3.6.3. Large Scale Flows: Continental Drift in the Nearby Universe?

The Aaronson et al. 1986 data set was restricted to the declination range of the Arecibo Observatory as they used the TF relation to measure the distances and V_c was obtained from 21-cm neutral hydrogen emission profiles in rotating disk galaxies. From that dataset Aaronson et al. (1986) set a limit of 500 km s^-1 on the random motion, due to the large scale mass distribution, of clusters of galaxies. This argues that, for distant clusters, only small corrections are needed to transform from observed velocity to cosmic velocity. Hydra-Cen is located in the Southern Hemisphere (and hence unobservable from Arecibo) at an observed velocity of 4500 km/s. The Aaronson et al. limit of 500 km s^-1 thus is only a 10% perturbation in the cosmic velocity of Hydra-Cen.

The Aaronson et al. 1986 paper was published in March. Two months prior to that was a conference entitled "Galaxy Distances and Deviations from Universal Expansion" which was held on the Kona coast of the Big Island of Hawaii. The meeting was noteworthy in two respects 1) boring presentations could be compensated for by a quick dip in the ocean, 100 yards away and 2) the first data were presented that the Hydra-Cen supercluster had a much larger peculiar velocity than was allowed for by the Aaronson et al. 1986 data. In every respect, this data would change our perception of the local Universe forever and ultimately lead to heated debates in the professional journals as to exactly how noisy the local Hubble flow was. To date, this situation has not been resolved (see the excellent and comprehensive review of Strauss and Willick 1995).

The principle fault of the Aaronson et al. sample was its limited sky coverage. A proper mapping of the large scale flow pattern requires an all-sky sample. In the early 80s, a team of astronomers led by Sandy Faber of the University of California at Santa Cruz developed a method that was similar to the TF method but which could be applied to elliptical galaxies. Recall that the TF method works under the assumptions that the circular velocity V_c of a rotating disk galaxy is driven by its total mass. If there is little variation in the ratio of Mass to Luminosity (M / L) then V_c is a measure of intrinsic luminosity. Elliptical galaxies are non-rotating and are supported by the internal velocity dispersion (_v) of its stars. More massive ellipticals have deeper gravitational potentials and hence higher values of _v. If M / L for ellipticals has little variation then _v is an indicator of intrinsic luminosity. There are two sources of error associated with this method that were originally too large to make it competitive with the TF relation. These error sources are:

anisotropy in the orbits of the stars in ellipticals. Just as the TF relation demands that rotating disk galaxies are circularly symmetric, the _v method demands that the orbits of the stars are isotropic. This is because we only measure the radial component of _v. If the orbits are anisotropic then our value of _v will depend on the orientation of the long axis of these anisotropic orbits with respect to the observer.

_v has a radial dependence in elliptical galaxies which means that observational determinations of _v are aperture dependent. The same problem exists in the TF relation as measurements of V_c are also aperture dependent as the rotational velocity of a galaxy is a function of radial distance from the center. For spiral galaxies, however, V_c can be determined from an aperture which is larger than the galaxy itself and hence contains the whole rotation curve. This is the main advantage of 21-cm neutral hydrogen observations. For elliptical galaxies, its impossible to measure _v through a large aperture because the signal comes from the integrated brightnesses or all the stars and hence is dominated by the central highest surface brightness regions of the elliptical.

The first source of error can not really be overcome although dynamical models of ellipticals are consistent with a low degree of anisotropy. The second problem was solved by the combined talents of Alan Dressler, David Burstein, Roger Davies and Donald-Lynden Bell who were all members of Faber's elite team. After looking at the _v data for ellipticals for many years, they concluded that improvement in Faber's original method (e.g., Faber and Jackson 1976) could occur if apertures which enclosed a constant surface brightness from one elliptical to the next were selected. Using this method to define the aperture in which to measure _v gave rise to the D_n- relation for measuring relative distances between elliptical galaxies. The scatter in this relation is similar to that in the TF relation and thus D_n- is now competitive with the TF relation.

The D_n- method has the advantage that only optical telescopes are required and hence it can be applied to any elliptical galaxy in the sky. The D_n- group (see Dressler's 1994 book Voyage to the Great Attractor for a detailed summary of the work of this group) made measurements in regions of the sky that Aaronson et al. could not cover. This fuller sky coverage revealed disconcertingly large deviations from expansion motion (see Dressler et al. 1987). For the Aaronson et al. group, these results were most distressing because they meant that a reliable determination of H₀ from cluster data was probably not possible. Furthermore, if Hydra-Cen is moving with respect to the CMB then it cannot be the sole source of the observed DA and a more distant mass concentration is required if the motion is gravitational in origin. What is even more interesting was the possibility that the entire region from the Milky Way to the Hydra-Cen supercluster was moving at 600 km s^-1. This motion is referred to as bulk flow and suggests a kind of plate tectonic model for the nearby Universe in which large regions are streaming, at a constant velocity, towards distant mass concentrations. But, what mass concentration could produce acceleration over such large of scale?

Additional analysis of the elliptical galaxy data by Lynden-Bell etal (1988) lead to a model in which the idea of bulk flow was replaced by an infall pattern that was driven by a rather large mass concentration. This mass concentration has been dubbed ''The Great Attractor''(GA) and is the subject of Dressler's book. The putative GA lies behind the Hydra-Cen supercluster at a kinematical distance of 4350 km s^-1. Infall of Hydra-Cen, the Local Supercluster and our Galaxy toward the Great Attractor (GA) then accounts for the observed positive peculiar velocities. Thus, the LG feels both the accelerations of the Virgo cluster and the GA and the relative normalization of these two vectors depends on and the relative overdensities / of the two mass concentrations.