2.6. Great Attractors, Dipoles and all that
2.6.1. The Great Attractor and Others
Lynden-Bell et al. (1988) identified a large scale enhancement in the distribution of galaxies as being a possible cause of the large scale Hubble flow deviations discovered in the S7 survey of elliptical galaxies. They dubbed this "the Great Attractor".
The remarkable flows in the direction of the Great Attractor, as deduced from the elliptical galaxy survey, receive some support from a survey of spirals for which distances have been determined using the IR Tully-Fisher distance indicator (Aaronson et al., 1989).
The existence of this attractor had been guessed at already by Lilje, Yahil and Jones (1986) who identified a quadrupole component in the Virgocentric flow on the basis of the Aaronson et al. (1982) survey of the Virgocentric flow. The long axis of the quadrupole might have been expected to point towards the center of the Virgo cluster, but it was found to point in the direction of the Hydra-Centaurus complex. The quadrupole has since been confirmed by the independent data set of Stavely-Smith and Davies (1989). The importance of the quadrupole component is that it is unambiguously gravitational in origin. The quadrupole component of the force exerted by neighboring masses falls if as r-3 and so the quadrupole imposes constraints as to where the mass causing the distortion of the flow is located.
The Great Attractor can be modelled by density distribution r-2 centered about a point 3,500 km s-1 away from us toward the Hydra Centaurus region with the total mass of ~ 1016 M. The detailed interpretation of the Local Group motion relative to the MWB is analyzed by Lynden-Bell, Lahav and Burstein (1989). There are local contributions to Local group motion as well as contributions from the directions of Perseus and Hydra-Centaurus. See the discussion on the "POTENT" method of reconstructing the cosmic density distribution and peculiar velocity fields.
Because the attractor lies in the Galactic plane, it is difficult to map. The search for the attractor lead other groups (Scaramella et al., 1989; Lahav et al., 1989; Raychaudhuri, 1990) to the discovery of a "super attractor" far beyond the Hydra-Centaurus system, (~ 140h-1 Mpc.) having about ten times the size of the proposed "Great Attractor". This super attractor is a strong concentration of rich clusters of galaxies. Its distance is, however, so great that it is unlikely to be responsible for the tidal forces that lead to the quadrupole distortion of the Virgocentric flow.
The existence of a Great Attractor poses a number of problems for theories of the origin of large scale structure. Bertschinger and Juskiewicz (1988), for example show that the Great Attractor is a 7 fluctuation in CDM models. Playing with the bias factor in CDM models can solve the problem, but creates other problems. It seems that only Peebles' (1987) isocurvature baryonic models with 0 = 0.4 and a very flat power spectrum (n = - 1) can solve the problem.
2.6.2. Dipole Convergence
The Local Group, and its surroundings are moving at what by cosmological standards is a high speed relative to the cosmic frame defined by the microwave background radiation. Although Rubin, Ford and collaborators (Rubin and Ford, 1976a, b; see section 2.4.1) had reported a high velocity of the Local Group relative to a sample of distant galaxies, the first unequivocal detection of the motion came from the microwave background dipole anisotropy (Smoot, Gorenstein and Muller, 1977; Lubin et al., 1983, 1985; Fixen et al, 1983). The COBE satellite's Differential Microwave Radiometers currently measures a dipole with an amplitude of 3.3 ± 0.2 mK pointing in a direction towards l = 265° ± 2°, b = 48° ± 2° (Smoot et al, 1991).
What was surprising was that the motion inferred from the microwave background dipole did not point in the direction suggested by Rubin et al. (1976a, b), and nor did it point in the `natural' direction, towards the Virgo Cluster of galaxies, our nearest significant mass concentration. There was indeed a significant component towards the Virgo cluster and this was viewed as a part of the general Virgocentric Infall (Davis and Peebles, 1983). The source of the main component of the motion was not apparent until the discovery by Lynden-Bell et al. (1988) of a systemic motion of the Local Group towards what has become known as the Great Attractor.
