2. Revisions and additions.
About 1935 Hubble undertook a systematic morphological study of the
approximately 1000 brighter galaxies listed in the Shapley-Ames
Catalogue, north of - 30°
declination, with a view to refining his original classification
scheme. This work begun with the 100-inch reflector and later
continued with the 200-inch reflector was practically completed at the
time of Hubble's death in 1953; his notes collected and organized by
Sandage will be published in the near future with a collection of
about 175 photographs illustrating the morphological features
characterizing the various galaxy types. Through the courtesy of Dr.
A. R. Sandage the main revisions introduced by Hubble to his system
can be briefly described here.
a) The most important addition was the
introduction of the S0 and
SB0 types regarded as transition stages between the ellipticals and
spirals at the branching off point of the tuning-fork.
S0 objects have the smooth
appearance of ellipticals, but a
luminosity distribution more like that of spirals, although no
spiral arms are visible. They are characterized by a sharp, bright
nucleus in the center of a more or less uniform disc or ``lens'' having
a rather sharp outer rim and surrounded by a faint, diffuse ``envelope''
with indefinite boundaries; the diameter of the lens in which dark
crescents or rings of obscuring matter are often observed is usually
about one-third of that of the envelope (see Plate III).
There is apparently a continuous
transition from ``late'' ellipticals, such as NGC 4270, NGC 4958, etc. and
``early'' S0's such as NGC 3630, NGC 4564, etc., then through
various stages of the S0 sequence
to the ``earliest'' regular spirals.
HUBBLE distinguished two groups of
- S0 (1): smooth lens and envelope; early examples are NGC 1201,
NGC 1332, and late examples NGC 3065, NGC 4684.
- S0 (2): Some structure in the envelope in the form of a dark zone
and ring; examples are NGC 4459 and NGC 4111, the latter seen edgewise.
In edgewise objects the presence of
rings manifests itself by the
appearance of ``ansae'' simulating Saturn's ring, such as in NGC 4215,
NGC 7332, etc. (Plate IX). The transition stage
S0 and Sa shows incipient spiral structure in the envelope.
SB0 objects are characterized by a bar
through the central lens,
sometimes broad and hazy, sometimes narrow and sharp; the envelope may
form faint outer rings, sometimes conspicuous, sometimes vague or
imperceptible (Plate IV). Hubble distinguished three groups of SB0
- SB0 (1): a bright lens, with broad, hazy bar and no ring,
surrounded by a larger, fainter envelope, e.g. as in NGC 3384, NGC 4262,
etc., some have circular envelopes, e.g. NGC 4203.
- SB0 (2): a broad, weak bar across a primary ring, with faint
outer secondary rings, e.g. NGC 2859.
- SBO (3): well developed bar and ring pattern, with the bar
stronger than ring, e.g. NGC 4643, NGC 5101.
b) Many of the objects originally
classified as SBa were then reclassified as SB0 and the
SBa sub-division had to be redefined as follows:
SBa: smooth bar and lens, with poorly
developed, closely coiled arms in envelope and either massive and
structureless or filamentary and partially resolved.
Hubble further distinguished two or
three groups of barred spirals, one in which the arms extend from the
rim of a ring crossed by a bar, e.g. NGC 2217, NGC 5566, NGC 5701
(SBa), NGC 4999, NGC 5950 (SBb), and one in which the arms
start at the ends of the bar without
ring, e.g. NGC 2798, NGC 4290, NGC 7743 (SBa), NGC 1300,
NGC 5430, NGC 6951 (SBb). In another group still the
ring is formed of closely coiled
filamentary arms, e.g. NGC 3185, NGC 4037, NGC 4385, NGC 4389. This
distinction remains well marked at the SBb stage, showing well
developed, partially resolved arms going through more than one
revolution, but there is no spiral structure in the lens; primary
rings consist of spiral arms, secondary rings are seldom found. It
was not followed up into the SBc sub-division, characterized by wide
open, well resolved spiral arms with absorption lanes and spiral
structure in the lens, but no ring pattern.
Sandage has also recognized two types
of normal spirals, one in
which the arms start at the rim of a ring structure, the other in
which they start from a central nucleus. Some objects of the ringed
type had previously been described by Shapley and
Paraskevopoulos 1 as
``plate spirals'' and an investigation of ring structures was made by
There are further differences in the
multiplicity of the spiral
pattern, some objects showing only two main regular arms, others
having a great many tightly coiled whorls. Reynolds had already
pointed out 2 that some spirals have ``massive'' arms,
e.g. M31, M33,
while others have ``filamentary'' arms, e.g. M81, M101. The importance
of this distinction was acknowledged by Hubble
[B] but could not be
incorporated in the classification.
c) An extension of the normal spiral
sequence beyond the stage Sc
was proposed by Shapley who used the notation Sd for objects such as
NGC 7793 (20), showing a very small, bright nucleus
knotty irregular spiral arms. This notation could also, and perhaps
more appropriately, be applied to highly disorganized and complex
spirals of low surface brightness, such as NGC 4395-4401 (see Plate
A parallel extension of the barred
spiral sequence beyond the
stage SBc was introduced by de Vaucouleurs
3 through the recognition
of spiral structure in the Magellanic Clouds and objects of similar
type, such as NGC 1313 (32), NGC 4027, NGC 4618, etc., which
may be noted as SBd or SBm.
Still ``later'' stages extending the
spiral sequence into the irregular types may be represented by objects
such as IC 2574 or
NGC 2366, IC 4662, etc. These objects are characterized by
brightness, high degree of resolution and sometimes outstanding
emission nebulosities similar to 30 Doradus in the Large Cloud. They
always show an abundance of blue supergiants and strong emission lines
in their spectrum.
