Though the nomenclature was very simplistic, the morphological classification of galaxies can be traced back to the time of the Herschels (1781-1847), whose telescopes were large enough to allow visual recognition of distinct differences in the large-scale characteristics of the so-called "white nebulae". Different degrees of central concentration, apparent flattening, and mottling were clearly distinguishable. More complex structure was seen in a few of the brighter cases, but it was Lord Rosse who, in 1845, added the attribute "spiral" to some members of the Herschels' white nebulae. This is when morphology began to get interesting.
The advent of photography in astronomy at the end of the 19th century firmly established the reality of the spiral morphology. Photography also provided a greater appreciation of the complexity and range of galaxy morphology that must have proved almost daunting to anyone wishing to understand how the different forms are related to one another, if at all. As the number of good quality images grew in the first part of the 20th century, different classification schemes were naturally attempted. An excellent review of the steps which led to the recognition of the main types is provided by Sandage 45, who gives all of the early references as well as references to a number of classification systems which fell into early dis-use. Examples of most of the main types were already known by 1920.
The classification systems in use today are all in some way related to that described in Hubble's 34 paper. This system, which originally included only ellipticals, spirals, and irregulars, focused on a few basic characteristics and ordered galaxies in a manner that was eventually found to correlate with some basic measured parameters. The sequence was best defined for spirals since three classification criteria were available: the relative strength of the bulge, the openness of the arms, and the degree of resolution of the arms. Van den Bergh 62 commented that the firm establishment and later addition of the SO class by Hubble 35 destroyed the "simple beauty" of the original system. It is interesting also that in spite of a great deal of recent research, there have been no firm correlations found between ellipticity and other properties of E galaxies, leaving the value of this criterion as part of the "sequence" in doubt (e.g., Tremaine 59). The later division of irregulars into two subclasses, "Irr I" and "Irr II", was proposed by Holmberg 33.
Hubble's final revision to his system is illustrated and described by Sandage 44. No other classification system has ever been so beautifully illustrated. This was complemented recently by the Atlas of Galaxies Useful for the Cosmological Distance Scale by Sandage and Bedke46, and by the published "mini-atlases" of Dressler and Sandage 27, Sandage and Brucato 48, Sandage, Binggeli, and Tammann 50 and the Revised ShapleyAmes Catalogue 49 (=RSA).
De Vaucouleurs 18, 19, 20 presented a personal revision of Hubble's system that provides a better description of what a galaxy looks like without being too complicated. The system uses the concept of a classification volume, rather than a simple multi-pronged "tuning fork", and is recognized mainly for its addition of stages later than Sc, called Sd and Sm, and for the notation (SA, SAB, and SB) used to denote continuity of the bar characteristic. These revisions to Hubble's system have been largely accepted by most astronomers. The Sm class is particularly important because it recognized the Magellanic Clouds not simply as "irregulars" but as very late spirals with no spheroidal component 25, 43.
The classification system proposed by Morgan 40 (see also Morgan, Kayser, and White 41) was designed as a means of tying galaxy morphology to the then current ideas of stellar populations. The degree of central concentration was used to define a one-dimensional spectral classification system (population group) based on form alone. Secondary dimensions, defined by the "form family" and the tilt index, tied the system in an indirect way to Hubble's system. Morgan also introduced some types not recognized fully by Hubble. The best known of these was the cD class, although its discovery is claimed by Vorontsov-Velyaminov 67. The Morgan system has recently been used in a study of the spiral-to-elliptical galaxy ratio in two nearby galaxy clusters 70. However, it has not been used in any recent major catalogues.
One limitation of Hubble's system was recognized by van den Bergh 60, 61, who demonstrated the existence of luminosity effects on the contrast and development of spiral arms. He assigned not only modified Hubble types to galaxies, but also luminosity classes symbolized in a manner similar to stellar luminosity classes. The luminosity classes were originally applied to Sb, Sc, and Irr galaxies, but the latter objects, lacking spiral structure, had to have their luminosity classes estimated from surface brightness alone. Van den Bergh's modified Hubble system was later drastically changed; he disagreed with Hubble's placement of SO's in the "transition region" between ellipticals and spirals for reasons related to flattening and bulge-to-disk ratio. He instead proposed placing SO's in a sequence parallel to spirals 62 (RDDO system). The two sequences use bulge to disk ratio as a classification criterion, and he identified transition cases between SO's and normal spirals which appeared to be spirals with little star formation in the arms. These were given the term "anemics", and were assumed to be gas-poor.
Because the RDDO system builds around the expected evolutionary scenario in clusters that stripping can deplete spirals of the gas needed for ongoing star formation, nurture is explicit: galaxies which may once have been spirals evolved to a completely different type called SO's owing to an interaction with the cluster environment. SO's may not be born, but are made by this interaction. This is a controversy that has not yet been resolved. A primary problem, noted by Burstein 9, is that there are as yet no known examples of SOc galaxies, i.e., galaxies of the SO type which have a bulge-to-disk ratio as small as those seen in many Sc galaxies. Another problem, noted by van den Bergh 64, is that the SO class is a mixed bag of possibly unrelated types of objects.
Vorontsov-Velyaminov 67, referring to his work on the Morphological Catalogue of Galaxies (MCG) in the 1960's, has held steadfastly to the view that galaxy morphology is too complicated to be represented adequately by any of the available Hubble or Hubblelike systems. He developed in the MCG a purely descriptive classification with symbols geared to almost every detail of morphology. This makes for a complex symbolism but is still useful because it allows the isolation of specific categories of objects that the broad Hubble classes do not adequately represent.