I. CLASSIFICATION
3. Revised classification. In
the course of a survey of bright
southern galaxies with the 30-inch Reynolds reflector at Mount Stromlo
[80] a classification and
notation system has been developed to
include in a consistent scheme all or most of the recent revisions and
additions to the standard classification. The classification follows
as closely as possible the revised Mt. Wilson-Palomar scheme, but
includes the additional types and sub-types suggested by the Harvard,
Lick and Mt. Stromlo work; the notation is based on the original Mt.
Wilson system, but has been modified and supplemented as required by
the introduction of new types and sub-types and so as to permit a
finer description of structural features. Classification and notation
are illustrated in Fig. 2 which may be
regarded as an extension of Hubble's tuning-fork diagram.
Fig. 3. A 3-dimensional representation of
the revised
classification scheme and notation system. From left to right are the
four main classes: ellipticals E, lenticulars S0,
spirals S and irregulars I. Above are the ordinary
families SA, below the barred families SB; on the
near side are the S-shaped varieties S (s), on the far side
the ringed varieties S (r). The shape of the volume indicates
that the separation between the various sequences SA (s), SA
(r), SB (r), SB (s) is greatest at the transition
stage S0/a between lenticulars and spirals and vanishes at
E and Im. A central cross-section of the classification
volume illustrates the relative location of the main types and the
notation system. There is a continuous transition of mixed types between
the main families and varieties across the classification volume and
between stages along each sequence; each point in the classification
volume represents potentially a possible combination of morphological
characteristics. For classification purposes this infinite continuum of
types is represented by a finite number of discrete ``cells''. Compare
with Plates I and II.
Actually
Fig. 2 may be considered as a
plane projection of a three
dimensional representation, the idea of which was first conceived
during conversations with Dr. Sandage in August 1955. It is
schematically illustrated in
Fig. 3 which may be consulted in
conjunction with Fig. 2 to follow
the detailed description of the
adopted classification and notation (compare also with Plates I and
II). The three-dimensional classification appears necessary to
represent objects of mixed characteristics in correct relation to the
main sequences as well as the progressive divergence of the various
sequences from E to S0/a through the S0 stages and
their ulterior convergence from S0/a to I through the
S stages.
a) Isolated galaxies.
The present classification scheme and
notation system rest on the following principles:
Four classes are retained: ellipticals
E, lenticulars S0, spirals S, irregulars I,
which coincide with the main divisions introduced by Hubble.
The two families of lenticulars and
spirals, originally denoted ``normal'' S and ``barred'' SB
by Hubble, are now designated ``ordinary'' SA and ``barred''
SB, so as to permit the use of the compound symbol SAB
for ``intermediate'' objects of mixed characteristics. The symbol S
alone is used when a spiral object cannot be more accurately
classified as either SA or SB because of poor resolution,
unfavorable tilt, etc. The ``ordinary'' spirals are not more ``normal''
than those of the ``barred'' family which are at least as common
(cf. Sect. 3c).
In the new class of lenticulars the
two families are denoted SA0 and SB0, depending on the
absence or presence of a bar structure across the central lens;
``intermediate'' objects with a very weak bar
are noted SAB0. The symbol S0, used for SA0 in
Hubble's notation, is now used for a lenticular object which cannot be
more precisely classified as either SA0 or SB0; this is
often the case for edgewise objects 1.
Two main varieties are recognized in
each of the lenticular and spiral families, the ``annular'' or
``ringed'' type, denoted (r), and the ``spiral'' or
``S-shaped'' type, denoted (s). Intermediate types are noted
(rs). In the ``ringed'' variety the structure includes circular
(sometimes elliptical) arcs or rings (S0) or consists of spiral arms
or branches emerging tangentially from an inner circular ring (S). In
the ``spiral'' variety two main arms start at right angles from a
globular or little elongated nucleus (SA) or from an axial bar
(SB).
The distinction between the two
families A and B and between the
two varieties (r) and (s) is most clearly marked at the
transition stage S0/a between the S0 and S
classes. It vanishes at the transition stage between E and
S0 on the one hand, and at the transition stage between S
and I on the other (cf. Fig. 3).
Four sub-divisions or
stages are distinguished along each of the
four spiral sequences SA (r), SA (s), SB (r),
SB (s), viz. ``early'',
``intermediate'' and ``late`` denoted a, b, c as
in the standard classification, with the addition of a ``very late''
stage, denoted d. Intermediate stages are noted Sab,
Sbc, Scd. The transition stage towards the magellanic
irregulars (whether barred or not) is noted Sm, e.g. the
Large Magellanic Cloud is SB (s) m.
