5. RECENT DEVELOPMENTS IN CLASSIFICATION AND MORPHOLOGY
Recent developments in this field have been concerned with (1) new major catalogues of standard Hubble types; (2) slight modifications to existing classification systems; (3) the identification of new types of galaxies and improved understanding of older types; (4) identification of the major orbit resonances in grand design spirals and ringed galaxies; (5) widespread application of electronic detectors with high quantum efficiency to large numbers of galaxies of many types, and over a wide range of passbands; and (6) computer classification.
Major catalogues with morphological data continue to be produced. These are summarized in reviews by Corwin 17 and Buta 13. The RC3 combined several of these catalogues and other smaller lists into the largest database of Hubble morphological type information ever compiled. The 17,700 types given in this catalogue are based solely on photographic sources and are on the classification system of de Vaucouleurs 19. Other major catalogues include the Virgo Cluster Catalogue 5 and the RSA; these give types on Hubble's revised system 44, with additions and revisions, some described below.
No major new classification systems have been proposed since 1976, though modifications to existing systems have been suggested. For example, Kormendy 38 suggested that tenses be distinguished from rings using their own notation within the framework of the de Vaucouleurs revised Hubble system. He suggested denoting inner lenses by (1) and outer lenses by (L) in the same classification positions where inner and outer rings would be specified. Kormendy also suggested a different approach to morphology, the idea of characterizing galaxies in terms of a small number of "distinct components" (bars, rings, etc.) rather than a large number of morphological "cells". The objective of the approach is to make deductions concerning secular evolution from the ways these components might be expected to interact (see Table 1 of Kormendy 39). Kormendy suggested that such an approach leads to the possible conclusion that bars are not permanent features of galaxies but may evolve under certain circumstances to a lens. Whether this evolution actually takes place or not is still uncertain.
Another revision to the classification systems is the recognition of "dusty E's." The misclassification of these objects as S0's is a noteworthy problem of catalogues emphasized by Ebneter, Djorgovski, and Davis 28 (and references therein; =EDD). The presence of such "features" in a type of galaxy which was by definition featureless led EDD to suggest a more physical classification of E's is now warranted. A lovely montage and catalogue of dust-lane ellipticals is provided by Bertola 3.
Sandage and Brucato 48 pointed out that the original classes called Irr I and Irr II in the Hubble Atlas are not satisfactory because they combine widely differing objects into the same bin, namely "Irr". To distinguish galaxies which are not E, S0, or S but which have an amorphous appearance to the unresolved light, sometimes with imbedded resolved stars, they proposed the term "amorphous" galaxies. Sandage and Brucato emphasize that these objects are similar to, but not precisely like, the Irr II's in the Hubble Atlas, and that some similar objects classified as I0 by de Vaucouleurs may be peculiar spirals or S0's. One of the hallmarks of the amorphous class is a well-developed early-type absorption spectrum spread throughout the disk.
In the case of spirals, many aspects of the "grand design" and "flocculent" spiral morphologies have now been quantified 31. These are aspects of spiral structure morphology that are not directly built into Hubble classifications. Flocculent galaxies lack bimodal symmetry and have a spiral-like structure composed only of small pieces of arms. Grand design galaxies generally have a two-armed structure and the arms are longer and more continuous than in flocculent galaxies. To account for these differences and for combinations of the two pattern types in many galaxies, Elmegreen and Elmegreen 30 proposed a system of 10-12 "arm classes" or AC's to highlight a systematic orderliness of spiral arms. The AC's are not exactly the same as van den Bergh luminosity classes because they emphasize symmetry and arm length, rather than arm contrast.
The identification of the locations of specific dynamical orbital resonances in spiral galaxies has seen much progress in recent years. Research has focussed on two classes of objects: grand design spirals by the Elmegreens, and ringed galaxies by myself. The paper by Elmegreen and Elmegreen 29 summarizes how to recognize the primary orbit resonances in a relatively typical grand design spiral, NGC 1566. For this purpose, purely morphological methods guided by expectations from spiral structure theory are used. The features considered are spiral arm kinks, gaps, spurs, bifurcations, endpoints to star formation ridges, dust-lane crossover points, interarm star formation, and the ends of a weak bar. If consistency can be found between the positions of these features and those inferred for specific resonances from a rotation curve, then the pattern speed of the wave can be derived with reasonable confidence. However, even in an extreme grand design case like NGC 1566, the resonance features are very weak. It takes a great deal of tenacity, for example, for the reader to study and identify clearly all of the features summarized in Table 1 of Elmegreen and Elmegreen 29.
