As noted by Sandage (1975), the first step in studying any class of objects is a classification of those objects. Classification built around small numbers of shared characteristics can be used for sorting galaxies into fundamental categories, which can then be the basis for further research. From such research, physical relationships between identified classes may emerge, and these relationships may foster a theoretical interpretation that places the whole class of objects into a global context. There is no doubt that such an approach greatly contributed to the development of the science of biology, and this is no less true for galaxies.
The genesis of galaxy classification is to take the complex combinations of structures described in the previous section and summarize them with a few type symbols. Sandage (1975) describes the earlier classification systems of Wolf, Reynolds, Lundmark, and Shapley that fell into dis-use more than 50 years ago. The Morgan (1958) spectral type/concentration classification system, which was based on a connection between morphology (specifically central concentration) and the stellar content of the central regions, was used recently by Bershady, Jangren, & Conselice (2000) in a mostly quantitative manner (see also Abraham et al. 2003). Thus, Morgan's system has in a way survived into the modern era but not in the purely visual form that he proposed. Only one Morgan galaxy type, the supergiant cD type, is still used extensively (section 10.6). Van den Bergh's luminosity/arm morphology classification system is described by van den Bergh (1998; see section 6.5).
The big survivor of the early visual classification systems was that of Hubble (1926, 1936), as later revised and expanded upon by Sandage (1961) and de Vaucouleurs (1959). Sandage (1975) has argued that one reason Hubble's view prevailed is that he did not try and account for every superficial detail, but kept his classes broad enough that the vast majority of galaxies could be sorted into one of his proposed bins. These bins were schematically illustrated in Hubble's famous "tuning fork" 1 (Hubble 1936; reproduced in Figure 2) recognizing a sequence of progressive flattenings from ellipticals to spirals. Ellipticals had only two classification details: the smoothly declining brightness distribution with no inflections, and no evidence for a disk; and the ellipticity of the isophotes, indicated by a number after the "E" symbol. (For example, E3 means the ellipticity is 0.3.) Spirals were systems more flattened than an E7 galaxy that could be subdivided according to the degree of central concentration, the degree of openness of the arms, the degree of resolution of the arms into complexes of star formation (all three criteria determining position along the fork), and on the presence or absence of a bar (determining the appropriate prong of the fork).
Figure 2. Hubble's (1936) "tuning fork" of galaxy morphologies is the basis for modern galaxy classification.
The S0 class at the juncture of the prongs of the fork was still hypothetical in 1936. As "armless disk galaxies," S0s were mysterious because all examples known in 1936 were barred. These were classified as SBa, but this was a troubling inconsistency because nonbarred Sa galaxies had full spiral patterns. Hubble predicted the existence of nonbarred S0s to fill the gap between type E7 2 and Sa and cure what he felt was a "cataclysmic" transition.
It was not long before Hubble himself realized that the tuning fork could not adequately represent the full diversity of galaxy morphologies, and after 1936 he worked on a revision that included real examples of the sought group: nonbarred S0 galaxies. Based on fragmentary notes he left behind, Sandage (1961) prepared the Hubble Atlas of Galaxies to illustrate Hubble's revision, and also added a third dimension: the presence or absence of a ring. This was the first major galaxy atlas illustrating a classification system in a detailed, sophisticated way with beautifully produced photographs. Hubble's revision, with van den Bergh luminosity classes (Sandage & Tammann 1981), was updated and extended to types later than Sc by Sandage & Bedke (1994).
Because Sandage (1961) and Sandage & Bedke (1994) describe the Hubble-Sandage revision so thoroughly, the details will not be repeated here. Instead, the focus of the next section will be on the de Vaucouleurs revision as outlined in the dVA. The reasons for this are: (1) the de Vaucouleurs classification provides the most familar galaxy types to extragalactic researchers, mostly because of extensive continuing use of the Third Reference Catalogue of Bright Galaxies (RC3, de Vaucouleurs et al. 1991); and (2) the de Vaucouleurs classification is still evolving to cover more details of galaxy morphology considered significant at this time. It should be noted that both the de Vaucouleurs and Hubble-Sandage revisions are strictly applicable only to z 0 galaxies and that it is often difficult to fit objects having z > 0.5 neatly into the categories defining these classification systems. High redshift galaxy morphology is described in section 13.
1 As recently noted by D. L. Block (Block et al. 2004a), this diagram may have been inspired by a similar schematic by Jeans (1929). Back.
2 van den Bergh (2009a) shows that E0-E4 galaxies are more luminous on average than are E5-E7 galaxies, suggesting that all E7 galaxies (and not many have been recognized) are actually S0 galaxies. Van den Bergh argues that genuine E galaxies may be no more flattened than E6. Back.