Astronomy Supplement 22.

A Universe of Galaxies

Latest Modification: March 9, 1999

Table of Contents

Galaxies seemingly stretch away into space without end. Photographs taken with large telescopes reveal literally billions of galaxies; in fact, in some directions they outnumber the foreground stars of our Galaxy. After galaxies had been recognized as islands of stars in an immense black sea of space and time, interest grew in their arrangement in the Universe and their relationships to each other. Certainly the galaxies of the Local Group are not uniformly spread out in space. On the largest scale of the Universe, therefore, are galaxies uniformly or nonuniformly distributed? And how does that distribution change with time if at all? The simplest premise with which to begin is that on the largest scale galaxies are uniformly distributed in space. Any number of arguments can be advanced to make such a contention seem realistic, and that is more or less what has been assumed to be true about the large-scale organization of the Universe.

A second facet of this question comes from studies of clusters. Clusters of galaxies are gravitationally bound collections of hundreds to thousands of galaxies. Astronomers find that within a cluster galaxies are moving relative to each other at velocities of thousands of kilometers per second. Such high velocities implies that members of a cluster are bound together by much larger gravitational forces than can be accounted for by the mass of visible galaxies. So a paramount question is in what form and where are the colossal amounts of matter necessary to keep clusters from flying apart? And what effect does this dark matter have on the large-scale structure of the Universe?

22.1. The Local Group of Galaxies

22.1.1. What Is a Group of Galaxies?

As was mentioned above, gravity can bind galaxies together in pairs, or in groups containing several tens of galaxies, or in clusters containing hundreds to thousands of galaxies. Our Galaxy, along with about 20 nearby galaxies, belongs to a physical group of galaxies known as the Local Group (Table 22.1). Most of the galaxies in the Local Group are dwarf elliptical systems along with three spirals and four irregulars. The Local Group may actually consist of more than these 20 or so galaxies, since interstellar matter in the disk of our Galaxy could hide a number from our view, particularly if they are faint dwarf elliptical galaxies.

[Table 22.1]

The largest galaxies in the Local Group are three spiral galaxies: the Andromeda galaxy (M31; Figure 22.2), our Galaxy, and the Triangulum galaxy (M33; Figure 22.3), in that order of size. Smallest are the dwarf elliptical galaxies. The Local Group spans a region of space about 3 Mly across. And in three dimensions most of the galaxies are located around either our Galaxy or the Andromeda galaxy, as shown in Figure 22.1. This is presumably a result of the large mass and the consequent gravitational pull of our Galaxy and the Andromeda galaxy. All the galaxies in the Local Group are moving relative to each other, and it is probably that the small ones orbit the two large spirals, which in turn orbit each other with a very long period.

[Figure 22.1]

In addition to the galaxies listed in Table 22.1, there are at least eight additional galaxies lying within about 5 Mly of the Milky Way. Some astronomers have argued that these galaxies are also gravitationally attached to the Local Group, but there is still uncertainty about this point. There is a growing sense among astronomers that maybe all galaxies are part of small groups (such as the Local Group), large clusters, or superclusters. In reality, there may be few, if any, isolated galaxies.

Although it is not possible to present all that astronomers know about the galaxies of the Local Group, the following discussion will give some sense of the state of knowledge. As is true for stars, most galaxies can be fitted into one of several classes, but upon closer inspection, they all have their individual distinguishing characteristics, such that no two are precisely alike.

22.1.2. Andromeda Galaxy, M31

Visible to unaided eye as a faint, hazy patch of light, the Andromeda galaxy (Figure 22.2) is the largest member of the Local Group and is an Sb type normal spiral like our own Galaxy. Its longest angular dimension covers nearly 5o on the sky, the same as the distance between the two end stars in the bowl of the Big Dipper. The central plane of the galaxy is inclined at an angle of about 12o to the line of sight.

