Next Contents Previous


The Hubble classification of galaxies is a remarkable achievement; although it was devised at a time before Baade's great work on stellar populations, its basic value has survived through the spectacular developments of the past quarter-century; and we are still able to perceive quite clearly the genius of its formulator. The progression from elliptical systems at one extreme to little-concentrated spirals at the other represents, without any question, an arrangement according to some physical property or properties.

However, examination of the forms of the brightest galaxies showed Hubble - as it shows us today - that these bodies present themselves to us in an exceedingly wide variety of appearance; and it is not possible to describe uniquely the forms of all galaxies by the types of the original Hubble system. This fundamental complexity has led Sandage, de Vaucouleurs, and others to devise elaborated modifications of the Hubble system; the work of Sandage, in particular, in collating and expanding fragmentary material left by Hubble is of basic importance.

A few years ago, an attack on the classification problem of galaxies was begun by Morgan and Mayall (1957) from a somewhat different point of view: a classification of galaxies according to their spectroscopic characteristics. Their work was extended later by Morgan, to the devising of a new form classification which makes use of the spectroscopic results; this classification depends, in effect, on one of the criteria used by Hubble: the degree of central concentration of luminosity. This parameter is especially well correlated with the spectra, in the case of galaxies of the three brightest magnitudes of luminosity. In accordance with a suggestion first made by Shapley some years ago, the order of the progression of types has been reversed.

The most sensitive region of galaxies for the discrimination of stellar population characteristics is the bright, inner part. The change from B, A, and F spectral characteristics to those of class K is accompanied by a systematic change in the appearance of the nuclear region. The amorphous central region of spirals similar to M31 can be considered typical of the K-type population encountered in giant elliptical galaxies. The outer, fainter regions of M31 are systematically earlier in spectral type; this is well shown in unpublished spectrograms obtained by H. W. Babcock at the Lick Observatory. It is also shown, in the case of our Galaxy, by McDonald spectrograms of transparent and semiobscured regions in the galactic nuclear bulge; the hydrogen lines are considerably stronger - and the spectral type is earlier - in regions where obscuring clouds are present.

During the past year a careful comparison of new spectroscopic material with galactic forms has resulted in the development of a highly simplified modification of the system developed by Morgan; this simplified system has certain advantages in interpretation over the more complicated Yerkes system; however, the latter is still to be preferred from the point of view of the amount of information included.

In the simplified Yerkes system, galaxies are classified into seven groups, according to the degree of central concentration of luminosity. The simplified system is illustrated in Figure 1.

Figure 1

Figure 1. Simplified Yerkes classification system (Y') for galaxies. Diagram by J. W. Tapscott.

The (S) and (B) sequences resemble those of the Hubble system; however, there are considerable divergences in the case of individual galaxies because of lack of correlation between the appearance of the arm structure and the nuclear region. Since the latter appears to be the most sensitive criterion of the stellar population of the inner parts of a galaxy, the abscissa in Figure 1 is closely correlated with the spectral type of the inner region.

Figure 1 can be divided into three general areas: (1) Certain highly irregular galaxies lacking symmetrical nuclear concentration of light; the spectra of galaxies in this category show that such systems are exceedingly rich in hot stars and gas. (2) An area including galaxies of the strongest central concentration of luminosity; these consist of the giant ellipticals, elliptical-like objects in which the luminosity distribution differs from that of pure ellipticals, and barred systems in which effectively all the luminosity is concentrated in the bar itself and an amorphous ring surrounding it; the spectroscopic evidence indicates that galaxies in this area owe most of their light in the blue and violet regions to yellow giant stars. (3) An intermediate area between the two extremes; here are located the systems of intermediate concentration of luminosity; the spectroscopic evidence indicates that these are the most highly composite systems of all.

Two differing sequences are illustrated for the spirals of stronger concentration of luminosity; in one case, (S), a relatively small nucleus grows in size and develops brilliant, amorphous characteristics; in the other, (S'), the nuclear region remains more nearly the same size and tends to be bounded by a ringlike structure. The (S') sequence begins at concentration class 4.

The question of continuity along these two spiral sequences is of interest; that giant elliptical galaxies may be considered as an end-point for the (S) sequence is suggested by certain instances of galaxies classified as ellipticals by Hubble which have been found recently to be surrounded by armlike segments. NGC 474, which is illustrated in Figure 2, can be considered to be a type example of this intermediate class. The illustration is from the blue print of the National Geographic Society-Palomar Observatory Sky Survey. I am indebted to Dr. I. S. Bowen for permission to reproduce this plate.

