal elliptical galaxies, based on photometric parameter correlations (Kormendy 1985; Kormendy et al. 2009). That is, cE galaxies are true `dwarf ellipticals'. These are to be contrasted with a large number of Virgo cluster objects classified as type `dE' by Bingelli et al. (1985) that have photometric properties more akin to dwarf irregulars than to normal ellipticals. As I have already noted in Lecture 1 (see Section 4), Kormendy & Bender (2012) have interpreted the Bingelli et al. (1985) dE, dE,N, dS0, and dS0,N types as environmentally driven morphologies, former irregular and very late-type galaxies that were stripped, harassed, or otherwise windblown to lose their gas.
 |
Figure 78. Examples of compact elliptical galaxies. |
7.12. Blue compact dwarf galaxies
These are small, high surface brightness galaxies experiencing a strong starburst (Fig. 79). As defined by Gil de Paz et al. (2003), blue compact dwarf (BCD) galaxies have (μB - μR)peak ≤ 1, μB < 22 mag arcsec-2, and MK > -21. On average, BCDs have B - R = 0.7 ± 0.3, MB = -16.1 ± 1.4, and logLHα = 40.0 ± 0.6. BCDs have been identified in the Virgo cluster by Bingelli et al. (1985; three examples are shown in Fig. 79), but are also known in field environments. In the Local Group, IC 10 has been identified as a BCD (Richer et al. 2001). Figure 79 shows an SDSS colour image of BCD NGC 2537.
 |
Figure 79. Examples of blue compact dwarf galaxies. |
The BCD class has brought to light some of the most extreme cases of star-forming galaxies. For example, the BCD I Zw 18 was once thought to be a genuine young galaxy with a high level of star formation and an optical morphology greatly affected by stellar winds and supernovae from an earlier starburst. However, deep imaging with HST revealed a definite older stellar population in I Zw 18 (Contreras-Ramos et al. 2011).
7.13. Ultra-compact dwarf galaxies
These are very compact, star-like galaxies with luminosities comparable to dE galaxies (Drinkwater et al. 2000, 2003). Hilker (2011) reviews the properties of a significant population of ultra-compact dwarf galaxies (UCDs) in the Fornax cluster. These objects have absolute magnitudes in the range -13.4 < MV < -11.4, and are brighter than the brightest Milky Way globular cluster but fainter than compact ellipticals. One interpretation of these objects is that they are the threshed nuclei of dE,N galaxies (that is, environmentally driven morphologies where a dE,N galaxy has lost all of its stars, except for the `N').
For investigations of the connection between morphology and secular evolution, there can probably be no better sample than isolated galaxies. For such galaxies, internal processes would have to be largely driving evolution.
A significant effort into establishing a reliable isolated galaxy sample was made by Karachentseva (1973) and was further improved by the Analysis of the Interstellar Medium in Isolated Galaxies (AMIGA) project (Verdes-Montenegro et al. 2005). Out of about 1000 galaxies in a final sample, Sulentic et al. (2006) found that the majority are Sb-Sc spirals, a few of which were shown in Fig. 4. The most striking thing about many of these galaxies is how regular and well-defined their spiral patterns are. AMIGA galaxies are selected to not have had a close encounter with a major galaxy in more than 3 Gyr. Do these galaxies support the idea that spiral structure can arise spontaneously, independent of an interaction?
7.15. Ultraviolet galaxy morphology
The ultraviolet (UV) is an important window into star formation in galactic disks. The recent launch of the Galaxy Evolution Explorer (GALEX, Martin et al. 2005) has provided an extensive database of images at 0.15 μm (far-UV, or FUV) and 0.22 μm (near-UV, or NUV). Gil de Paz et al. (2007) outline what can be learned from such images:
Figure 80 shows the GALEX FUV image of M 51 compared to a normal B-band image. The FUV image shows mainly hot stars younger than 108 years, while the B-band shows older stars in addition to dust and star-forming regions. Most interesting is how the companion galaxy, NGC 5195, and the extensive tidal debris around it, are mostly invisible at FUV wavelengths. This galaxy does not have a UV upturn that would make it more prominent in the FUV image.
 |
Figure 80. GALEX UV image of M 51 as compared to a typical B-band image. |
The NUV morphology of a spiral like M 83 strongly resembles an Hα image but with an older background disk.