[Figure 22.2]

On short-exposure photographs, Andromeda's small, brilliant nucleus looks quite similar to the one detected by infrared and microwave observations of our Galaxy's nucleus. Two spiral arms appear to wind out of the nuclear bulge for several turns, trailing the direction of rotation. In recent years astronomers have identified hundreds of H II emission nebulae, associations of young blue stars, and some 400 open clusters--all marking the position of the arms in the disk.

The galaxy's dust lanes exhibit chaotic organization and do not appear to fit any spiral pattern well. Instead they are most noticeable as nearly a ring around the nucleus. Atomic hydrogen gas extends well out beyond the optical image of the galaxy; it does not extend into the center but stops about 12,000 ly from it. However, the hot, ionized hydrogen gas clouds (H II emission nebulae) taper off at about 50,000 ly, and also have a hole in their center about 12,000 ly in radius. This doughnut-shaped distribution of cold dust clouds, atomic hydrogen, and hot, ionized hydrogen clouds is now known to be characteristic of spiral galaxies (Figure 21.7). The absence of appreciable amounts of interstellar matter in the nucleus, the hole in the doughnut, suggests that star formation has ceased there. Although there is still great uncertainty as to whether the abundances of the chemical elements are similar to those in our own Galaxy, they do not appear to be dramatically different.

Spheroidal-population red giants and a number of planetary nebulae are the brightest objects in the galaxy's central regions; the outlying portions are dominated by bright, blue, spiral-arm-population supergiants. In the intermediate areas between the arms, which are relatively free from obscuring material, there is a mixture of different age disk-population stars. These interarm regions are sufficiently transparent for us to see remote galaxies through them. In the galaxy's halo, about 400 globular clusters can be readily identified.

From the work of the Einstein Orbiting Observatory, about 80 sources of X-rays have been found. They appear to be comparable to the brightest X-ray sources in our Galaxy. Figure 22.2 shows both a wide- and a narrow-angle view of the nuclear bulge in Andromeda. As can be seen, some X-ray sources are in spiral arms and some are in the nucleus. Although in both places X-ray sources signal the occurrence of violent events, possibly dying stars, they are not exactly the same kind of event. Finally, about 20 X-ray sources are in globular clusters.

As in our Galaxy, the central bulge of the Andromeda galaxy rotates much like a solid body, whereas the outer disk shows an almost constant rotation velocity out to great distances from the center. This uniformity, along with the extension of atomic hydrogen well beyond the optical image, suggests that Andromeda possesses a very large and massive halo containing a dark component of unseen matter like that for our Galaxy. In almost everything our stellar system and the Andromeda galaxy are basically alike: major features, physical characteristics, and composition of celestial objects.

22.1.3. Triangulum Galaxy, M33

The two spiral galaxies besides our Galaxy in the Local Group, Andromeda and Triangulum, do not possess the most elegant or dramatic spiral patterns, but they have the advantage of being nearby. The Triangulum galaxy (M33) is a Hubble-type Sc galaxy (Figure 22.3), and we see it almost face-on. About a tenth the size and mass of the Andromeda galaxy, it contains only a few tens of billions of stars.

[Figure 22.3]

Although smaller than M31, the Triangulum galaxy emits almost as much energy in the form of radio waves as does Andromeda. This may not be surprising: With more interstellar matter in an Sc galaxy such as M33 than in an Sb such as M31, more stars should be forming, as is evidenced by the larger H II regions in Triangulum. Supporting the hypothesis of an enhanced star-formation rate is the fact that red giants, stars having evolved off the main sequence, are about 100 times more prevalent in Triangulum in a region the size of the solar neighborhood than they are in Andromeda or our Galaxy.

Emission of 21-cm photons by hydrogen atoms comes from all parts of the disk, but it is somewhat heavier from what appears to be a multiarm--though broken and disordered--spiral pattern. Studies of the ionized hydrogen emission nebulae show a multitude of filaments, loops, and arcs as part of the structure of individual H II nebulae. This supports the 21-cm results that show that the galaxy has a rather chaotic spiral structure.