Along each of the non-spiral sequences
the signs + and - are used to denote ``early'' and ``late''
subdivisions; thus E+ denotes a ``late'' E,
the first stage of the transition towards the S0 class
2. In both the SA0 and SB0 sub-classes three
stages, noted S0 -, S0 °,
S0 + are thus
distinguished; the transition stage between S0 and Sa,
noted S0/a by Hubble, may also be noted
Sa-. Notations such as Sa+,
Sb-, etc. may
be used occasionally in the spiral sequences, but the distinction is
so slight between, say, Sa+ and Sb-,
that for statistical purposes it is convenient to group them together as
Sab, etc. Experience shows that this makes the transition
subdivisions, Sab, Sbc, etc. as wide as the main
sub-divisions, Sa, Sb, etc. 3.
The classification of ``irregulars'' is
still somewhat in doubt; objects obviously related to the magellanic
type, but which do not show clearly the characteristic spiral structure,
are noted I (m); those with an elongated core and asymmetrical
branches are probably later stages of the SB sequences, e.g.
NGC 4449, NGC 6822, etc., while those more nearly symmetrical
and without bar-like core are probably later stages of the SA
sequences, e.g. IC 1613, IC 2574, etc. (see Plate VII) the distinction
vanishes in the
ultimate, chaotic dwarf stages of low surface brightness, e.g. the
Wolf-Lundmark nebula or the Sextans system 4.
A faint, outer ring-like structure is
often observed in lenticulars and early-type spirals which is not
clearly related to a specific type, i.e. with some variants it appears
about equally in all four sequences near the transition stage
S0/a. This particularity which seems therefore more
characteristic of a certain stage of evolution than of any definite line
of evolution, is denoted by an (R) preceding the symbol of
the class. Examples of (R) SA are NGC 1068
(M77), NGC 4736 (M94), NGC 7217, etc. of (R) SB:
NGC 1291, NGC 1326, NGC 2859, etc. Some objects have both an outer
(R) structure and an inner (r) pattern; a good example
is NGC 6753, noted (R) SA (r) ab (22).
In the spiral sequences supplementary
data of some interest are the multiplicity of the spiral pattern
and the character, ``massive'' or ``filamentary'' of the spiral
arms. These are included whenever possible as subscripts respectively
after the family and variety symbols. Thus M33 is noted
SA (S) 2 cm and described as ``an ordinary,
late-type, S-shaped spiral with two main massive arms''. When one or
more additional and weaker arms are present this is noted 2 + 1
(e.g. M99), 2 + 2 (e.g. M51), etc. An asymmetrical, regular spiral with
one arm stronger than the other as is often the case among late-type
barred spirals (e.g. NGC 7479) is noted 1 + 1 (not 2). Branching arms
not clearly assignable to a definite multiplicity are noted
1+, 2+, etc. as the case may be; this is a
frequent occurrence among late-type spirals. A complicated spiral
pattern of multiplicity higher than 4, as often observed in late-type
spirals, is simply noted n (e.g. NGC 2903, NGC 7793); this,
however, occurs also among early-type spirals of the ringed variety
(e.g. NGC 4736, NGC 7217).
Edgewise systems which can
usually be classified only as E7, S0,
S (a, b, c, d) or I are noted (sp) for
``spindle''. For ellipticals the traditional notation E0,
E1,...E7 may be preserved, although it is
not homogeneous with the rest of the classification scheme and
notation system used for the other classes; it is really almost
superfluous when measured dimensions and flattening are given.
An advantage of the present notation
system is that it is possible to retain significant information on
objects poorly resolved by dropping one or more symbols which from right
to left refer to stages,varieties, families and classes of increasing
generality. Thus S may be either SA or SB,
S0 either SA0 or SB0, SA is either
SA (r) or SA (s),..., Sa is either SAa or
SBa,..., etc. Hence the nature and number of symbols used can be
selected according to the amount of detail visible in any given object
with any particular telescope. Conversely the degree of completeness in
the symbolism gives an indication of the relative degree of resolution
available; this, of course, depends on both the size of the telescope
and the size of the nebula.
b) Pairs and multiplets.
These are discussed in detail by
Zwicky. It is nevertheless of some importance for the present
discussion to determine the separation
between neighboring objects within which their mutual interaction
begins to cause significant structural distortion and thus sets a
limit to the validity of the classification system established in the
first instance for isolated galaxies. A rough classification of pairs
is as follows:
- P (a) : pairs of widely separated objects showing no clear
sign of interaction (occasional optical pairs are included in this group);
- P (b) : close pairs showing clear signs of interaction,
such as tidal distortion of their outlying parts;
- P (c) : colliding pairs where the ``main bodies'' are in
actual contact and show extensive disruption of their internal structure.