I shall discuss the nature of the Great Attractor elsewhere in these lectures, all that should concern us here is the detection of the same motion relative both to a sample of Elliptical Galaxies and relative to the cosmic frame in which the microwave background radiation is isotropic. That motion has been confirmed in many subsequent surveys using different types of objects: spiral galaxies, Abell clusters, optically selected galaxies, .... Note that it is not only the Local Group that is moving, but at least the local neighbourhood, as is evidenced by the quietness of the local Hubble flow (Brown and Peebles, 1988). Also, the fact that the long axis of the quadrupole distortion of the flow of galaxies in the Virgo supercluster points in the same direction is also of importance (Lilje et al., 1986; Stavely-Smith and Davies, 1989).
It is an important goal of cosmology to reconstruct the entire local flow from the distribution of known material in the universe. The question is where is the matter that is exerting this pull on the Local Group and its environs? The answer of Lynden-Bell et al. (1988) was the Great Attractor, a vast aggregation of galaxies largely hidden from our view by the Galactic plane. The attempts to confirm this hypothesis have met with some difficulty. We see for example the claim of Rowan-Robinson et al. (1990) that the entire effect can be explained in terms of known systems of galaxies, and that there is no need to invoke any special unseen objects like the suggested Great Attractor. There are also claims that there is a detectable influence on the Local Group motion from distances far beyond the Great Attractor (Plionis and Valdarmini, 1990).
The issue is referred to as dipole convergence. As the influence of galaxies from ever larger shells is counted, the direction of motion of the Local Group should converge in magnitude and direction to the microwave back ground dipole direction. Several corrections have to be made, the most important being a correction for the galaxies and clusters that are hidden from sight because of the Galactic Plane. There is another correction that has to be made: we measure the microwave background anisotropy from the point of view of the Solar System frame of reference whereas the dynamical forces act on the Local Group mass center and on the Galaxy. A failure to achieve convergence could mean one of several things. It could mean that the light distribution does not trace the mass, it could mean that we have not gone far enough in distance, or it could simply be that we do not know the motion of the Sun relative to the mass center of the Local Group.
If the distances to the galaxies are known, we can sum up the contributions from shells of ever increasing size to the local force field. Note that in order to fit the amplitude of the dipole, we must assume a value of 0 / b5/3. The bias parameter b comes in because we look at the fluctuations in the light distribution, not the mass distribution.
Light and gravity both fall of as R-2, and so, if light traces mass, a dipole in the gravitational force field should show up as a dipole in the light distribution. Thus for a sample of objects drawn from a catalogue, we can deduce the contributions of various parts of the Universe to the local force simply by examining the distribution of the light and making the necessary assumption that light traces mass. This has been done for virtually every available catalogue of galaxies: spiral galaxy samples, elliptical galaxy samples, the IRAS-based samples and for samples of galaxy clusters.
Lahav (1988) summed up the light in diameter limited subsamples of a catalog of optically identified galaxies as a function of limiting optical diameter He found a light dipole in roughly the expected place, though he found convergence in direction was achieved at a relatively nearby distance. An enhancement in the galaxy density in the direction of the Great Attractor is clearly seen in his plots of the distribution of galaxies on the sky. This has also been done for the IRAS survey (Strauss et al., 1988; Yahil, 1988), and for the QDOT survey (Rowan-Robinson et al., 1990), and it is clear that convergence in direction is being achieved, though the dipole direction does move around on the sky quite a lot data from ever large shell is considered.
