3.3 Hubble's Major Modification Between 1936 and 1950
The S0 class was not isolated observationally until after 1936, although Hubble had earlier come to believe such a class was necessary. In The Realm of the Nebulae he wrote, ``The junction [between E and S types] may be represented by the more or less hypothetical class S0. Observations suggest a smooth transition between E7 and SBa [on the barred side of the tuning-fork diagram], but indicate a discontinuity between E7 and Sa (on the spiral side] in the sense that Sa spirals are always found with arms fully developed [whereas SBa in the 1936 system had no arms, by definition].'' The difference, then, was primarily a matter of definition (cf. plate II of Realm), but it did represent an asymmetry between the ordinary and barred spiral families because no ``armless ordinary spirals'' had been found whereas armless barred spirals did exist (cf. NGC 2859, Hubble Atlas, p. 42; NGC 2950, p. 42; and NGC 4643, p. 42, all originally classed as SBa but now classed as SBO).
Because of this asymmetry, Hubble undertook a special search for nonbarred examples of ``armless spirals'' (the modern S0 class). They were subsequently found on long-exposure plates taken in the Mount Wilson survey of the Shapley-Ames galaxies north of = - 15°, carried out by Hubble between 1936 and 1950 with the 60- and 100-inch reflectors.
Elliptical and S0 galaxies are easily confused on small-scale plates, which explains why the S0 class remained unrecognized for so long. Neither class has arms, and both show smooth intensity distributions with no resolution into bright supergiant stars. The criterion that distinguishes the two classes is a difference in the radial intensity distribution I(r). E galaxies have a steep intensity gradient (Hubble 1930, de Vaucouleurs 1948, 1953, 1959a), whereas S0 galaxies have an extensive exponential outer envelope, superposed on a central E-like distribution (cf. de Vaucouleurs 1956, 1959a; Liller 1960; Johnson 1961; Hodge and Webb 1964; Hodge and Merchant 1966). This envelope is similar to the underlying exponential disks of spirals (cf. de Vaucouleurs 1958, 1959b, 1962, 1963a, 1964; Freeman 1970).
Although it is clear that Hubble discovered the S0 class and studied some of its members, he published no discussion of the important phenomenon. The first literature references were made by Spitzer and Baade (1951), based on Baade's conversations with Hubble. Summaries were given by de Vaucouleurs (1956, 1959a) based on prepublication notes from the Hubble Atlas, and by Sandage (1961) where the class is discussed and illustrated.
A description of Hubble's revised system as it existed in 1950, written by Hubble between 1947 and 1950, is useful, and is reproduced from the Hubble Atlas:
The sequence of classification, as originally presented, consisted of a series of elliptical nebulae ranging from globular (E0) to lenticular (E7) forms, and two parallel series of unwinding spirals, normal (S) and barred (SB). Each of the latter series was subdivided into three sections, termed early, intermediate, and late, and designated by the letters a, b, and c, respectively. Thus the early, normal spirals were represented by the symbol Sa, and the early barred spirals by SBa.
The data available in 1936 seemed to indicate a smooth and continuous transition from elliptical nebulae to barred spirals, and, in fact, the first section of the latter series, SBa, exhibited no spiral arms. The corresponding section of the normal series, Sa, contained so many nebulae with fully developed spiral arms, that, where arms could not be definitely recognized, their presence was assumed, and the failure to detect them was attributed to effects of orientation or other causes. The procedure was unsatisfactory because it introduced subdivisions in the parallel series of spirals that were clearly out of step. Moreover, the transition from E7 to Sa appeared so abrupt that, if real, it might be regarded as cataclysmic.
With accumulating data, and especially with the increasing number of good photographs with the 100-inch reflector, the situation has clarified. Numerous systems are now recognized which are later than E7 but which show neither bars nor spiral structure. These nebulae fill the supposed gap between E7 and Sa and remove the excuse for postulating a cataclysmic transition. [These transition galaxies are designated S0. They are actually found in nature and are no longer a hypothetical class, as was once believed; see The Realm of the Nebulae, pages 45-46, and the legend to figure 1, page 45, of the Yale University Press edition of 1936. A.S.]
A similar group of objects corresponds to the section [called SBa in the 1936 classification]. This situation emphasizes the desirability of redefining the sections of both series in a more comparable manner.
The revision might be made in various ways but only that actually adopted will be described. First, two new types, S0 and SB0, have been introduced to include objects later than E7 but with no trace of spiral structure. Second, the series of true spirals, as before, are subdivided into the three sections Sa, Sb, Sc, and SBa, SBb, SBc. In the case of the normal spirals, the change amounts to a subdivision of the former section Sa into the two sections S0 and Sa. In the case of the barred spirals the entire former section, SBa, is now termed SB0, and the former section SBb is subdivided into the two sections SBa and SBb. The revisions are summarized in the following table:
Old Class New Class Old Class New Class Sa SBa SBO SBb Sb Sb Sc Sc SBc SBc
The introduction of the new types leads to a revision of the original assignment of symbols. The original SBa nebulae are now described as SB0, and the original SBb nebulae are redistributed between SBa and SBb. Among the normal spirals, the Sa objects are redistributed between S0 and Sa. Otherwise the system remains unchanged.
The transition stages, S0 and SB0, are firmly established. In both sequences, the nebulae may be described as systems definitely later than E7 but showing no spiral structure. The next stages, Sa and SBa, are represented by nebulae which show incipient spiral structure. Fully developed spirals are distributed over the two later stages of each sequence according to the relative extent of the unresolved central region, and the degree to which the arms are resolved and unwound.