22.1.4. Large and Small Magellanic Clouds

The closest extragalactic neighbors (both are visible to the naked eye) are two objects in the southern skies called the Magellanic Clouds. They were named in honor of the explorer Ferdinand Magellan, whose observations of them in 1522 introduced these galaxies to the Western world. The Large and Small Magellanic Clouds are a physically related double system immersed in a common envelope of neutral hydrogen gas. They are considered satellites of and move in large orbits around our Galaxy.

The orbital plane for the Large Magellanic Cloud appears close to perpendicular to our Galactic plane; its closest approach is about a quarter of a million light years. Figure 22.4 shows the Magellanic Clouds relative to the Milky Way. Also shown is an arc of atomic hydrogen containing some tens of millions of solar masses of gas that appears to follow a great circle through the Large Magellanic Cloud. The structure is known as the Magellanic Stream and may represent gas pulled out of the Large Magellanic Cloud by our Galaxy. This is certainly feasible if our Galaxy is as much as 10 times more massive than we presently suppose.

[Figure 22.4]

The central plane of the Large Magellanic Cloud is tilted nearly 90o to our line of sight; we also see the Small Magellanic Cloud at an oblique angle. Both galaxies have a ragged, disklike structure somewhat flattened by rotation. In them there are stars of all descriptions and ages, including thousands of Cepheid variables as well as gaseous nebulae, star clusters, and several supernova remnants. Very prominent in the Large Magellanic Cloud are numerous blue and red supergiants and obscuring interstellar dust. Thus some star formation is still going on.

In fact, young blue stars are seen in some globular clusters in addition to the usual red giants. This means that these globular clusters formed within the last billion years or so. The Magellanic Clouds, as seems to be true for other irregular galaxies, appear to have been forming stars at a slower and more uniform rate over their lifetimes, whereas our Galaxy and other spiral systems went through an early burst of star formation and have since slowed down considerably.

22.2. Clusters of Galaxies

22.2.1. Small Groups of Galaxies

The clumping of matter in the Universe is certainly evident on all levels of size we have discussed so far. Starting with atomic nuclei and moving to larger scales, we find dense clumps of matter, such as people, Earth, stars, clusters of stars, galaxies, and clusters of galaxies.

Galaxies occur in a wide variety of gravitationally bound systems, ranging from binary pairs through small groups containing a few tens of galaxies to rich clusters composed of hundreds to thousands of galaxies. These clouds of galaxies in turn vary widely in the number and spacing of member galaxies. From small groups to rich clusters the typical separations of bright galaxies vary from a few tens of thousands of light years to a few million light years (a few tenths of our Galaxy's diameter to a few tens of our Galaxy's size). We must talk about typical separations among bright galaxies because these are the ones we see; the number of faint dwarf galaxies lying between the bright ones is simply not known.

In the category of small groups of galaxies probably the most common are the rather loosely packed groups containing fewer than 10 bright galaxies with typical separations of a few million light years (or a few tens of diameters for a large spiral). Examples of small groups of galaxies are those centered on the bright spiral galaxies M81 (Figure 22.5) and M101 (Figure 20.3). Table 22.2 lists some properties of these two small groups as compared with our Local Group. Within about 30 Mly there are a number of small groups similar to those listed in Table 22.2, and fewer than 25 percent of all galaxies within that distance are not members of a small group. Thus it is tempting to ask whether all galaxies are, or were at one time, members of some kind of grouping. It is probable that only a lack of enough observing time on large telescopes prevents astronomers from answering this question. With the ability of Space Telescope to observe galaxies 10 times fainter than the best Earth-based telescopes can, the astronomical community may have an answer to this question in the not too distant future.