B13 shows a montage of GALEX NUV images of galaxies of different types. In early-type galaxies, significant UV flux may occur if the galaxy has a star-forming nuclear ring, as in NGC 1317 and NGC 4314. In an intermediate type barred galaxy like NGC 3351, the inner ring, nuclear ring, and outer arms are conspicuous, while the bar is invisible. In the SBc galaxy NGC 7479, both the bar and the arms are seen, while the late-type galaxies NGC 628 and NGC 5474 have extensive UV disks. Most interesting was the discovery of an extensive UV disk around the diminutive Sm galaxy NGC 4625, which suggests that the galaxy is currently forming most of its stars (Gil de Paz et al. 2005). Finally, the intriguing star-forming galaxy NGC 5253 shows bright UV emission even from its extended diffuse disk.
7.16. The morphology of the interstellar medium
The ISM in galaxies has distinctive morphological qualities that are surely, in some cases, tied to secular evolution. The Hi Survey of Nearby Galaxies (THINGS, Walter et al. 2008) provides some of the best information on the Hi morphology of normal galaxies. In such galaxies we see:
Other galaxies, possibly because of their environment, have Hi disks comparable in extent to their optical disks, such as NGC 1433 (Ryder et al. 1996) and NGC 5850 (Higdon et al. 1998). In these cases, the Hi morphology follows the optical morphology closely. I have already described previously the Hi morphology of galaxies in the central regions of the Virgo cluster, where stripping has not only truncated the Hi disk significantly, but also has erased morphological structures so thoroughly that the diversity of Hi morphologies is greatly reduced.
The morphology of molecular hydrogen, H2, is also of interest. The Berkeley-Illinois-Maryland CO Survey of Nearby Galaxies (BIMA-SONG, Helfer et al. 2003) is one of the most extensive databases of molecular galaxy morphology available. Using CO 2.6 mm emission as a tracer of H2, BIMA-SONG maps reveal the following:
CO morphology does not necessarily mimic Hi morphology and the CO disk may not extend as far as the Hi disk.
Figure 81 shows several examples of Hi and CO maps of nearby galaxies, with a corresponding optical image for comparison.
 |
Figure 81. (Top rows): Hi maps of NGC 628, NGC 2841, NGC 2903, and NGC 3031 from THINGS (Walter et al. 2008); (bottom rows): CO maps of NGC 628, NGC 1068, NGC 2403, and NGC 2841 from the BIMA-SONG (Helfer et al. 2003). |
Especially interesting morphology is revealed by 8 μm emission maps obtained with IRAC. This mid-IR wavelength is sensitive to the warm dust associated with spiral arms and shows no near-side/far-side asymmetry, as the comparison in Fig. 82 shows. IRAC maps at 8 μm reveal the distribution of dust directly, instead of partly depending on how the dust layer projects against the bulge. In optical images, the spiral structure of M 81 appears to nearly terminate in a large pseudoring in the outer parts of the bright bulge. Spiral structure inside this pseudoring appears mainly in near-side extinction arcs. In contrast, at 8 μm there is complex spiral structure inside the apparent pseudoring that continuously winds outward.
 |
Figure 82. A B - I colour index map of M 81 as compared to an IRAC 8 μm image (from B13). |
7.17. High-redshift galaxy morphology
There is no doubt that high-redshift galaxy morphology has come of age during the past 10-15 years. The resolution provided by HST with the Wide Field Camera and the Advanced Camera for Surveys (ACS) has provided morphological information on thousands of high-redshift galaxies. Using both spectroscopy and photometric techniques, significant redshift ranges can be isolated to examine galaxy evolution firsthand.
