Looking out across the sky, astronomers find structures at both Solar System scales and at scales far greater than any individual cluster of galaxies. These largest structures, are actually referred to as the large-scale structure of the universe. On the scale of galaxies and larger, structure is hierarchical — that is to say, galaxies are part of structures, which are part of larger structure, and so on. We actually find clusters of galaxy clusters, called galaxy super clusters. This progression doesn't go on forever, however, there is a scale at which on average the distribution of matter begins to be smooth (even though it's laid out in "particles" or units of galaxies). This scale is about 100 Mpc or 330 million light years.
Gravity has a long reach and can cause galaxies to aggregate on an enormous scale. Astronomers naturally speculated that even larger structures than galaxy clusters might exist. In the 1940s, Clyde Tombaugh, the discoverer of Pluto, plotted the positions of galaxies in the sky and made the first map of a structure of galaxies that spanned much of the northern sky. He showed it to Edwin Hubble, who refused to believe the observation, no doubt because no one at that time expected to see such large structures. Part of this reluctance comes from the belief that if we look at the universe on a large scale everything is the same everywhere. This is the assumption of isotropy and homogeneity. Isotropy is the idea that the universe looks roughly the same in and direction, and homogeneity is the idea that the universe is roughly the same at any location. You can see that these are quite different assumptions. It's easy to point a telescope in a different direction, but it's impossible to relocate ourselves to a distant location in the universe! What wasn't understood at the time of Tombaugh and Hubble was just how big a scale we have to look at the universe with to verify this assumption.
Twenty years later, French astronomer Gerard de Vaucouleurs showed that the Milky Way is in fact part of an enormous flattened structure called the Local Super cluster. A projection of the brightest galaxies in the northern and southern sky reveals a high concentration of galaxies running in a strip across the sky, nearly at right angles to the Milky Way. This flattened structure of galaxies is analogous to the flattened disk of stars in the Milky Way. Other astronomer were skeptical of de Vaucouleurs' work and he didn't have redshifts to verify that the structure existed in three dimensions.
The Local Super cluster contains the Local Group, the Virgo and Coma clusters, and about 100 other clusters. It measures about 20 Mpc across by 2 Mpc thick and contains a total of 1016 solar masses. In other words, it would take light 65 million years to cross the Local Super cluster and this super cluster contains ten thousand trillion stars! The center of the Local Super cluster is near the Virgo cluster; the Milky Way is near the outskirts. The gravity of the Virgo cluster is so strong that we — the Milky Way and the entire Local Group — are being pulled into it at a speed of 250 km/s. You should not lose sleep over this, however, it will take many billions of years to get there.
Super clusters range in size from 50 to 100 Mpc across — this is a physical structure more than 300 million light years across. By comparison, galaxy clusters are usually in the size range of 1 to 5 Mpc. On the largest scales, astronomers see knots or concentrations of galaxies, linear or filamentary features called walls, and regions where the galaxy density appears to be very low. Much modern research is devoted to understanding this complex web of large-scale structure. The mathematics of large scale structure is reasonably well described by a fractal, where there is more and more structure on smaller and smaller scales.
Sky maps of galaxies are not ideal for studying large-scale structure because they only represent two dimensions. Nearby and distant galaxies are very far apart yet they can appear close together in projection. The most detailed information on large-scale structure thus has come from painstaking red shift surveys of galaxies in narrow "slices" of the universe. Rather than try the almost impossible task of measuring distances for thousands of galaxies, astronomers use recession velocity (or equivalently, red shift) as a distance indicator. Position on the sky plus red shift allows galaxies to be placed in three-dimensional space. The first of these surveys was carried out by Margaret Geller, John Huchra, and their collaborators at the Harvard-Smithsonian Center for Astrophysics. A series of separated slices can be used to reconstruct the distribution of galaxies over a large volume. Technology is advancing this science very quickly. By the 1990's, red shift surveys had measured over 100,000 galaxies and telescopes like the Sloan Digital Sky Survey telescope at Apache Point have increased this number to over a million. Astronomers are mapping out larger and larger regions of the universe, heirs to the pioneering cartographers who mapped our planet hundreds of years ago.