[Figure 22.5]

[Table 22.2]

22.2.2. Rich Clusters of Galaxies

Considerably larger and more complex than small groups of galaxies are rich clusters, containing hundreds to thousands of galaxies. Even in the early part of this century rich clusters were clearly evident to the pioneers of extragalactic astronomy. Today several thousand rich clusters have been identified--out to distances of about 4 billion ly--which contain not more than 10 percent of all galaxies. Therefore, more galaxies appear to be members of small groups than of rich clusters. Fourteen representative rich clusters are listed in Table 22.3. (The actual number of galaxies in a cluster may be much larger than noted if the dwarf galaxies could be seen.)

[Table 22.3]

Anyone who has study one of the biological sciences is familiar with the idea of cataloging and classifying in those fields. Although such activities may seem to be no more than mundane tasks "to help tidy up the subject," they are critically important activities. Cataloging and classifying help to establish the scope of the particular science and are the foundation upon which more specialized research will be carried out later. To help organize clusters of galaxies as a field of study, astronomers have devised a classification scheme for them analogous in principle to stellar spectral classification (Section 13.3) and Hubble's galaxy classification sequence (Section 21.2). Many cluster properties have been taken into consideration in developing a classification scheme, such as

Using these various criteria, clusters can be placed in one of three classes: regular, intermediate, and irregular, where the name describes the general appearance of a cluster as elliptical and the like do for galaxies (Table 22.4).

[Table 22.4]

Regular clusters are all rich, high density clusters having galactic populations numbering on the order of a thousand galaxies or so. These clusters are roughly spherical in shape, with a high concentration of galaxies toward their centers. Regular clusters are composed largely of dust-free galaxies--the E and S0 Hubble types--with spirals less than 20 percent of their membership. Their centers are dominated by giant elliptical galaxies, often referred to as supergiant diffuse (cD) galaxies, with extended halos, and few, if any, have spiral galaxies in their centers.

An example of a regular cluster is the rich cluster known as A2199 [number 2199 in George Abell's (1927-1983) 1958 catalog of rich clusters], which is dominated by NGC 6166 (largest galaxy in Figure 22.6). This cluster is about 600 Mly from our Galaxy, a distance that implies that NGC 6166, a supergiant elliptical galaxy, is almost 2 Mly in diameter, or one of the largest galaxies known. Surrounding the galaxy is a large cloud of gaseous matter that is emitting radio and X-ray photons. The X-ray emissions are apparently the result of thermal processes, indicating that the gas is extremely hot (20 to 200 million K), rather than due to nonthermal processes. Surprisingly, the gas contains an excess quantity of iron that has lost 24 of its 26 electrons. This suggests that the gas is not primordial, which would be predominantly hydrogen and helium. But it is gas that has been enriched with heavy elements produced in supernova outbursts, and thus it comes from galaxies in the central regions of the cluster that has collected around NGC 6166. The cluster A2199 is not the only regular cluster in which hot gas fills the cluster's center. Other examples are the Coma cluster (Figure 22.7), the Perseus cluster (Figure 22.8), and the Corona Borealis cluster (Figure 20.1).

[Figure 22.6]

[Figure 22.7]

[Figure 22.8]

Figure 22.9 reveals X-ray emissions by gas in the cluster A85 in the constellation Cetus by contour lines superimposed on a negative reproduction of the cluster as seen in visible light. The presence of hot gas in regular clusters provides a natural explanation for the many trail radio galaxies, such as NGC 1265 (Figure 21.18g) in the rich Perseus cluster. Apparently these radio galaxies are really plowing through a gas that sweeps the radio-emitting region back behind the visible galaxy.

[Figure 22.9]

Intermediate clusters, whose general features are outlined in Table 22.4, are (as their name implies) intermediate between regular and irregular clusters. Of the types of galaxies found in intermediate clusters, over half are lenticular (S0) ones, the "armless spirals," with about 15 percent being elliptical galaxies and 30 percent spirals. They, like the regular clusters, tend to contain mostly ellipticals and S0s in their central regions, with spirals more toward the periphery of the cluster. The brightest galaxy or galaxies are usually normal-looking giant ellipticals.

