Galaxy morphology became evident when large and effective telescopes began to be used to observe the sky. In the late 18th century, English astronomer William Herschel built 18.7 in speculum metal reflectors which he used to `sweep' the sky for anything out of the ordinary, such as `nebulae'. Herschel and those who followed him saw different kinds of nebulae: those that appeared to lie within the band of light called the Milky Way (`galactic nebulae'), and those which were found mainly away from the Milky Way (`non-galactic nebulae'). For a while, all types of nebulae were thought to be distant but unresolved stellar systems, a popular but largely speculative idea at the time. Prior to the 1870s, all illustrations of nebulae were visual sketches based on what was seen through an eyepiece. One of the most detailed early sketches was that of M 51 made by Herschel's son John in the 1820s (Hoskin 1982).
The non-galactic nebulae seen by Herschel had a variety of interesting shapes, ranging from round to highly elongated, and showed varying degrees of central brightness. Herschel invented a simple descriptive classification of these objects based on brightness, size, shape, and central concentration. This approach was also used by John Herschel to describe all nebulous objects compiled to the 1860s (Herschel 1864).
Even though the Herschel telescopes could reveal thousands of non-galactic nebulae, the finer details of galaxy morphology were largely elusive. Galaxy morphology `came alive' in 1845 when William Parsons, 3rd Earl of Rosse, observed nebulae with a much larger telescope, the 1.8 m `Leviathan of Parsonstown'. This telescope was constructed on the grounds of Birr Castle in central Ireland, and was the largest telescope in the world for nearly 75 years. The Leviathan observations are replete with visually seen details of galaxy morphology, and from the sketches that were made, one can tell that the observers had seen spiral arms, bars, rings, dust lanes, star-forming regions, tidal features, even Magellanic barred spirals. In the extensive set of notes published by the 4th Earl (Parsons 1880), one can determine that spiral structure was seen in 75 `nebulae' later found to be galaxies. Figure 5 shows what could be viewed as the first galaxy morphology atlas: a compilation of Birr Castle Leviathan sketches made by a variety of observers.
 |
Figure 5. The first galaxy morphology atlas, based on visual sketches made with the Birr Castle `Leviathan' (Parsons 1880). |
Although the Leviathan of Parsonstown was the most powerful telescope of its day, it was not capable of guided photography and therefore could not get long exposure images of things like spiral nebulae. Nebular photography became possible in the 1880s with the availability of silver bromide dry emulsion plates and telecopes designed for accurate guiding. Isaac Roberts (1829-1904) obtained in 1888 the first long-exposure photograph of the Andromeda Nebula, which first revealed the spiral structure in the faint outer parts of the nebula (Roberts 1893). As the number of plates accumulated, the first classification systems emerged. Max Wolf (1863-1932) published a simple system of letters to describe 17 different types of non-galactic nebulae, ranging from amorphous inclined types to patchy, well-developed spirals (Wolf 1908). This system was used over a period of 30 years by a number of well-known nebular researchers, including Hubble who thought it was a useful temporary system until accumulation of more data allowed something better to come along.
A big photographic survey described by Curtis (1918) helped set the stage for the Hubble classification system. Photographs of hundreds of nebulae, galactic and non-galactic, were taken over a nearly 20-year period with the the 36 inch Lick Crossley reflector by, in addition to Curtis, well-known photographers James E. Keeler and Edward E. Barnard. The main specific galaxy morphology Curtis recognised was `ϕ-type spirals', later renamed barred spirals by Hubble (1926). Curtis believed all non-galactic nebulae were spirals and that any that didn't look spiral would eventually be found to be such. Hubble (1922) disputed this conclusion; he noted that genuine bright but definitely non-spiral non-galactic nebulae existed. After obtaining and inspecting many available plates, Hubble (1926) published a new classification system to replace the Wolf (1908) system. This system placed galaxies on a sequence ranging from amorphous elliptical-shaped objects to well-developed, patchy-armed spirals. The spiral part of the sequence was split between non-barred (`normal') and barred spirals. However, Hubble believed that his 1926 system was flawed because the transition from the flattest-looking E galaxies to Sa spirals looked too sharp to be real. He hypothesised that there had to be armless but highly flattened disk-shaped galaxies in the transition from types E to Sa. This was shown in his famous `tuning fork' illustration in his book, The Realm of the Nebulae, published in 1936 (Fig. 6, left). It is thought that Hubble was inspired to illustrate his classification this way because this is how Sir James Jeans illustrated it in his 1928 book, Astronomy and Cosmogony (Block et al. 2004). Sandage (2005) has also noted that the arrangement of galaxies on a sequence ranging from amorphous to highly structured spirals was outlined independently by Reynolds (1920), but never referenced by Hubble.