Figure 2

Figure 2. Reproduction of NGC 474 (at right) from blue print of Palomar Sky Atlas. Copyright National Geographic Society - Palomar Observatory Sky Survey.

In the case of the (S') sequence, there is some evidence that galaxies similar to the illustrated example at E' may be considered an end-point, similar to the elliptical end-point for the (S) sequence.

Figure 1 has certain incompletely two-dimensional characteristics; there exists a continuous array of galaxies intermediate in appearance between the (S) and (B) sequences. A continuous range of forms also exists between types E and E'. The S' sequence appears to be more nearly one-dimensional and in some respects should be considered as lying outside the principal plane of Figure 1. However, even in this case, further investigation may modify the apparent discreteness of the S' sequence.

The new spectroscopic material on which the classification outlined in Figure 1 rests is given in Table 1. For the forty galaxies included, McDonald spectrograms were used exclusively for the revised spectral types. The columns give (1) the NGC number, (2) the revised MK spectral type as determined in the blue-violet region, (3) the form class on the Yerkes system (Y), and (4) the simplified form class on the revised (Y') system. The latter is stripped down to the bare essentials of a concentration class and a form-sequence indicator; for the E and E' systems an inclination class has been retained.

(McDonald Slit Spectrograms)

Galaxy MK Y Y'

NGC 224 K gkS5 6S
NGC 891 G g?S7 5S
NGC 1398 K kB2 6B
NGC 2775 GK kDS4 6S
NGC 2841 K kS4 6S
NGC 3031 K gkS4 6S
NGC 3115 K kD7 E'7
NGC 3368 GK gkS4p 6S
NGC 3379 K kED3 E3
NGC 3384 K kD5 E'5
NGC 3489 GK kSD5 6S
NGC 3626 G kDS5 6S
NGC 3627 G gS4 5S
NGC 4111 GK kD7 E'7
NGC 4214 OB aI 1I
NGC 4254 FG fgS1 4S
NGC 4258 G gS5 5S
NGC 4303 FG fS1 3S
NGC 4321 FG fgS1 4S
NGC 4374 K kE1 E1
NGC 4406 K kE2 E2
NGC 4429 K kD5 E'5
NGC 4435 K kDE6 E'6
NGC 4438 K gkI 6I
NGC 4449 BA aI 1I
NGC 4459 K kE3 E3
NGC 4472 K kE2 E2
NGC 4473 K kE5 E5
NGC 4477 K kED1 E1
NGC 4486 K kE1p E1p
NGC 4490 A aI 1I
NGC 4501 GK gS4 5S
NGC 4526 K kD6 E'6
NGC 4594 K kDS6 6S
NGC 4636 K kE1 E1
NGC 4649 K kE2 E2
NGC 4762 K kD7 E'7
NGC 4889 K kE4 E4
NGC 5194 F fgS1 4S
NGC 7814 K kD7 E'7

NOTE: Some of the spectrograms used in Table 1 were obtained jointly with N. U. Mayall.

The relationship between the numbered concentration classes and those expressed by small letters in the Yerkes system is given in Table 2.

1 2 3 4 5 6 7
Y class aI aS fS fgS gS gkS kE
aIp afS fB fgB gB kS kD
  aB   gkB (kB)
  afB   (kB)  

The principal changes are as follows: the (S) and (B) sequences begin at concentration class 2; and concentration class 7 consists entirely of E, E', and B' systems. In the Y' system, concentration class 1 is composed entirely of irregular galaxies similar to NGC 4214, 4449, and others of highly irregular form.

The drastic simplification represented by the Y' system entails serious disadvantages in at least one respect: a considerable percentage of all galaxies brighter than the thirteenth magnitude cannot be satisfactorily classified on the simplified system, and for this reason it is of the greatest importance to retain in a complementary fashion the additional information furnished by the Y system.

We may summarize the general physical significance of the classification illustrated in Figure 1 as follows: concentration classes 1 and 7 represent two end-points, cleanly separated with regard to stellar population: class 1 is characterized by high percentages of hot stars and gas; class 7 is characterized by yellow giant stars and little or no observable gas.

Next Contents Previous