Irregular clusters, the third category of clusters, should not be confused with irregular galaxies in Hubble's galaxy classification scheme. They are marked by an irregular shape with a low density of galaxies that show little tendency to concentrate toward the cluster's center. Spiral galaxies are more frequently found in irregular clusters, and they amount to over half of all galaxies in an irregular cluster. In fact, the brightest galaxies are generally spirals, and they are fairly uniformly spread from the center to the outer edge of a cluster. Radio and X-ray radiation have been detected coming from irregulars, but it is weaker than from regular clusters. Also, it seems to be associated with individual galaxies in an irregular cluster rather than with gas lying between them.

22.2.3. The Rich Cluster in Virgo

The nearest of the rich clusters, an irregular cluster, is that in Virgo, containing several hundred bright galaxies spread over a few tens of millions of light years and lying some 64 Mly from us (Figure 22.10). Its recessional velocity is about 0.3 percent of the velocity of light. About one hundred square degrees of sky is covered by the Virgo galaxies--roughly the area covered by a medium-sized book held at arm's length.

[Figure 22.10]

Most of the galaxies in the Virgo cluster are large normal spirals and dwarf elliptical systems. Its brightest members are giant ellipticals, such as M87 (also in Figure 21.18) and M84 (Figure 22.10). The X-ray emission from the Virgo cluster is concentrated around individual galaxies, particularly M84 and M86. The strong radio galaxy M87 in the Virgo cluster is also a strong source of X-rays. Another irregular cluster of galaxies, about 570 Mly from us, is the Hercules cluster (Figure 22.11).

[Figure 22.11]

22.2.4. Why The Differences Among Rich Clusters?

The masses of rich clusters of galaxies are on the order of 1015 Msun, being somewhat smaller for irregular than for regular clusters. Cluster luminosities range between 1012 and 1013 times that of the Sun, with as much as 1 percent in X-rays. As a result, the average mass-luminosity ratio is about 200, which indicates that clusters of galaxies have larger masses compared to light output than is typically true for individual galaxies (Table 21.2).

There is a strong concentration of elliptical galaxies toward the center of regular clusters, with a dominant supergiant elliptical galaxy located there. Several lines of evidence, including X-ray emission, point to the existence of a large mass of hot gas in the centers of regular clusters. This suggests that over the cluster's life the massive central galaxy has grown to colossal size by capturing gas, stars, and even whole galaxies from the other galaxies in the cluster. The hot gas filling the center is a by-product of this galaxy cannibalizing process. Thus the regular clusters are evolving more rapidly in comparison with irregular clusters, where change is much slower.

Assuming that all clusters are approximately the same age, why should one type of cluster differ from another, and why should one evolve more rapidly than the other? What appears to determine how much a given cluster has changed since birth, is both the density of matter at the time the cluster formed and how fast that matter was moving. A larger initial density of matter means that more and larger galaxies should form leading to more frequent collisions between member galaxies. Lower velocities lead to more vigorous collisions between members, so that high-density, low-velocity clusters should evolve dynamically more rapidly than low-density, high-velocity ones. Presumably then, high-density clumps of matter in the early Universe became regular clusters over time, while, low-density clumps stayed pretty much the same and are the irregular rich clusters of galaxies we see today.

22.2.5. Active Galaxies in Clusters

Active galaxies (discussed in the preceding chapter) play an important role in our understanding of the organization of galaxies into clusters and the evolution of cluster structure. For example, if active galaxies were found only among the 25 percent or so of galaxies that are not obviously members of clusters of galaxies, then astronomers would draw a very different conclusion about the evolution of galaxies and the Universe from that if active galaxies were found in clusters of galaxies. Evidence to date shows that active galaxies do occur in clusters, but whether they are in all, more, some, or just a few is not known for sure. This is because the most distant quasars are so much brighter than normal galaxies that a cluster of galaxies about a quasar is not easily detected.