 |
Figure 6. The original Hubble (1936) `tuning fork' classification (left), and a schematic of the revised tuning fork outlined in the Hubble Atlas (Sandage 1961; Buta et al. 2007). |
Hubble had planned to revise his classification system further but died before completing it. Allan Sandage used Hubble's notes to prepare the monumental Hubble Atlas of Galaxies (Sandage 1961). This firmly cemented Hubble's ideas into astronomy. Inclusion of S0 galaxies, as well as splitting each tuning fork prong into ringed and non-ringed varieties, made the classification more complicated. Van den Bergh (1976) commented that the addition of S0s destroyed the `simple beauty' of the original 1926 classification. The revised Hubble-Sandage (RHS) system was later expanded and further revised in the Carnegie Atlas of Galaxies (Sandage & Bedke 1994).
S0 galaxies, probably the most enigmatic type in the whole Hubble sequence, are thoroughly described in the Hubble Atlas and Carnegie Atlas. An excellent example of an S0 is NGC 2784, shown in Fig. 7, while the RHS classification of S0s is summarised in Fig. 8. The three main components of an S0 are the nucleus, the lens, and the envelope. A lens is a distinct feature that appears as a well-defined region having a shallow brightness gradient interior to a sharp edge. The enhancement can be very slight as in NGC 2784, or more distinct as in NGC 1411 (Fig. 8, top middle). The RHS subclassification of nonbarred S0s, S01, S02, and S03, depends on structure differentiation, while that for barred S0s: SB01, SB02, and SB03, is based on the development of the bar. The differences between nonbarred and barred S0s are important. For example, dust rings are a common feature of type S03, while they may not factor in at all in type SB03.
 |
Figure 7. NGC 2784 displays the main elements - nucleus, lens, and envelope, that define an S0 galaxy (Sandage 1961). |
 |
Figure 8. These galaxies show the main S0 galaxy categories descibed in the Hubble and Carnegie atlases of galaxies. |
Around the same time as the Hubble Atlas was being prepared, W.W. Morgan (1906-1994) proposed a classification system that combined galaxy form with stellar population defined by (Morgan 1958):
population group: a, af, f, fg, g, gk, k for dominant spectral types A, AF, F, FG, G, GK, and K. This is estimated solely from central concentration;
form family: S (spirals), B (barred spirals), E (ellipticals), I (irregulars), D (like S0s), plus others;
In Fig. 9, a Morgan sequence of population groups, based on classifications from Morgan (1958) and illustrated using SDSS colour images, shows the effectiveness of his approach. The sequence a through k is a colour sequence from bluish to yellow-orange. The system did not have the impact that Hubble's did, perhaps because the population groups were closely analogous to Hubble types Sa, Sb, and Sc. For example, in Fig. 9, NGC 3389 (type aS4) is Hubble type Sc, NGC 3583 (type fgS4p) is Hubble type Sb, and NGC 4260 (type gkB4) is Hubble type Sa.
 |
Figure 9. Examples of the stellar-population/form class classification of Morgan (1958). |
The most recognisable and important of Morgan's form classes is the cD galaxy, a supergiant version of the D form family found in the centres of rich clusters (Fig. 10). These objects, also known as `brightest cluster members', were extensively studied by Schombert (1986, 1987, 1988).