B13 reviews many papers on high-redshift galaxy morphology. One of the first results found from the early surveys is that high-redshift galaxies reveal morphological categories that would fit poorly within the various modern classification systems. Unusual irregular shapes dominate the high-z population. Normal spirals and ellipticals can be recognised to z ≈ 0.6, but as z → 1, the number of irregular-looking objects becomes more significant. Clump clusters, linear and bent chains, `tadpoles', catastrophic rings, and mergers are identified in papers by Elmegreen et al. (2004, 2007), van den Bergh et al. (1996, 2000), and Cowie et al. (1995). Clump clusters and chains are shown in Fig. 83.
 |
Figure 83. Several high-redshift morphological categories, from B13 and references therein. The number in parentheses is the redshift. |
7.18. The Sloan Digital Sky Survey
The SDSS (Gunn et al. 1998; York et al. 2000) is without a doubt one of the most important assemblages of morphological information on galaxies since the Palomar Sky Survey. The survey includes morphological, photometric, and spectroscopic data for a million galaxies, and opened up the new era of huge extragalactic digital databases of medium-high resolution imagery. The SDSS also represents the advent of large-scale colour imagery for galaxies of all types.
Figure 84, top, shows several colour-type relations based on effective colour indices derived as outlined by Buta et al. (1994) and Buta & Williams (1995). All show the same general trend: colours of E-S0 galaxies are the same within the scatter, but there is a smooth decrease in colour indices from stage S0/a to Im. This morphology-colour relation is beautifully illustrated with SDSS colour images in Fig. 84, bottom. The transition from redder to bluer colours begins with spiral arms, leaving only bulges and bars remaining relatively red. However, as type advances, bulges decrease in relative importance and bars become bluer. By type Sm, the old stellar background is muted against the bright blue star-forming disk.
 |
Figure 84. (Top row): quantitative colour versus type relations, from Buta et al. (1994); from left to right, the colours are (B - V)eo, (U - B)eo, (V - R)eo, and (V - I)eo, where the subscripts mean the colours are within an effective (half-power) aperture and are corrected for extinction. (Bottom): a colour tuning fork based on SDSS colour images. |
The most famous result from SDSS multi-colour imaging is the `galactic Hertzsprung-Russell diagram', or colour-absolute magnitude diagram. This is shown in schematic form in Fig. 85. Galaxies that are made almost uniformly of old stars lie along the `red sequence', the familiar colour-absolute magnitude relation for early-type galaxies that was used extensively for distance scale studies in the 1980s. Star-forming galaxies, including spirals and irregulars, lie mainly in the `blue cloud'. The `green valley' is the name given to the zone intermediate between the red sequence and the blue cloud, and it is here where redder, early-type spirals are usually found. Galaxies in the green valley are thought to be evolving from the blue cloud to the red sequence, positioning themselves according to their absolute luminosity.
 |
Figure 85. A `galactic Hertzsprung-Russell diagram' showing the three main groupings of galaxies, the `red sequence', the `blue cloud', and the `green valley'. (Adapted from Wikipedia). |
7.19. Galaxy Zoo and citizen science
As I noted earlier, the SDSS is a goldmine for galaxy morphology. Not only are the images of high quality for classification, but the sheer number of images, on the order of a million, is beyond the capability of a small number of experts. An important question is, how to tap the information contained in the survey in a reasonable amount of time? This is what Galaxy Zoo (GZ) was designed to deal with: outsource galaxy morphology and classification to the Internet, and allow non-professional volunteers to participate. The history of how the project got started, and a summary of its results so far, is provided by Fortson et al. (2011).