Among nearer galaxies, about 20 percent of all strong radio galaxies are found in rich clusters, many of these being the supergiant elliptical galaxies that dominate the centers of regular clusters. Seyfert galaxies, N galaxies, BL Lacertae objects, and quasars, however, are relatively rare in rich clusters, but some of each type are located in small groups containing presumably normal galaxies. To give a sense of where active galaxies are located in distance, Figure 22.12 shows a histogram of the redshifts for typical samples of normal galaxies, radio galaxies, and quasars (Figure 22.13). In general, normal galaxies are nearby, radio galaxies are farther away, and quasars are very distant. Because of the look-back effect, the important point is that each one appears to be prevalent at a different age for the Universe.

[Figure 22.12]

[Figure 22.13]

22.3. The Missing Mass Problem

In 1933, Fritz Zwicky (1898-1974) noted that there did not appear to be enough mass in the form of galaxies to bind them gravitationally into a cluster. Zwicky's observation introduced the problem of "missing mass" into the study of clusters of galaxies (recall our earlier discussions about the missing mass problem in our own Galaxy).

Attempts to estimate the mass of an entire cluster of galaxies have led to conflicting figures. One common method of estimating the mass is to average the Doppler line shifts arising from internal motions of individual galaxies in a cluster. The spread of values about the average can be shown to be a measure of the cluster's mass. Another method of estimating the mass uses the observed luminosities of individual galaxies with their mass-luminosity ratios in Table 21.2. Knowing the number of different galaxy types in a cluster, we can convert the amount of light they emit to mass and add up the individual masses to find the cluster's mass.

The dynamical mass, found by analyzing the observed radial velocity differences, is many times greater than the luminous mass, obtained from the light emitted. For the latter, the mass derived includes only luminous matter within a cluster, but the former method includes matter that may be extraneous to individual galaxies in a cluster. The luminous mass of all galaxies in a cluster amounts to only 3 to 5 percent of the mass needed to provide gravitational stability. In other words, if the gravitational binding of a cluster were truly as weak as the luminosities of the galaxies in it suggest, then cluster galaxies would not be concentrated into a relatively tight unit that is a few million light years in diameter. Instead, cluster galaxies would disperse, spreading out over tens of millions of light years. This result does not depend on the classification of a cluster, but is true for all three classes of clusters.

In what form might this dark matter be? Perhaps a large amount of undetected matter exists in clusters as intergalactic material lying between galaxies, or as subluminous galaxies, or as a yet undetected dark halo component in cluster galaxies. As for intergalactic matter, it if were cool atomic hydrogen, then we should be able to detect it by its emission of 21-cm photons. If it were molecular hydrogen, there would be the possibility of detecting its ultraviolet spectrum with orbiting ultraviolet observatories. No evidence exists for extensive quantities of either atomic or molecular hydrogen in clusters that could account for this puzzling discrepancy.

A third possibility for intergalactic matter is that the dark matter is in the form of a very hot gas. A hot gas should emit X-ray photons but no 21-cm radiation. As noted in the preceding subsection, dozens of clusters are powerful X-ray sources. The richer the cluster, the greater is the X-ray emission. The source of X-ray emission is apparently hot intergalactic gas, but estimates are that the mass of hot gas is about equal to the mass of all the cluster's galaxies. Therefore, hot intergalactic gas is not sufficiently massive to prevent the cluster from eventually coming apart.

If the so-called missing mass in clusters were in the form of subluminous galaxies, there would have to be millions to billions of them for each bright galaxy, and they should be visible collectively as a faint glow spread over the entire cluster. Such an effect is not observed, so subluminous galaxies are not likely to be the answer to the missing-mass problem.

Finally, if the missing mass were to be found in extended halo containing dark matter, such as is suspected for our Galaxy, the large central galaxies of the cluster should have rapidly cannibalized the other galaxies earlier in the life of the cluster. In clusters where several large central galaxies exist, these large galaxies should have merged long ago because they would be 10 times more massive than they are thought to be. Hence clusters ought to appear very different from the way they do now. Although evidence grows that dark matter does exist in extended halos in some galaxies, extended galactic halos raise many perplexing questions if they are the location of the missing mass. The unanswered problem of the missing mass in clusters of galaxies is as important, if not more so, to our understanding of the Universe as any problem in astronomy today.

