2. Structure

2.1 Morphological distinction

Dwarf ellipticals span a range of at least 10⁴ in luminosity (from M_B -18 to -8) along a sequence of increasing mean surface brightness with increasing luminosity (see Sect. 2.2.2). At the bright end of this sequence, the relatively high surface brightness of a dE can mimic a normal elliptical, and quantitative analysis of surface brightness profiles is necessary to attempt a dE versus E distinction. This has proven difficult (see Sect 2.2.1). From a purely morphological (classificatory) point of view, there will always be a certain number of bright ``intermediate'' (E/dE) types, irrespective of a possible discontinuity in various measurable properties between E's and dE's (Binggeli and Cameron 1991; Prugniel 1994; Vader and Chaboyer 1994).

A similar problem exists for the dE versus Irr distinction at the faint end; in fact, for the whole range M_B < -16. Dwarf irregulars in this range can appear very smooth and dE-like - presumably when they happen to be ``sleeping'', i.e. at a low or zero star formation rate. In the Virgo cluster, there is a broad dE/Irr class of galaxies comprising roughly 10% of the whole dwarf population, with increasing percentage faintwards (Sandage and Binggeli 1984; Sandage et al. 1985b). It has become clear over the years that this is not just a problem of classification (as it might be in the case of E versus dE): there appears to be a continuum of intrinsic properties such as gas content, metallicity, and star formation rate (Sect. 4) among dwarf galaxies. There are truly intermediate types which are probably in a transitional stage from Irr to dE (Sect. 7.6). A prototype dE/Irr in our neighborhood is the Phoenix system (van de Rydt et al. 1991). The Andromeda satellites NGC 205 and 185, too, are well-known ``peculiar'' dE's that contain dust and gas (e.g. Hodge 1971). Other, more distant examples of ``mixed morphology'' have been discussed by Sandage and Hoffman (1991) and Sandage and Fomalont (1993).

The following features are also relevant for the dE morphology (cf. Sandage and Binggeli 1984):

(1) Nuclei. Most bright dwarfs (M_B < -16) show a distinct luminosity spike in their center, commonly referred to as the central nucleus. These nuclei are not (yet) resolved at the distance of the Virgo cluster, i.e. they have a stellar appearance, but local resolved analogs, such as the nuclei of NGC 205 or M33, suggest that they are dynamically separate supermassive star clusters. The brightest nuclei can reach up to 20% of the total light of the parent dwarf galaxy (see Fig. 2) . The ratio of nucleated-to-normal dE's is monotonically decreasing with decreasing luminosity; faint dwarfs usually do not have a nucleus (Sandage et al. 1985b). The presence of a nucleus is conveniently indicated by appending an ``N'' to the type: dE,N.

Figure 2. Surface brightness profiles of five dwarf ellipticals with prominent nuclei. The solid lines represent the sum of a King profile to the outer regions (with parameters rc, µ0, and log(rt / rc)), and a central point source (convolved with the approppriate PSF) fit to the inner regions such that the observed surface brightness is nowhere exceeded. The percentage of the total light contributed by the central light excess (nucleus) with respect to the King fit is given under the heading fex. From Vader and Chaboyer (1994).

(2) Dwarf S0 types. Sandage and Binggeli (1984) introduced the dwarf S0 type, which is a rare variation of the dE class. In their azimuthally averaged surface brightness distribution the dS0's are indistinguishable from bright dE' s (Binggeli and Cameron 1991). Some dS0's do show a pronounced two-component structure that is reminiscent of classical S0's, but there are a variety of other reasons why a dwarf was (and may be) called dS0 rather than dE (e.g. the presence of a bar feature, twisting isophotes, or simply high apparent flattening; see Binggeli and Cameron (1991). The dS0 class is very inhomogeneous and since it is so small (there are only about 25 dS0's in the Virgo cluster, as compared to 800 dE's) it will be mixed into the dE class, if not otherwise stated.

(3) Huge, low-surface brightness types. This is another class of galaxies first isolated in the Virgo cluster by Sandage and Binggeli (1984): very extended systems of extremely low surface brightness and almost no gradient, mostly classified as dE, but often also as intermediate (dE/Irr) or even irregular. These galaxies fall off the canonical dE sequence (see Sect. 2.2.2). The most extreme examples in the Virgo cluster have been found by Impey et al. (1988) using a special photographic technique developed by D. Malin. Similar galaxies exist in the Fornax cluster (Bothun et al. 1991). None are known in our local neighborhood.

Typical dwarf ellipticals, as well as some related types, are shown in Fig. 1a-n.

Figure 1. A collection of dwarf galaxy members of the Virgo cluster, adapted from Binggeli (1994a). The common scale is indicated on top. Typical dwarf ellipticals in the left row (c, f, i, l) along with the two ``dS0'' variants (a, d) are confronted with a low-luminosity, compact E (b), and with various types of smooth (g, h) or clumpy (k, m) dwarf irregulars. Two intermediate (transitional) types are also shown (e, n), the latter of which is a ``huge, low-surface brightness'' type. The individual names and types are as follows: a = NGC 4431 (dS0(5),N), b = NGC 4486B (E1), c = IC 3328 (dE1,N), d = IC 3435 (dS0(8),N), e = NGC 4344 (S pec, N:/BCD), f = IC 3457 (dE4,N), g = IC 3416 (ImIII), h = UGC 7636 (ImIII-IV), i = VCC 1661 (dE0,N), k = IC 3453 (ImIII/BCD), l = VCC 354 (dE0), m = VCC 1313 (BCD), n = IC 3475 (ImIV or dE2 pec)