 |
Figure 10. The Morgan cD class of supergiant galaxies. |
In 1953, non-Palomar firebrand Gerard de Vaucouleurs (1918-1995) carried galaxy morphology into the southern hemisphere, developing his own interpretation of the Hubble-Sandage classifications on the way. Figure 11 shows de Vaucouleurs's (1959) `classification volume' (the VRHS, or `three-pronged swirling two-handled tuning fork'). The long axis defines the `stages' E, E+, S0-, S0o, S0+, S0/a, Sa, Sab, Sb, Sbc, Sc, Scd, Sd, Sdm, Sm, Im.
 |
Figure 11. The de Vaucouleurs (1959) revised Hubble-Sandage classification system. |
De Vaucouleurs viewed galaxy morphology as a continuous sequence of forms. He was also artistically talented and made a sketch (see Kormendy's contribution, this volume) of a cross-section of his classification volume in 1962. The sketch shows the arrangement of families (apparent bar strength) and varieties (presence or absence of an inner ring) near stage Sb. Families and varieties can be thought of as continuous secondary traits (de Vaucouleurs 1963); Fig. 12 shows the use of the underline notation for these characteristics.
 |
Figure 12. Family and variety in the VRHS as continuous characteristics (from B13). |
The stage is the primary dimension of the VRHS. Elliptical galaxies are amorphous systems with a smoothly declining brightness gradient (Fig. 13). They are not disk-shaped. Elliptical galaxies have two VRHS stages: E and E+. The number after the letter E, as in E2, is the flattening index n = 10(1 - b / a), a numerical specification of how elliptical the galaxy is. The En sequence is not, however, physically significant. Type E+ is a transition stage to the S0 class. These are E-like galaxies showing slight traces of differentiated structure, usually subtle evidence of lenses or faint outer envelopes. Type E+ has also been used by de Vaucouleurs as a `home' for Morgan cD galaxies.
 |
Figure 13. An elliptical galaxy (E) and a `late elliptical' galaxy (E+). |
In the VRHS, S0 galaxies have three stages: S0-, S0o, and S0+, in a sequence of increasing structure. Type S0- generally has barely differentiated structure, often in the form of subtle lenses. Type S0o tends to have stronger lenses and is more obvious as an S0. Type S0+ often has well-defined ring structures, both inner and outer, but can also have subtle spiral structure. Types such as SB(r)0o and SB(s)0+ are possible, as are SAB types. Note that this sequence is based on development of structure and NOT on bulge-to-total luminosity ratio.
In the VRHS, spirals have 9 stages: S0/a, Sa, Sab, Sb, Sbc, Sc, Scd, Sd, Sdm, and Sm. These are still recognised using Hubble's three criteria: the relative size of the bulge, the degree of openness of the spiral arms, and the degree of resolution of the arms into knots. Small-bulge early-type galaxies, especially barred spirals, are the biggest violators of these rules. The VRHS introduces extreme late-type spirals: Sd, Sdm, and Sm, to the Hubble sequence. Figure 14 shows how well these types fit into the sequence with clear, easily distinguishable characteristics. The intermediate spiral types (like Sab, Sbc, etc.) are almost as common as the main types. The most common spirals in magnitude-limited samples are of types Sb-Sc. A special hallmark of the VRHS is the recognition of the Magellanic Clouds as extreme late-type barred spirals of the type SB(s)m that show a characteristic one-armed asymmetry and offset bar (de Vaucouleurs & Freeman 1972). Magellanic irregulars form the last major stage along the VRHS, and are often barred (i.e., classified as type IBm or IB(s)m, implying a subtle spiral variety). Examples highlighting the S0, spiral, and irregular sequences are shown in Fig. 14.