Many parameters are obviously involved,
such as separation, diameters, masses, velocities, etc. in the
interaction of two galaxies. Only the first two are more less directly
accessible in most cases and more precisely the ratio x' of the
projected separations s'1,2 between the nuclei
G1, G2 of a pair to the sum of their
apparent photographic radii r1, r2. The
true separation s is not known in individual cases, but for a
random distribution of orientations of mean (G1
G2) with respect to the line of sight
the mean value of s' / s is
/ 4
0.79. For 40 pairs south of - 35° observed with the Reynolds reflector
on Mt. Stromlo 5 the following mean values were obtained:
P (a) mean x' = 3.65 (range: 1.6 - 9:)
n = 30,
P (b) mean x' = 1.30 (range: 0.6 - 1.6)
n = 8,
P (c) mean x' = 0.3: (range: 0.2: - 0.4:)
n = 2.
Allowing for scale factors the average
projected and actual separations are:
P (a) mean s' = 18 kpc (range 8 to 45 kpc)
mean s = 23 kpc,
P (b) mean s' = 6.5 kpc (range 3 to 8 kpc)
mean s = 8 kpc
on a distance scale in which Hubble's expansion parameter is 180
km/sec per Megaparsec. The mean critical separation (x'
1.6, x
2.0) within which tidal distortion becomes notable is then of the
order of 10 kpc. Among the bright galaxies south of -35° only 10 out
of more than 400, or less than 2.5%, are within this limit. Hence, the
revised classification should be applicable to at least 97 or 98% of
the external galaxies, of which about 95% can be included in the
normal scheme of Fig. 2 and the
remaining 2 or 3% must still be
regarded as special or ``peculiar'' cases reserved for further analysis.
c) Frequency of revised types.
A classification of over 200 southern
galaxies in the revised system
[80] yields the following apparent
relative frequencies for a
sample essentially selected according to apparent magnitude and
surface brightness (all objects listed in the Shapley-Ames Catalogue
south of - 35°).
The regular spirals SA, SB
make up just over fifty percent of the sample, ellipticals and
lenticulars close to a quarter each. The
discrepancies between earlier statistics (cf. Sect. 1b) may therefore
have hinged mainly on whether the lenticulars were counted as
ellipticals or spirals the ratio S/E changing correspondingly
from 1 / 1 to 3 / 1. Note also that the ordinary spirals appear no
longer more abundant than the barred spirals despite the
inclusion of intermediate types (SAB) with the SA group in
the table.
Table 1. Frequency of
revised types
Class/Family
| E
| S0
| SA
| SB
| I
| Pec
|
Frequency (%). .
| 23.4
| 21.0
| 24.4
| 26.3
| 3.4
| 1.5
|
The detailed frequency in each of the
sub-divisions along the regular classification sequence is given in
Table 2, in which E / S0 = E
+, S0 - and S0/a = S0 +,
Sa -.
Table 2. Frequency of
sub-divisions
Type
| E
| E/S0
| S0
| S0/a
| Sa
| Sab
| Sb
| Sbc
| Sc
| Scd
| Sd
| Sm
|
Frequency
| 22.0
| 9.5
| 10.5
| 9.0
| 4.5
| 6.5
| 7.5
| 7.5
| 10.5
| 8.5
| 2.0
| 2.0
|
A plot of the frequencies
(Fig. 4)
gives a fairly smooth curve indicating maxima at Sc and E
separated by a minimum at Sa. The low
apparent frequency of the Sd, Sm stages is certainly due
to a large extent to the low absolute luminosity and surface brightness of
galaxies at these stages, but it is not yet possible to correct the
apparent frequencies for such selection effects. On the other hand
the high frequency of ellipticals cannot be due in general to an
abnormally high absolute luminosity, so that it probably reflects the
actual great abundance of this class 6.
Fig. 4. Frequency distribution of revised types in the Mount
Stromlo survey.
1 The symbol (S) may be used when even the distinction
between S and S0 is no longer possible, as in the case of
very small, poorly resolved objects.
2 An example is NGC 3115 which, although selected as the type
object for E7 in Hubble's original scheme, shows an absorption
marking and a beginning of nuclear differentiation on 200 inch plates.
3 No E-, nor I- or
I+ are at present clearly recognized, but such
notations may later prove useful for a still finer classification based
on accurate surface photometry.
4 Objects of low surface brightness (when it is not obviously
due to local obscuration, e.g. as for IC 10, IC 342, etc.) are almost
invariably dwarf systems: this particularity may be noted by a (d)
preceding the symbol of the class which is usually either E or
I. The existence of dwarf lenticulars and spirals remains to be
demonstrated. Criteria are still lacking for the assignment of objects
to the giant and supergiant groups on purely morphological grounds.
5 The photographic diameters used in the Mt. Stromlo survey
[80] are corrected for tilt
effects, i.e. reduced to ``face-on'',
and define a system of dimensions intermediate between those of
Reinmuth [62] and of Shapley
[66].
6 It should be emphasized that these provisional data based on a
rather small sample of 200 objects are still subject to further
revision; more definite and detailed results will eventually be
forthcoming through the current reclassification of 1250 bright
galaxies in the Shapley-Ames Catalogue.