6.3. Double Systems in Groups and Clusters
Table 10 listed the 12 nearest pairs of galaxies, together with information on which double systems are located in the de Vaucouleurs (1975) groups. Only one pair, number 217, appears to be outside the limits of the de Vaucouleurs groups. In a recent large scale image (Supplement Plate XV), this object appears to be a single dwarf galaxy of irregular form (extragalactic HII region) instead of an actual double system. The situation in which all (or almost all) pairs of galaxies are also group members appears paradoxical. However, neither the statistics of pair separations nor the form of the correlation function p(r) contradicts this statement. We note that the de Vaucouleurs groups present a sufficient variety of structural and dynamical properties that the presence of many pairs among them (for example, the interacting pair M51 in G5) cannot be disputed.
Viewing double galaxies as structural elements in systems of higher multiplicity (Karachentsev, 1981e), we were able to outline the very closest de Vaucouleurs groups and the largest well-studied clusters in Virgo and Coma. Such an attempt has errors due, at a minimum, to the uncertain assignment of pairs to groups or clusters. Upon exclusion of several false double systems we obtain a subsample of 25 pairs, of which 10 were located in de Vaucouleurs groups. On examining these pairs within the entire population of double galaxies we note the following trends.
The mean separation between components of pairs in groups and clusters is 16 ± 4 kpc, a factor of two less than for pairs in general. This deficit of wide pairs in systems may be explained by disruption due to tidal action from their surroundings, or might be explained as the action of the selection criteria in the presence of an excessive background of galaxies.
The mean orbital mass-to-luminosity ratio for double galaxies in larger systems, <fc> = 4.0 ± 2.7, does not exceed typical values obtained from internal motions in galaxies. This would suggest that dark matter in groups and clusters is not clumped on scales ~ 20 kpc, or is not associated with the individual members of systems. Such a situation might be expected if the excess virial mass in clusters is massive cosmological neutrinos (Doroshkevich et al., 1980).
The most striking property is the preferential distribution of double galaxies in the surroundings of groups and clusters. The mean distance of pairs from the centers is 2.0 ± 0.4, in units of the characteristic radius of the system. Such a segregation in radius between single and double members of systems is shown quite well by each de Vaucouleurs group and by other examined clusters. For example, figure 45 shows the distribution of galaxies in the group G10. Of 16 members, half are in isolated pairs indicated by the circles. The numbers indicate the radial velocities of the galaxies in kilometers per second.
Continuing this analysis, we were convinced that the selection criteria could not entirely explain the deficit of double systems in the central regions of the denser de Vaucouleurs groups. We could not identify a single interacting pair which failed the isolation criteria due to another group member appearing too close in projection. Apparently the observed concentration of double systems on the edges of groups and clusters has its origin, not in selection effects, but in physical causes.
The differences in the location of single and double members of systems may have arisen in the epoch of formation, when the development of large-scale structure in the Universe began. The formation and isolation of systems of galaxies with respect to the surrounding field probably involved a considerable amount of time (billions of years), during which the central regions of systems acquired extended peripheral regions containing a relatively larger number of double galaxies. In another scenario the major reason for this segregation might be the evolutionary process of the disruption of wide pairs via the tidal effects of the surrounding galaxies.
The observational data now on hand are not sufficient to allow a choice between the various origins of the segregation of double and single members of systems with radius. However, considering the probable role of tidal effects in disrupting wide pairs we can estimate the total system mass Ms. Following the tidal theory of Hodge and Michie (1969) for pairs of galaxies, leading to estimates of the mass Ms, we have
where M12 is the pair mass, rpc is the distance of the pair from the center of the system, and r12 is the separation of the pair members. In order to increase the statistics, we have combined five de Vaucouleurs groups at once. In this synthetic group, and in the Virgo and Coma clusters, we have calculated the masses using the maximal separation between pair members. For the observed objects the ratio of estimated system mass to total pair mass (Ms/M12) is: 1.3 × 102 (groups), 1.5 × 103 (Virgo), and 3.0 × 104 (Coma). As might be expected, these values reflect the increasing mass of systems with increasing scale. The satisfactory agreement between the tidal and virial estimates of the masses of groups and clusters of galaxies is confirmation of the role of tidal effects, especially in the disruption of wide pairs.
The existence of very large redshift surveys in large areas of the sky (Huchra et al., 1983) introduces new possibilities for analysing the distribution of double galaxies in groups, clusters and superclusters. As a first step, it is interesting to study the distribution of pairs in groups identified by Geller and Huchra (1983) and also in the nearest clusters from the Zwicky catalogue.
Struble and Rood (1981) drew attention to the fact that several nearby Abell clusters contain among their brightest members an unusually large number of double galaxies. They proposed that a high percentage of pairs among such clusters may have a significant effect on the rate of dynamical evolution of the clusters. In order to test whether there are systems of galaxies rich in pairs, we examined the likelihood of encountering double galaxies in de Vaucouleurs groups and in the nearest Zwicky clusters. To identify pairs in Zwicky clusters we used the list of 334 nearby CGCG clusters with measured radial velocities (Baiesi-Pillastrini et al., 1984). Within the boundaries of these clusters in the radial velocity interval
we found 96 catalogue pairs. Among these we found 12 cases in which there are two or more pairs within the contours of a single cluster. The observed probabilities of encountering double galaxies in various clusters were compared using a 2 criterion with the expectation being the same fraction of double systems in all clusters. This null hypothesis agrees very well with the observed situation both in the nearby Zwicky clusters and in the nearby de Vaucouleurs groups, but differs from the results of Struble and Rood.