 |
Figure 14. Examples of the S0 and spiral stages along the VRHS sequence. |
Spiral and S0 stages are still recognisable in edge-on galaxies, but Fig. 15 shows one case that is distinctive. NGC 3115 is an original Hubble E7 type that was reclassified as type S01 in the Hubble Atlas. It is now recognised as an S0 with a `thick' disk, although it looks more like an E galaxy with an embedded disk (Fig. 15).
 |
Figure 15. Edge-on S0 galaxy NGC 3115 (V-band). |
The outer ring classification is the final original dimension of the VRHS. Closed outer rings are symbolised by (R) preceding the other type symbols. The hallmark of VRHS classifications is the recognition of outer pseudorings (R') made of variable pitch angle outer spiral arms. Examples are illustrated and described in Lecture 2 (Section 5).
The genesis of a galaxy classification is shown in Fig. 16. The main features of the galaxy are labelled and its basic type is (R)SAB(r)0/a. However, its bulge is inconsistent with the type S0/a as is often the case for early-type barred galaxies.
 |
Figure 16. Classification of NGC 3081. |
The VRHS has had some additions and revisions in recent years. For example, inner and outer lenses were added to the classification by Buta (1995) using notation suggested by Kormendy (1979). Inner lenses are found roughly in the same location as inner rings, while outer lenses are found where outer rings often are seen. Figure 17 shows that there is a morphological continuity between rings and lenses, and one possible interpretation is that some lenses are highly evolved former star-forming rings.
 |
Figure 17. Examples showing the continuity of ring and lens morphologies. Other examples may be found in B13. |
The effect of total mass and luminosity on galaxy morphology was characterised by van den Bergh (1960a, b) in terms of luminosity classes. (An illustration of the luminosity class standards, van den Bergh 1998, is given in B13). The classes are applicable only to intermediate to late-type spirals and are analogous to those used in stellar spectral classification: luminosity class I: supergiant spirals; II: bright giant spirals; III: giant spirals; IV: subgiant spirals; and V: dwarf spirals and Magellanic irregulars. The main idea is that more luminous spiral galaxies have the most well-organised structure (Fig. 3). In massive, luminous spirals, the structure is more symmetric and more ordered. In lower-mass spirals, the structure is more chaotic; in fact, in dwarfs the spiral structure can be so weak or absent that surface brightness is used as the criterion of luminosity classification.
Dwarf galaxies in the Virgo cluster were the subject of an exceptional study by Binggeli et al. (1985). The main types of dwarfs classified by these authors are: cE - compact ellipticals; dE - dwarf ellipticals; dE,N - nucleated dwarf ellipticals; dS0 - dwarf S0s; dS0,N - nucleated dwarf S0; BCD - blue compact dwarf; and `large dE' or large dwarf ellipticals (all illustrated in B13). If the dEs and dS0s were actually dwarf versions of normal E and S0 galaxies, their existence would widen the parameter space at the left end of the VRHS. However, recent studies have indicated that dEs and dS0s are likely not connected to actual E or S0 galaxies but to Magellanic irregulars. Kormendy & Bender (2012) have suggested that dEs and dS0s are environmentally modified Magellanic irregulars, and that the true dwarf members of the E galaxy class are objects like the compact Es in Fig. 78. Kormendy & Bender have proposed renaming all dE and dS0 galaxies `spheroidals'. This connects these objects directly to galaxies referred to as `dwarf spheroidals' and dwarf irregulars in the Local Group. Detailed studies of Local Group dwarf spheroidals and irregulars reveal systems with a complex star formation history (Mateo 1998).
Dwarf spirals are a controversial subject; only a few genuine cases are known. IC 3328, a dE, was shown to be a true dwarf spiral by Jerjen et al. (2000). NGC 3928 was recognised as a dwarf spiral by van den Bergh (1980). It is shown relative to the supergiant spiral UGC 6614 in Fig. 18.
 |
Figure 18. Dwarf spiral NGC 3928 as compared to the supergiant spiral UGC 6614. |