3. EXTENSIONS AND ADDITIONS

Over the past 10-15 years, the imaging capabilities for galaxies have been greatly improved, especially with the advent of CCD's. Not only this, but the great Sky Surveys produced with Schmidt telescopes since the 1950's have greatly increased the number of galaxies for which detailed morphology can be clearly studied. This exposive growth in the number of available images has in recent years led to great curiosity about the apparently secondary and more subtle aspects of galaxy structure, and to a few possible revisions to the classification systems. Some of these are described here.

a. Rings and Pseudo-Rings

It is well-known that in addition to the inner rings used for the variety dimension in the RHS, two other types of ring are occasionally found: nuclear rings, sometimes seen in the nuclei of barred galaxies, and outer rings, occasionally seen enveloping the main bodies of early-type spirals and lenticulars. Although these features may not be as widely present or as easily discernible as family, variety, and stage, their characteristics and their physical interest make it worthwhile to recognize them with their own type symbols. In the revised Hubble classification notation, outer rings and pseudo-rings do have type symbols, (R) and (R'), respectively, to be used preceding the main part of the classification. However, for nuclear rings and pseudo-rings, there is currently no notation, and the first possible extension or addition to the RHS that I propose is the use of (nr) and (nr') for such structures. Note that this classification is not entirely straightforward: some nuclear rings appear to be pseudo-rings only because of dust (e.g. NGC 1512, Lindblad and Jorsater 1981; NGC 4314, Benedict 1980). Table 1 illustrates the possible use of this notation and some examples.

The second possible extension or addition to the RHS is perhaps more subtle. In a recent paper (Buta 1986a), I have suggested a possible refinement to the classification of outer rings which is motivated by the recent theoretical work of Schwarz (1979, 1981). From the appearance of over 300 outer rings and pseudo-rings in SB and SAB galaxies on the SRC IIIa-J Southern Sky Survey films, it appears that at least two distinct subtypes exist: the first type (R'₁) shows the characteristics that the arms which emerge from each end of the bar seem to wind 180° and return to the opposite side, dipping in to form a broad figure eight pattern (Figure 2a, NGC 3504); in the second type (R'₂), the arms go an extra 90° before intersecting each other, and no figure eight characteristic is made (Figure 2c, ESO 294-16). Between these extremes, intermediate cases are found which appear to show aspects of both types (Figure 2b, ESO 509-98).

Figure 2. New Outer Pseudo-Ring Morphologies and Notation.

These distinctive morphologies were discovered by Schwarz (1979) in simulations of barred galaxy gas dynamics, and it is a small triumph for the theory that they were found in real galaxies after they had been predicted to exist. The predictions came at an especially opportune time because the availability of the deep IIIa-J SRC Southern Sky Survey films made it possible to detect outer rings and pseudo-rings in galaxies that otherwise would have been very difficult to study on the ESO Quick-Blue Survey or the Palomar Sky Survey plates. For reasons described in section 4, the mixed morphology (R₁R'₂) was not predicted to exist, but its apparent existence in real galaxies underscores once again the "fuzziness" characteristic of galaxy morphological cells.

The relative frequencies of the different ring patterns have not been firmly established. De Vaucouleurs and Buta (1980) selected carefully defined samples from the Second Reference Catalogue of Bright Galaxies (de Vaucouleurs et al. 1976, hereafter RC2) to judge the frequencies of inner rings. These occur in about 25% of spirals, but if inner pseudo-rings are considered, the total frequency would exceed 50%. In a study exclusively dealing with barred spirals, Kormendy (1979) found that 76% of early-to-intermediate barred spirals possess inner rings. Outer rings appear to be most frequent near stage S0/a (de Vaucouleurs 1975). The tentative results of Buta (1986a) suggested that the R'₁ subtype is perhaps three times as frequent as the R'₂ subtype, although only 55% of recognizable outer pseudo-rings could be categorized in either way. No definitive statistics exist yet for nuclear rings, whose detection is strongly influenced by selection effects.

