Next Contents Previous


The distribution and clustering properties of galaxies in clusters are known to depend strongly on the Hubble type (Dressler 1980). The Virgo cluster is no exception of this. Fig. 3 shows the projected distribution of Virgo cluster members devided into the main morphological types: giant early types (E + S0), dwarf early types (dE + dS0), spirals and smooth, magellanic-type irregulars (S + Im), and clumpy irregulars (blue compact dwarfs, BCD). Only galaxies with known velocities are depicted, although this is a restriction only for the dEs, as all other types are essentially complete with respect to kinematic data (all dEs are shown in Fig. 9, Sect. 6, in a different context).

Figure 3

Figure 3. The distribution of the main morphological classes of galaxies in the Virgo cluster: E + S0, dE + dS0, spirals + magellanic irregulars, and clumpy irregulars (BCDs), shown in four panels. Only galaxies with known radial velocities are shown. The irregular-shaped contour is indicating the area of the Las Campanas Survey of the Virgo cluster (Binggeli et al. 1985). The crosses mark the position of M87. Figure from Binggeli et al. (1993).

From Fig. 3 we note: (1) late-type galaxies are much more dispersed than early types - independent of luminosity, (2) The southern M49 subclump is very spiral and irregular-rich, unlike the northern M87 / M86 subclump structure, (3) the E-NW axis of the cluster, which is aligned with the projected jet direction of M87, is nicely traced out by the Es and S0s, (4) there is a prominent asymmetry in the distribution of dEs with respect to M87 (of which more in Sect. 6 below). For more details, including the Dresslerian morphology-density relation for the Virgo cluster, the reader is referred to Binggeli et al. (1987).

These morphological differences are even more pronounced in the kinematic space. In Binggeli et al. (1993) we have collected and statistically analysed the radial velocities of ca. 400 Virgo cluster members. Most measurements are from the optical Center of Astrophysics Redshift Survey (Huchra et al. 1983, Geller & Huchra 1989) and the Arecibo H I survey of late-type Virgo members by Hoffman et al. (1987, 1989). Special efforts to get the velocities of a number of dE galaxies, which play a key role in our analysis, are due to Bothun & Mould (1988). In fact, all 800 odd Virgo members still lacking a velocity are dEs: their notoriously low surface brightness renders spectroscopy essentially unfeasible, at least at present.

Figure 4

Figure 4. The heliocentric velocity distributions of Virgo cluster members divided into the four main morphological classes shown in Fig. 3. The velocities of M86 and M87 are indicated in the panel for early-type dwarf galaxies.

The basic kinematic data for the main morphological types, reproduced from Binggeli et al. (1993), are listed as velocity means and (r.m.s. or 1sigma) dispersions in Table 1, and shown as distributions in Fig. 4. From these we note the following: (1) the velocity distribution of late-type members (spirals and irregulars) is significantly broader than that of early types (giant and dwarf E + S0s), i.e. late types are more dispersed in space and velocity, (2) the velocity distribution of spirals and irregulars is distinctly non-Gaussian, though it is fairly symmetric with a low-velocity and a high-velocity wing, (3) the velocity distribution of dwarf ellipticals is non-Gaussian and non-symmetric, being skewed towards low velocities, (4) the velocity of M87 is off the cluster mean by +200 km s-1 (as its projected position is also off a naively determined global cluster center), but is coinciding with the peak (median) of the velocity distribution of dEs, (5) the velocity of M49 is not significantly different form the cluster mean, while the velocity of M86 is even negative, coinciding with the low-velocity tail of the velocity distribution of dEs.

Table 1. Heliocentric velocities of Virgo cluster members

sample N < v > sigmav
(km s-1) (km s-1)

E + S0 75 1017 589
dE + dS0 93 1139 649
S + Im 188 1031 737
BCD 29 1110 795

All 399 1064 699

M87 1258
M86 -227
M49 969

These kinematic features, in connection with the projected spatial distributions discussed before, have been interpreted in the following way (cf. Huchra 1985, Binggeli et al. 1987, 1993). The broad velocity distribution of spirals and irregulars likely means that most of these galaxies are not yet relaxed (virialized); if they are only bound to the cluster, one indeed expects a velocity dispersion that is higher by approx sqrt2 than the dispersion of the presumably older, relaxed E + S0 population. The existence of low and high-velocity wings in the S + Irr distribution is a strong indication for infalling/expanding shells of late-type galaxies around the core of the cluster. It is quite plausible that nearly all spirals and irregulars are late, or even future arrivals, and hence are not yet virialized. The surrounding low-density field of the cluster, where these types of galaxies predominate, is subject to a global clustercentric velocity perturbation with a characteristic infall pattern (Rivolo & Yahil 1983, Tully & Shaya 1984). Field late-type galaxies are constantly fed into the cluster. Based on H I properties, it seems even possible to discriminate beteween spirals that have already fallen through the cluster core and spirals that are still in approach: the former, which are typically found in the central cluster area, are naturally identified with those spirals that are strongly H I-deficient (Haynes & Gionavelli 1986) and have very small H I disks (Cayette et al. 1990).

A clear asymmetry in the velocity distribution of a cluster of galaxies is almost certainly an indication of ongoing subcluster merging (e.g. Schindler & Böhringer 1993). The present asymmetric velocity distribution of Virgo dEs is taken as sign of the merging between the M87 and M86 subclumps (or rather, the infall of the M86 subclump into the more massive M87 subclump), for which there is additional evidence from X-ray observations, as will be discussed in Sect. 6. Both giant galaxies are obviously the centers of huge swarms of dwarf ellipticals, which is why the velocity of M87 is coinciding with the peak of the dE velocity distribution and not with the cluster mean, while M86 is apparently falling into, or through the M87 subclump from the back, hence with a high relative (negative) velocity, dragging along a smaller swarm of dwarfs, some of which are the most blueshifted galaxies in the sky (cf. Sect. 6). Finally, the well-behaved velocity of M49, coinciding more or less with the cluster mean, would suggest that the M49 subclump is approximately at the same distance as the M87 / M86 core structure, and that their supposed future merging will take place in the plane of the sky. This view is supported by the lack of a significant difference in the distance moduli of the M49 and M87 / 86 subclumps based on a host of different distance indicators (cf. Federspiel et al. 1998).

Next Contents Previous