Next Contents Previous

3.2 The Early Hubble System

From these data, supplemented by plates taken by Hubble, Humason, and Duncan with the Mount Wilson 60- and 100-inch reflectors, and building on the earlier proposals (Section 2), Hubble (1926) formulated the linear classification system which has evolved into the present standard systems. Galaxies are divided into three major classes (ellipticals, spirals, and irregulars). The spirals are separated into two families, the ``normal'' spirals and the barred; and are separated along each family into types a, b, and c according to the three criteria listed in Section 1, which generally (but not always) vary together along the form sequence from early Sa (SBa) to late Sc (SBc). The system has been described in many places (e.g., Hubble 1926, 1936; de Vaucouleurs 1959a; Sandage 1961; Baade 1963, chap. 2; Hodge 1966) and need not be repeated here.

Soon after publication of Hubble's 1926 paper, Reynolds (1927a, b) criticized the system because of a supposed inadequate number of classification bins. Impressed by the enormous variety of galaxian structures, Reynolds wrote: ``The problem I have always found in attempting a general classification of spiral nebulae is that one meets case after case where a special class is required for the individual object. Spectral classification of stars is a simple and straightforward matter compared with this.'' (2)

Wolf had earlier commented similarly in the first sentence of his 1908 paper ``Es gibt kein zwei Nebelflecken am Himmel, die sich gleichen.'' But it is precisely because Hubble's system specifically ignores the multitude of superficial details of arm structure (for example), and concentrates on the gross characteristics of pattern according to broad criteria, that the system has such merit. Nearly all galaxies can be put into a specific classification bin (in the revised system) without forcing, because the bins are large. This is true even for most of the interacting and peculiar galaxies discussed by Vorontsov-Velyaminov (1959), Arp (1966), Burbidge, Burbidge, and Hoyle (1963), and others. The underlying Hubble type is usually visible, albeit with peculiar features denoted by pec or p after the basic type. Baade's comment (1963, chap. 2) that the Hubble system has great merit is shared by most classifiers, but not all would agree with his belief that the logical extensions by de Vaucouleurs are unnecessary-extensions which retain the simplicity of the scheme, but narrow the class sizes and give notational recognition to the transition cases between ordinary and barred spirals.

Genuinely peculiar objects do exist, such as M82, NGC 3077, NGC 520, NGC 2685, NGC 3718 (illustrated in the Hubble Atlas 1961). These fall outside the system, but constitute only a few percent of any random sample.

An example of the universality of Hubble's system is given by a trial classification of galaxies in the Atlas of Peculiar Galaxies (Arp 1966). Of the 338 illustrated galaxies, 45 are so abnormal that they could be given no underlying basic Hubble type (except peculiar), but this is only 13 percent of the total sample, which itself is already highly selected as regards peculiarity.


2 Although Reynolds criticized Hubble's 1926 system as too ``simple,'' he had himself, seven years earlier, proposed a classification (Reynolds 1920b) that was similar enough to Hubble's E, Sa, Sb, Sc, and Irr types (but called classes I-V), that, in the absence of his 1927 repudiation, Reynolds would now have been considered as an early originator of part of the modern classification. Back.

Next Contents Previous