Plionis and Valdarmini (1991) study the dipole contribution from all galaxy clusters in the Abell-ACO catalogues having their 10th brightest galaxy brighter than in m10 = 16.4. The catalogue is 80% redshift complete. They calculate the dipole moment of the light distribution from the clusters as a function of depth in their catalogue, using the population of each cluster as an estimator of the total cluster light. In order to calculate the acceleration due to a given cluster, they rescale the light cluster light to a mass in such a way that they get the correct answer for the Coma cluster mass. Most of the effect comes, as expected, from a volume of radius r < 50h-1 Mpc., but there still appears to be a substantial contribution from the Shapley concentration of galaxy clusters and clusters in general out to the limit of their survey. What is perhaps surprising about this result is that the clusters should trace the mass distribution so well. Only a few percent of all galaxies are in such clusters and it would have been quite conceivable that the contribution to the local gravitational field from the clusters would be swamped by the contributions from all the other galaxies.
Recall the discussion of the QDOT survey and the fact that Rowan-Robinson et al. (1990) were able to obtain satisfactory dipole convergence using only clusters that had been identified in the QDOT survey. There was no need for `extra' mass hidden behind the Galactic plane. Rowan-Robinson et al. refer to highly extended (and overlapping) cluster halos which have power law density distributions ( r-1.6) extending out to 30h-1 Mpc. Only 1.5% of the cluster mass lies within the Abell radius! It is presumably these overlapping halos that the POTENT reconstruction is finding and labelling the Great Attractor with a density contrast of around 0.7.
The outstanding question is at what distance is convergence in direction achieved? The amplitude of the velocity can be fixed by selecting a mass to light ratio for the objects in the sample, and that is equivalent to fixing the parameter 0 / b0.6 with a bias parameter, b, appropriate for the sample. There are several aspects to this question. Firstly there is the question of the statistical significance of the inferred direction and the value of the velocity. There is then the question as to whether such an amplitude is expected in a given theory and the confidence with which we can determine 0 and b.
Several people (Kaiser and Lahav, 1989; Juskiewicz, Vittorio and Wyse, 1990; Lahav, Kaiser and Hoffman, 1989) have constructed models for Local Group peculiar velocity, vR, in Cold Dark Matter and Baryonic Dark Matter models. The data from the IRAS and elliptical galaxy surveys out to 10,000 km s-1 cannot tell the difference between these models with any confidence. Another approach (Regos and Szalay, 1989) is to use multipole expansions of the observed radial velocities in shells to shed light on the uncertainties in deriving the bulk velocity vector. It seems that distinguishing between what observers measure and what theorists talk about is a large part of the problem.
It would be important if there were a difference between the true peculiar velocity of the Local Group, v as inferred from the MWB dipole and the estimate vR (as inferred from the volume averaged motion in a sample of galaxies (Vittorio and Juskiewicz, 1987)). One of the uncertainties in converting the motion of the Solar System relative to the microwave background to the motion of the mass center of the Local Group relative to the cosmic frame is the unknown dynamics of our Galaxy relative to the Local Group mass center. There is even a contribution from the rotation of the Local Group which is only poorly determined (Moore, 1991).
2.6.3. Large Scale Flows and CBR Anisotropy
It is an important question as to whether these non-Hubble motions arose from density fluctuations that were consistent with observed limits on the anisotropy of the microwave background radiation temperature. The anisotropy on such scales results simply from the Sachs-Wolfe effect.
Juskiewicz, Górski and Silk (1987) and Martínez-González and Sanz (1989) calculate the minimal microwave background anisotropy associated with a given streaming motion. By `minimal' what is meant is that the density fluctuation spectrum is chosen so as to minimize the observed temperature fluctuations, subject to the constraint that they can also produce the velocity field. Although it is a near thing, no microwave background anisotropy experiments are yet in conflict with the observed non-Hubble motions.
Of course, it might be that the actual spectrum gives an observable temperature fluctuation. Doroshkevich and Klypin (1988) used the Zeldovich approximation to describe the evolution of velocity correlations on very large scales, and they also calculated the expected temperature anisotropies for the purposes of comparison with the RELICT experiment. They argued in favour of needing a feature in the spectrum of fluctuations on scale of 50-100 h-1 Mpc.