While red shift helps us understand the distances to clusters, this method of mapping the universe is not without flaws. Although the Coma cluster is spherical in shape, it appears elongated in red shift. This elongation occurs because the large mass of a cluster gives galaxies high orbital velocities within the clusters. The range in velocity, as some galaxies race towards us along their orbit and others race away along their orbit, spreads the red shift distribution of galaxies towards and away from the observer — astronomers call this the "finger of god" effect. In other words, red shift is a poor distance indicator of individual galaxies in a cluster. That said, the diversity of velocities in a single cluster can be averaged out giving us the true position of the cluster itself, even if we don't know the exact distance to individual galaxies.
A striking feature in surveys of galaxies and clusters over large volumes is the presence of voids in the galaxy distribution. The voids are nearly circular in cross section and are nearly empty of bright galaxies. The Local Super cluster has been mapped out in three dimensions by Brent Tully. Most of the volume is completely empty of luminous matter. As many as 98% of the galaxies occupy only 5% of the volume. Some voids are over 150 Mpc across, or nearly half a billion light years, which is a whole lot of nothing!
The mixture of super clusters, voids, and walls in the architecture of the universe has been steadily sculpted by gravity over billions of years. Astronomers have used colorful language to describe the structures they see. Historically, it was thought that the rich clusters were set in a uniform sea of galaxies, like meatballs. The discovery of strings and filaments of galaxies added a component of spaghetti. The first slices of the universe showed that the galaxies form large connecting structures around empty voids 20 to 100 Mpc across, like Swiss cheese. The most recent research indicates a single analogy that cleans up these various terms. The general emptiness of space — where galaxies are distributed in sheets and filaments and clusters are found at the junctions of the sheets — is like a network of soap bubbles. Astronomers have also used mathematics to describe large-scale structure. What is the typical shape, or topology, of the way galaxies cluster in space? A point has no dimensions, and in three-dimensional space this would correspond to galaxies that were scattered with no clustering at all. A line has one dimension, so this would be the appropriate description of galaxies in stringy structures. A plane has two dimensions, so this would describe galaxies that are all contained in sheet-like structures. Recent analysis indicates that the large-scale structure has a dimension of about 1.7, which is partway between 1 and 2. In geometry terms, a dimension of 1.7 corresponds to structures that are somewhat string-like and somewhat sheet-like.
This description actually only works for the modern universe though. The universe started out as a nearly homogeneous gas. It's taken the full 13.8 billion years of cosmic evolution to get the structure we have today. As we look further back in the past, the number of super clusters diminishes, the void shrink, and we see the universe going back to its more smooth original state. As the universe continues to evolve into the future, the structures we see will continue to contract; to gravitationally compress. More galaxies will merge, and more clusters will fall into super clusters. The structure we see is specific to the moment.
Even with all this structure, however, we still see the universe is isotropic and homogeneous at the largest scales. On scales of 10 or 30 Mpc, the observed variation in the number of galaxies exceeds the random variation expected from counting statistics (√N) — an indication of gravitational clustering. However, by a scale of 100 Mpc, the number of galaxies shows only a small fractional variation. By a scale of 300 Mpc, the universe is essentially smooth.
The study of large-scale structure reveals two important aspects of the universe on the largest scales we can measure. First, it is isotropic. We see a similar Hubble relation in every direction we look. We also see similar types of structures — clusters, super clusters, and voids — in every direction we look. Second, it is homogeneous. Imagine that you are looking down on a beach from a mile high. The grains of sand and pebbles would not be visible and the beach would appear smooth and featureless. Similarly, a mountain range would appear smooth as viewed from the Moon. Gravity has obviously formed lumps that range in size from planets and stars to clusters and super clusters of galaxies. But the universe is smooth on scales larger than about a billion light years, reflecting the smoothness of the original hot and dense state.
Clyde Tombaugh. Click here for original source URL.