A most important feature of galaxies is the great difference in stellar content between their spiral arms and the spheroidal component. The difference is evident from direct photographs: it was shown explicitly by Baade's resolution studies in 1944, and is apparent from the change of integrated colors along the classification sequence as the relative importance of the old (spheroidal) and young arm populations exchange dominance in going from S0 to Sm systems.
Because the difference is pronounced, it must be expected that the integrated spectral type should correlate with the light ratio of the spheroidal to young-disk populations, and therefore roughly with Hubble type, since this ratio is one of the classification parameters.
The earliest indication of such population differences was noted by Seares (1916a, b), who discovered that the central parts of galaxies are redder than the arm regions. After the start of the Mount Wilson and Lick systematic redshift programs, both Humason and Mayall noted gross spectroscopic differences among galaxies, and summarized their results in the discussion of their redshift catalog (HMS 1956). Starting in 1932, Humason regularly classified the dominant spectral class and showed that it correlated with Hubble type in the expected sense.
Morgan and Mayall (1957) reviewed these early results and advanced the work by showing that the composite spectral class correlates well with the degree of dominance of the spheroidal component alone, i.e., with the concentration of the luminosity toward the center. (5) Morgan (1958, 1959) developed this spectral-type concentration correlation into a classification system whose color-class notation explicitly isolates the spectral type that is expected on the basis of the correlations.
The Morgan system contains information on the state of stellar evolution in the central regions of galaxies. The classification complements information in the Hubble-de Vaucouleurs system which, in cases of conflicting criteria, emphasizes more strongly the strength of the spiral-arm population [see, e.g., NGC 4941, Sab: Hubble; SABab: de Vaucouleurs; Atlas, p. 10].
In the introduction to his 1958 paper, Morgan states: ``The correspondence between form and spectral appearance is especially close for two categories of galaxies: (1) irregular systems of the Magellanic Cloud-type and spirals having an insignificant central concentration of luminosity [Sd, Sm], and (2) giant ellipticals such as those found in the Virgo cloud and spiral systems such as M31, in which the major share of the luminosity of the main body is due to a bright amorphous central region. In the first category, early-type stars and emission nebulosities have a profound effect on the spectrum in the blue and violet regions [producing strong hydrogen absorption lines and bright emission lines such as found in galactic HII regions); in the second, the luminosity of the brighter parts is due principally to yellow giant stars.''
``Systems possessing an intermediate degree of central concentration of light (M51 [Plate 4 here on NGC 5194]) give spectroscopic evidence of an intermediate kind of stellar population: the degree of compositeness is very high, and it appears likely that most of the luminosity is due to a mixture of F to G main-sequence stars and K giants.''
Morgan's notation contains a concentration class ranging from a through af, f, fg, g, gk, and k, in the direction of increasing domination of the nuclear light (and consequently later composite spectral type according to the Morgan-Mayall correlation), and a form notation similar to Hubble's with E for ellipticals; S, spirals; B, barred spirals; and I for irregulars. Three new form classes are added: D for dustless systems dominated by the amorphous light; L for galaxies of low surface brightness (such as NGC 45 (Atlas, p. 37]); and N for systems having small, brilliant nuclei superposed on a considerably fainter background.
The addition of an inclination index ranging from 1 (circular) to 7
(spindle with a/b 
10) and the
symbol p denoting ``peculiar'' completes
the notation. Examples of well-known galaxies classified on the Morgan
system, many of which are shown in
plates 2 -
7, are NGC 5273 (gkD2
Morgan; S0/Sa Hubble), NGC 488 (kS2 Morgan; Sb Hubble), NGC 628 (fgS1
Morgan; Sc Hubble), NGC 5204 (fI-fS4 Morgan; Sc/Ir Hubble; SAm de
Vaucouleurs); M33 (fS3 Morgan; Sc Hubble) and NGC 4449 (aI Morgan; Ir
Hubble; IBm de Vaucouleurs).
There is a general correlation of the Morgan class with the Hubble-de Vaucouleurs types. A comparison has been made by de Vaucouleurs (1963b, Table 13). The major difference is in the definition of the D systems. The closest Hubble type is S0, although there is a rather large extension into the E's on one side of the scatter and the early SA and SB types on the other. The difference is of some importance because Morgan classifies the brightest galaxies in rich clusters as cD (c for supergiant), whereas the Mount Wilson observers do not.
Following Matthews, Morgan, and Schmidt (1964), the cD notation has been generally adopted by radio astronomers for the radio ellipticals, but these galaxies are still classified as giant E galaxies on the Mount Wilson system. The Mount Wilson procedure follows naturally from the discovery of Hubble and Humason in the 1930s that rich clusters of galaxies are composed almost entirely of E and S0 galaxies. The brightest member galaxies are of the same general type as the fainter very numerous E systems (restricting to the top 4-mag range), differing only slightly if at all in the occasional presence of a extended envelope (it is not yet known if these outer regions follow a power intensity law as in true E galaxies, or if there is a small exponential component as in the disks of true S0's).
The introduction of a separate D class in an otherwise continuous sequence of E forms among the brighter cluster members is then the chief difference between the Mount Wilson and the Yerkes classification of galaxies in rich clusters.
A summary of the Morgan system and its significance for studies of stellar content is given by Morgan and Osterbrock (1969).
5 The difference in scatter between the
Morgan-Mayall and the Humason
spectral-type, galaxy-class correlation shows clearly that the size of
the nuclear region does not vary uniquely and monotonically with
Hubble type, although there is a general trend. The point has
previously been made by Vorontsov-Velyaminov in his comments on the
Hubble system (cf. Introduction, Vol. 2, Morphological Catalog of
Galaxies, p. 4) and by
Freeman (1970)
in his plot of the
spheroidal-to-disk length ratio as a function of Hubble type.
Back.