Launched in 2006, GZ quickly became a model example of scientists giving something back to the general public: an opportunity to do science. The enlisted volunteers became known as `citizen scientists', and although not a new concept, the sheer number of such scientists who volunteered, more than 200,000 forming a tight-knit community of galaxy morphologists, was surely a wonder to behold. Eventually, two Zoo projects were activated: GZ1, which asked very basic questions about morphology, and GZ2 which was slightly more advanced. Some results from GZ1 are (see also Fig. 86):
One thing that GZ offered to its volunteers was a real chance for discovery. When examining so many images of galaxies that had not been studied in much detail before, someone was bound to find something new and unusual. This was the case with `Hanny's Voorwerp (Object)', a colourful cloud of ionised material and some active star formation located near a faint spiral galaxy, IC 2497. Figure 86 shows an HST image from Keel et al. (2012). Follow-up studies suggest that the Voorwerp is tidal debris from a past encounter between IC 2497 and another galaxy that drew out a tidal tail and triggered infall into a supermassive black hole, producing a transient quasar episode. The Voorwerp's light is an echo of the quasar phase acting on tidal debris (Lintott et al. 2009).
 |
Figure 86. Highlights from GZ1 (B13). Hanny's Voorwerp reproduced with permission from Keel et al. (2012). |
A second unusual type of object found in GZ is `green peas', which are compact star-forming galaxies having a high equivalent width of [Oiii] emission (Cardamone et al. 2009). A recent study by Amorín et al. (2010) showed that green peas are a distinct class of metal-poor galaxies, possibly affected by interaction-driven gaseous inflow over a short phase of evolution.
Another result from GZ is that bars and bulge-dominated galaxies tend to be found in denser environments than unbarred and disk-dominated counterparts (Skibba et al. 2012). Also, tidal dwarf galaxies, formed from ejected debris during a merger, were studied by Kaviraj et al. (2012) using the GZ merger sample.
7.20. Advanced galaxy morphology and classification
While GZ took large-scale galaxy classification to the public, which led to visual morphological information for hundreds of thousands of galaxies and useful information for follow-up studies, the need for more detailed and sophisticated types has led several professional groups to try and get more information on fine structure details with the ultimate goal of facilitating automated galaxy classification. The key to success in such studies has been the development of an interface that allows visual classification to be more efficiently and more accurately carried out. This can be done either in web-based fashion or off-web.
One of the first such studies was by Nair & Abraham (2010), who used SDSS g-band images to visually record morphological information on 14,034 galaxies brighter than magnitude g' = 16 and having z < 0.1. The survey went beyond RC3-style T-types to include recognition of rings, bars (regular and ansae types), spiral arm morphology, dust, and tidal features.
Another advanced morphological project was the Extraction of the Idealised Shapes of Galaxies from Imagery (EFIGI) survey (Baillard et al. 2011), which uses SDSS DR4 images of 4458 RC3 galaxies to get information on 16 morphological attributes, including features such as bulges, arms, bars, rings, dust, flocculence, hotspots, inclination, and environment. The procedure for doing the `morphometry' was for 11 astronomers to first classify a common subset of 100 galaxies to estimate relative biases and get each observer on a common scale compared to the scale of morphological T-types in the RC3. Then each observer individually classified 445 galaxies, or 0.1 of the final sample. The final sets were then homogenised to give final classifications on the RC3 scale. The visual classification was done using a sophisticated interface called `Manclass'. De Lapparent et al. (2011) analysed the statistical properties of the final EFIGI catalogue. The goal of the EFIGI catalogue is to set the stage for automated morphometry of the same attributes for a much larger sample.
Another sophisticated morphology project was the Wide-field Nearby Galaxy-clusters Survey (WINGS, Fasano et al. 2012). This survey provides morphological types of nearly 40,000 galaxies in 76 nearby galaxy clusters. The classification procedure began with 233 RC3 galaxies classified by two astronomers to evaluate the reliability of visual T-types. Then a single astronomer classified nearly a thousand randomly selected galaxies from the cluster sample to train an automated tool called MORPHOT, which uses neural networks and maximum likelihood techniques to extract types for the full cluster samples.
All of these studies highlight much of the future of galaxy morphology and classification, but most of all, they show that classical morphology still has relevance to modern extragalactic research and still has much to offer as we seek better understanding of galactic evolution.