An important characteristic of the
group of irregulars related to
the Magellanic type I(m) is their small diameter and low luminosity
which marks them as dwarf galaxies. Typical objects of this group are
NGC 6822, IC 1613, the Sextans system, the
Wolf-Lundmark-Melotte nebula, etc.4
In the poorest and smallest of them
emission objects may be few or
absent, but when their distance is small enough they are always well
resolved into blue supergiants and giants.
d) The existence of dwarf ellipticals
(dE) of very low surface
brightness was first brought to notice in 1938 through the discovery
of the Fornax and Sculptor systems by Shapley 5. These are close
enough to be resolved on blue plates at m 18; except for their very
low density they seem to share all other population characteristics of
normal ellipticals. Because of their very low surface brightness such
systems are difficult to detect and might be much more abundant in
space than their belated discovery suggests; nevertheless a special
search for them on Harvard plates failed to disclose other examples. A
few more ``Sculptor type'' systems have, however, been found since 1949
with the 48-inch Palomar Schmidt 6.
Dwarf galaxies of a possibly related
type have been investigated
in 1952 by Reaves  on
plates of the Virgo cluster taken by C.
D. Shane with the Lick 20-inch astrograph. These objects of which the
brightest example in the Virgo cluster is IC 3475 have the same smooth,
circular or little elongated appearance as the dwarf ellipticals and
likewise show little central condensation. Hubble and Reaves
suggested that they may be related to a group of dwarf spirals of low
surface brightness exemplified by NGC 3299 because their color was at
first thought to be bluish; however, recent observations indicate
them to be reddish and this strengthens the similarity with the
e) After all such additional types or
variants have been weaned out,
there remains a hard core of ``irregular'' or ``peculiar'' objects which
do not seem to fit in any of the recognized types. One group of such
objects consists of strongly interacting or colliding systems such as
NGC 1275, NGC
4038-39, NGC 5128, etc.7 which can often be
their distorted structure, abnormal spectrum and strong
radio-emission; these systems are discussed in the article by B. Y.
Mills. Double or multiple systems showing moderate interaction in the
form of connecting links, distorted spiral arms or irregular
filamentary extensions are discussed by F. Zwicky. Only isolated
galaxies or weakly interacting systems are considered in the present
Even among them, however, there remains
a small number of objects
not clearly assignable to any of the previous types; one possible
group, exemplified by NGC 3034, NGC 3077, etc. is characterized by an
early-type spectrum (A, F) contrasting with a reddish color
irregular absorption patches and filaments and a smooth,
unresolved nebulous structure indicating an absence of blue
supergiants and of discrete emission nebulosities (see Plate
XI). Another possible group, exemplified by NGC 5253 8 is
characterized by fairly strong emission lines and a complex structure
unlike that of the Magellanic irregulars or late-type spirals. Still
other puzzling objects such as Mayall's nebula 9, also show
strong emission lines but have a fairly smooth structure 10.
Fig.2. Revised classification: a plane
projection of the
classification volume. Compare with
Fig.3. The ordinary spirals
SA are in the upper half of the figure, the barred spirals
SB in the lower half. The ring types (r) are to the left,
the spiral types (s) to the right. Ellipticals and lenticulars
are near the center, magellanic irregulars near the rim. The main stages
of the classification sequence from E to Im through
S0 -, S0, S0 +, Sa,
Sb, Sc, Sd, Sm are illustrated,
approximately on the same scale, along each of the four main
morphological series SA (r), SA (s), SB (s), SB
(r). A few mixed or ``intermediate'' types SAB and S
(rs) are shown along the horizontal and vertical diameters
respectively. Structures predominantly of Type I population are dashed,
of Type II dotted. This classification scheme, as used in the Mount
Stromlo survey of southern galaxies, is superseded by the slightly
revised and improved system illustrated in
Fig. 3 and Plates I to XI.
1 H. Shapley and J.S. Paraskevopoulos:
Proc. Nat. Acad. Sci. U.S.A. 26, 31-36 (1940) = Harvard Rep. 184.
2 J. H. Reynolds: Observatory 50, 185 - 189 (1927);
with comments by Hubble, Observatory 50, 276-281 and by Reynolds,
Observatory 50, 308.
3 G. de Vaucouleurs: Observatory 74, 23-31 (1954). -
Astronom. J. 60, 126-140, 219-230 (1955).
4 See F. Hubble: Astrophys. Journ. 62, 409-433 (1925)
= M.W.C. 304. - W. Baade: Astronom. Nachr. 234, 407 (1929). -
F. Zwicky: Phys. Rev. 58, 478 (1940). - K. Lundmark:
V.J.S. 68, 382 (1933) = Lund. Medd. (I), Nr. 135.
5 H. Shapley: Bull. Harvard Coil. Obs. 1938, No. 908.
6 R. G. Harrington and A. G. Wilson: Proc. Astr. Soc. Pacific
62, 118-120 (1950). - A. G. Wilson: Proc. Astr. Soc. Pacific
7 See W. Baade and R. Minkowski:
Astrophys. Journ. 119, 215-231 (1954). - F. Zwicky:
Ergebn. exakt. Naturw. 29, 344-385(1956).
8 D.S. Evans: Observatory 72, 164-166 (1952).
9 R. T. Smith: Proc. Astr. Soc. Pacific 53, 1S7
(1941); also Astrophys. Journ. 119, 225 (Fig. 14) (1954).
10 Some early-type spirals are known to show abnormally wide
emission lines in the spectra of their nuclei, although the origin of
such high random velocities ( 5000 km/sec) is not known, this need
not be considered as an indication of a distinct type, but merely of a