b. Lenses

A lens is a relatively uniform or shallow zone in the luminosity distribution of disk galaxies which, on direct photographs, appears to have a fairly sharp edge. Several are illustrated in the Hubble Atlas. They were first recognized by Hubble in S0 galaxies, but were never formally given a type symbol. Instead, they were regarded as characteristic of S0 galaxies in general. They also do not have a type symbol in de Vaucouleurs's revision, although de Vaucouleurs has proposed using the code "L" for S0 galaxies, to imply "lenticular," and has presented information on lens sizes in the notes to the Reference Catalogue of Bright Galaxies (de Vaucouleurs and de Vaucouleurs 1964). Kormendy (1979, 1982) has done detailed studies of the frequency and other properties of lenses, and believes them to be an important "distinct component" of galaxies whose true frequency cannot be judged from RC2 types. He has criticized RC2 types because he believes these misclassify lenses as inner rings. He has also pointed out that even though all S0 galaxies might be denoted "L," many do not have lenses.

These disagreements reflect some of the subjectivity of galaxy classification and I wish to make two points about them. First, since the RC2 does not give quantitative information on the strength of luminosity enhancements associated with rings, one cannot determine whether a ring is strong or weak from RC2 types. However, if a ring is very weak, then in less than optimum seeing the enhancement may be washed out and one will see a "plateau" (or lens) in the light distribution. The distinction between calling a galaxy "inner-ringed" or "inner-lensed" then reduces to semantics in such circumstances, unless rings and lenses are genuinely distinct phenomena. A second point is that rings tend to be blue in color and fairly narrow, while lenses are red and broad. When both are present in the same galaxy, they overlap and have the same shape, implying a close causal connection.

This suggests at least two interpretations: either rings and lenses are variations on the same phenomenon, differing only in the mean age of the stellar population and the amplitude of the enhancement which is superposed on the usual decreasing background of the bulge plus disk; or they are distinct phenomena which happen to coincide in position and shape. The first interpretation would view lenses as being a direct result of ring formation, in the sense that the lens would represent the aging component of a ring whose star formation rate has been fairly uniform over a Hubble time, such that a more diffuse, broader component has been built up. This view finds some support in the integrated colors of rings (Buta 1986b, 1987a, 1988b), the existence of non-axisymmetric color gradients (Buta 1986b), the existence of lens analogues of each type of ring (de Vaucouleurs 1974; Kormendy 1979; Sandage and Brucato 1979; Buta 1986b), and the existence of rings lacking an old component (e.g. NGC 1398, Buta 1988b) or a young component (Gallagher and Wirth 1980). How diffuse an old ring will be will most likely depend on its age and velocity dispersion. Only for several lenses have velocity dispersions so far been measured, so that a definitive comparison cannot yet be made. As shown by Kormendy (1984), lenses tend to be fairly hot.

However, the nature of lenses is probably more complicated than might be inferred from the above arguments, because lenses usually contain a lot of light and at the same time have oval intrinsic shapes which make them characteristically bar-like. It is quite interesting how some galaxies show inner, outer, and nuclear rings just like those in SB galaxies, yet no obvious bar crosses the inner ring. In such a circumstance, we might conclude that the lens cannot be simply an "old" ring, but is the near-axisymmetric analogue of a bar; it is therefore distinct from rings, but capable of appearing to be associated with them in the same manner as an SB-type bar. Several interpretations have been proposed in this vein. Kormendy (1979, 1981, 1982) believes that lenses originate secularly from bar dissolution in SB galaxies. Although bars are believed to be fairly robust (Miller and Smith 1979b), Contopoulos (1983) has stated that strong bars can generate enough stochasticity inside corotation that stars on stochastic orbits will populate a less elongated, oval zone. On the other hand, Athanassoula (1983) believes that lenses represent disk instabilities just like conventional SB-type bars, the main difference concerning the degree of random motions.

Regardless of the physical interpretation of lenses, I agree with Kormendy (1979) that perhaps they should be given their own type symbols in the classification. He has proposed the use of (l) for inner lenses and (L) for outer lenses, with (nl) an obvious extension for nuclear lenses.

The relative frequencies of lenses are as poorly known as those of rings. Statistics for inner lenses (l) are most secure. Kormendy (1979) found using a carefully defined sample that inner lenses are present in 54% of early-type SB galaxies. The frequencies of nuclear and outer lenses remain to be determined.