12.7.2. Voids and the Segregation of Galaxian Properties

Progress in the collection of radial velocity data, both by optical and radio means, has made clear that the distribution of optically luminous matter may be outlined by regions of enhanced density mixed with "voids," regions largely evacuated of luminous matter. Density contrasts on the order of 10 to 100, between the large-scale high- and low-density regions, are common. Currently popular cosmological models require the universe to have a mean density equal to the critical value (i.e., one that will asymptotically reduce the expansion rate of the universe to zero). The majority of matter necessary to achieve critical density is not visible and probably not in baryonic form. Virial analysis of groups and clusters of galaxies indicates that if the universe has indeed the critical density, the dark matter is not as clumped as are the luminous galaxies. It then appears difficult to understand how the baryonic component - in this scheme, a dynamically minor component of the universe - would have separately evolved to the high degree of clustering that is observed. One possible solution to the problem is that the baryonic mass distribution, in spite of the highly contrasted picture offered by the bright galaxies, is actually only slightly nonhomogeneous. In that case, one must postulate that galaxy formation is a threshold process. If bright galaxies only formed in 2- to 3 peaks of the density fluctuations of pregalactic material, bright galaxy formation thrived in regions dominated by large-scale perturbations slightly above the average density. That process was inhibited, however, in large-scale regions with mean density slightly below the universal average. This scheme would result in a large apparent amplification of the density contrast; such "biasing" of the galaxy formation process is well described by Rees (1985).

We may then address the observational questions of whether "voids" are actually devoid of baryonic matter or are populated by less conspicuous conglomerates such as very-low-luminosity galaxies. The answers require the study of samples of intrinsically faint galaxies over large volumes. Because' their surface brightness is also low, such objects become increasingly difficult targets of optical surveys. A large number of them are HI rich, however, because a large fraction of their mash is in the form of HI (See Section 12.5).

Recent studies (Bothun et al. 1986, Oemler 1986) indicate that dwarf galaxies are far from filling the voids. Studies of larger samples (such as that in the Pisces-Perseus supercluster, mentioned in Section 12.7.1) suggest that the density contrast between low- and high-density regions diminishes with the intrinsic luminosity of the galaxian population. This effect is sketched in Figure 12.10, where the density contrast within a large solid angle encompassing a conspicuous void in the Pisces-Perseus supercluster region is illustrated separately for bright and faint galaxies. The relative excess or deficiency of galaxies is estimated in comparison with the expectation from a skywide determination of the galaxian luminosity function. It shows that the density contrast between high- and low-density regions diminishes for fainter galaxies, a trend in the direction indicated by biased galaxy formation schemes.

Figure 12.10. The ratio between the number of galaxies observed and the number which would be expected from a homogeneously distributed population, plotted as a function of radial velocity within a solid angle which encloses both a foreground "void" and part of the high-density enhancement seen in Figure 12.9 near 5000 km s-1. The ratio is plotted separately for bright and fainter galaxies (the label of each curve corresponding to a photographic absolute magnitude limit). Notice that both the depression associated with the void and the peak associated with the supercluster are milder when seen with a "filter" which excludes the brightest galaxies.

The effect illustrated in Figure 12.10 indicates that there is luminosity segregation between different local density regimes. Other forms of segregation according to morphological type, gas content, and total mass are present. Such variations are intimately related to the physical processes that led to galaxy formation. Some of these effects, such as the local density dependence of the gas content, can be exacerbated in particularly active environments, such as the cores of clusters, as was described in Section 12.6. Those secularly occurring effects, however, may be small-scale alterations to the inbred characteristics of the galaxian population.

As an alternative component of baryonic matter capable of making a contribution to the mass density distribution, the existence of not fully collapsed intergalactic clouds of HI has been postulated. Searches, both directed and serendipitous, for such objects (Krumm and Brosch 1984, Altschuler et al. 1987) have been unsuccessful. Whether diffuse HI is abundant in intergalactic space is, however, a relatively open question because the dominating excitation process of the hyperfine levels and the spectral signature of possible HI features are still largely uncertain.

An alternative mode of investigating the dynamic mass contrast, i.e., whether the overall mass distribution is traced by that of luminous galaxies, consists in studying the magnitude of large-scale peculiar motions of galaxies. If voids are indeed devoid of mass, then substantial deviations from a smooth Hubble flow, as discussed in Section 12.4.2, should be visible in the population of galaxies near their edges.