Ringed galaxies refer to normal galaxies classified in the de Vaucouleurs revised Hubble system with the symbols (R)SB(r), (R')SB(rs), (R')SAB(s), (R)SAB(r,nr) etc., that is, objects which have inner, outer, or nuclear rings or pseudorings. These rings are believed to define the locations of specific orbital resonances with a bar or oval, and if correct they are much more obvious optical features with a direct link to resonances than some of the features seen in the best grand design spirals. Thus, they provide a promising way of indirectly estimating pattern speeds of bars and ovals, of which very little is known. At the moment, there is a great deal of evidence that the outer rings and pseudorings of SB and SAB galaxies trace the location of the outer Lindblad resonance, or OLR. This follows from statistics of their shapes and orientations with respect to bars 10, from their relative sizes with respect to inner rings 2 and most of all from their morphology 15. The Catalogue of Southern Ringed Galaxies12 is designed specifically to understand the link between rings and resonances, and has been the basis for the studies of Buta 10 and Buta and Crocker 15.
A number of interesting findings have been made concerning cluster galaxies. A photometric study of brightest cluster members, or "BCM's", including gE, D, and cD types (Schombert 51, 52, 53) has led to a refined and quantitative classification of these galaxies based on luminosity profile shapes. Schombert has noted that the characteristic extended envelopes of cD galaxies are generally fainter than 10% of the night sky brightness and are not readily seen on PSS prints. Thus, the rather shallow luminosity profiles of cD's is what led to their recognition, in addition to their central location in clusters. It is the existence of a true extended envelope that distinguishes the cD from the D class.
As emphasized by Sandage and Binggeli 47 (=SB), the Virgo Cluster contains galaxies of virtually every known morphological type. Of particular interest has been the identification in Virgo of dwarf S0, or dS0 galaxies, which morphologically are like S0's but which are of considerably lower luminosity and surface brightness than normal S0's (see also Binggeli and Cameron 4). Most of the galaxies in Virgo fainter than B 14 appear to be dwarf E, or dE, systems. SB emphasize that the "great void" in luminosity below Sa, Sb, and Sc types is real - there are no convincing cases of dSa, for example. This confirms that the Hubble sequence is largely defined by giant galaxies. However, although no examples of dSa or dSb were found in Virgo, a promising example was found by van den Bergh 63 in the compact, apparent elliptical galaxy NGC 3928, a member of the Ursa Major Cloud of galaxies.
The luminosity class system of van den Bergh has been extended to classes V-VI and VI by Corwin (see introduction to RC3) to allow for a greater range of apparent surface brightnesses seen among dwarf and late-type galaxies on the SRC-J sky survey. The RSA luminosity class system was also refined by SB to allow for a greater apparent range of surface brightnesses seen among Im galaxies in the Virgo Cluster. Among the galaxies classified as Im V by SB are "huge" Im types having significant diameters (up to 10 kpc) and peak central surface brightnesses less than 10% of night sky in blue light. These are accompanied by similar huge dE systems. The data from a variety of sources of luminosity classes have been compared and combined in RC3 24.
Van den Bergh, Pierce, and Tully 65 (=BPT) have discussed the classification of 231 Virgo Cluster galaxies from CCD images. They propose a revision to the classification system of van den Bergh 62 to include Sd and Sm types, and demonstrate that the accuracy of luminosity classification is improved on digital images ((MTB) 0.7 mag) compared to classifications based on photographic plates published in the RSA ((MTB) 1 mag). Of particular interest in this work was the identification of a possible new class of galaxies, called "Virgo types." These galaxies have fuzzy outer regions and active star formation in their bulges or inner disk regions, and constitute 43% of 88 Virgo cluster spirals. In contrast, BPT find that the Ursa Major cluster includes only 2 "Virgo types" out of 35 spirals, suggesting real differences. BPT suggest that the early "Virgo types" represent a mild form of the Butcher-Oemler effect that persists at zero redshift.
In a study of the HI and optical properties of cluster galaxies, Bothun, Schommer, and Sullivan 6 identified a class of red, HI-rich, low surface brightness spirals. A sample of these objects is compiled by Schommer and Bothun 54, and two extreme examples of the class, NGC 3883 and UGC 542, were studied by van der Hulst et al 66. The types of these galaxies range from Sa to Sc in Schommer and Bothun 54, and NGC 3883 is quite distinctive for its size and appearance in Abell 1367. Van der Hulst et al 66 interpret these galaxies in terms of a threshold HI surface density for star formation and possible interrupted star formation activity or an altered IMF.
An important serendipitous finding from a study of a field of the Virgo Cluster was an object dubbed "Malin 1" 7. This galaxy appears small enough on PSS prints that it did not make inclusion into the UGC 42. However, on amplified deep IIIa-J plates, Malin 1 shows an extended, low surface brightness disk surrounding a small bright core. The object is not a member of the Virgo Cluster (it is 20 times as distant) and is now recognized as a new class of giant, HI rich, low surface brightness disk galaxies the likes of which had not been appreciated before. The properties of Malin 1 are further summarized and described by Impey and Bothun 36, and a second example of the class was reported by Bothun et al 8. These objects are now interpreted as disk galaxies whose HI surface density is so low that they evolve only slowly.
The study of interacting galaxies has led to the recognition of several new morphologies. Polar ring galaxies 69 are believed to be cases where a small satellite has been disrupted into a polar orbit around an S0. Hoag-type ringed galaxies 55 may be related cases where the central object is an EO system. X-galaxies are edge-on S0's displaying a distinct X-shape across the center that may also be related to polar ring galaxies 68. "Ocular" galaxies are interacting galaxies displaying an "oval-apex" structure resembling an eye 32. The latter objects are particularly interesting, because they represent a type of bar not distinguished within the Hubble system. A key feature is a double arm on one side, as illustrated in Figure 1 of Elmegreen et al 32.
Of particular interest to students of spiral structure is the discovery of a leading spiral arm in the interacting galaxy NGC 462216. The arm was first noticed by Byrd on a well-known commercially available photograph published in Shu 56. The galaxy is of type SA(r)ab and shows two major outer arms that wind clockwise, but inside the inner ring a single arm winding in the opposite sense is present. Since the "discovery" photograph was taken in blue light, Buta, Crocker, and Byrd 15 (=BCB) re-observed the galaxy in the Cousins I-band to test whether the arm is stellar or an artifact of dust. The leading arm was found to be a clear feature in the galaxy's old disk population. The fact that only a single leading arm is observed in this case, rather than two, is strong evidence that the arm was generated by a tidal interaction, as discussed in detail by BCB.
The widespread use of high quality CCD's, especially the large format TEK CCD's at KPNO and CTIO, has greatly increased the number of large-scale images available for classifying galaxies. What is particularly important is that a typical modern CCD can provide in a short amount of time images that are deeper in limiting surface brightness than the SRC-J sky survey, and yet still provide detailed information on the central regions of galaxies. Thus, they bypass the main problems of direct prime focus or Schmidt plates and have the potential of adding greatly to our knowledge of morphology. It is also clear that recent advances in infrared detectors make the development of a classification system in the 1-3 µ wavelength range a real possibility. The advantages of using near infrared images to type galaxies are their increased sensitivity to the dominant old stellar populations, which tends to enhance the visibility of features such as bars and bulges. The young component of galaxies which dominates blue light images for many spirals as well as dust will be less prominent and therefore not important for typing purposes. The number of "cells" required to classify galaxies should therefore decrease somewhat. However, going to the infrared will not change the pitch angle of spiral arms or the relative sense of the Hubble sequence. What is clear is that the number of "non-barred" galaxies will probably decrease, as can be gathered just the ESO-B and ESO=R sky survey charts.
Finally, in the future some catalogues of galaxies will probably include automatic classification 1, 58. This is an approach still under development, owing to the difficulty of defining some aspects of morphology, but once a satisfactory methodology is achieved, it has the potential of providing more consistent classifications than might be achieved visually.
Gérard de Vaucouleurs developed an extension of the Hubble classification of galaxy types involving a "three-dimensional" classification grid extending from ellipticals at one end to irregulars at the other. For those and lenticular galaxies, de Vaucouleurs' classifications are somewhat similar to Hubble's classifications, but for spiral galaxies distinctions are made between normal and ringed, barred and unbarred, and transitional types, and even the simpler structures have various subclassifications, leading to a more detailed "shorthand" description. Various methods are commonly used to illustrate the system. One is a drawing of the three-dimensional grid, showing artistic impressions of the appearance of various galaxies (an example of such a grid will be posted here if I can find one that is in the public domain). Another method is the publication of a catalog showing several hundred photographs of actual galaxies as examples of specific types (current editions include examples used by de Vaucouleurs, and numerous supplementary examples). When complete this page will include thumbnail images of every galaxy in the online de Vaucoulers Atlas (with a link to the appropriate page in that Atlas), and the NGC/IC/PGC designation of each galaxy linked to its entry on my website. A quick perusal of the images below shows that this page is in its early stages, but I intend to work at it on a regular basis and it should be complete by the end of the summer (the linked pages on my site may take a little longer, but are also a top priority).
Note: The obvious order for the table/images below would be according to the galaxy classification; but some galaxies defy classification, and for that and presumably other reasons the order of the plates in de Vaucouleurs' Atlas is not strictly in order of the classification. But rather than second-guess him and those who have updated and added to his original work I have left the images in the order presented on the external site.