22.4. Superclusters, Clusters of Clusters

22.4.1. Local Supercluster in Virgo

As we just saw, clustering of galaxies seems to be the rule rather than the exception, about three quarters of all galaxies being in clusters. Thus it is natural to wonder whether there is any evidence for clusters of clusters. Within 20 Mly of us there are many small groups much like the Local Group. And about 64 Mly away is the giant Virgo cluster shown in Figure 22.10. The proximity to the Virgo cluster of some 50 nearby small groups, including the Local Group, and apparently isolated galaxies suggests that they all form an enormous flattened cluster of clusters, which has come to be known as the Local Supercluster. The Local Supercluster's center coincides with that of the Virgo cluster (Figure 22.14), and its equatorial plane is almost perpendicular to our Galactic plane. Its diameter is about 130 Mly, and its collective mass is estimated to be about 1015 Msun. The Local Group, which is near one edge of the Local Supercluster, appears to be revolving around its center at about 400 km/s.

[Figure 22.14]

22.4.2. Finding Superclusters

Although many details of the Local Supercluster's structure are still uncertain, the fact that it is a structural element in the Universe seems secure. One means of searching for other superclusters and an even larger hierarchy of clumping, if it should exist, is to plot locations on the sky of galaxies that are brighter than a chosen limiting apparent magnitude. Presumably what one should see projected on the plane of the sky are all galaxies in a certain volume of space centered on our Galaxy.

The most extensive such survey was conducted at Lick Observatory over a 12-year period and includes more than one million galaxies brighter than nineteenth magnitude. Figure 22.15 is based on that survey and shows a volume of space extending to about 500 Mly, or about five times the distance of the Virgo cluster. Even without distinguishing distances for different galaxies, the map clearly reveals a clumping of galaxies in the form of knots, clouds, and long filaments. On scales of hundreds of millions of light years galaxies are found in interlocking chains, but even more surprising are the empty regions, up to tens of millions of light years in size, in which there appear to be no galaxies.

[Figure 22.15]

To advance beyond the results of Figure 22.15, distances of galaxies must be known in order to find their true arrangement in space. From Hubble's velocity-distance law, if we know a galaxy's redshift, we can determine its distance. Before converting redshift to distance, one must know the value of Hubble's constant, a number not so well known. Redshift alone can suffice as a measure of distance, so that, much as we use Kepler's third law to construct a relative-distance scale for the Solar System, we can construct a relative scale for the Universe with just the redshift. And by measuring redshifts for galaxies brighter than a chosen limiting apparent magnitude, we can in principle determine the relative space distribution of galaxies for a volume of space set by the limiting apparent magnitude.

22.4.3. Coma, Perseus, and Hercules Superclusters

Such studies as those just discussed have been done and reveal the existence of several other superclusters. One such supercluster contains the Coma cluster of galaxies. The Coma cluster (about 30 Mly across; Figure 22.7) is apparently part of a vast complex of galaxies spanning some 200 Mly of space. The method of demonstrating this result is shown in Figure 22.16, in which redshift velocities are plotted against angular position on the sky out to more than 20o from the Coma cluster and all the way to the rich cluster A1367 in the constellation Leo. It seems evident that the two rich clusters, Coma and A1367, are members of the same supercluster and that a large cosmic void lies in front of it.

[Figure 22.16]

In a similar fashion, such diagrams have been done for regions around the rich clusters in Perseus (Figure 22.8) and Hercules (Figure 22.11). Figure 22.17 shows that the Hercules supercluster is about 330 Mly in diameter with the center of the supercluster some 700 Mly distant. The astonishing feature is the immense cosmic void, easily 300 Mly deep, in front of the supercluster. The void is as large as the Hercules supercluster itself.

[Figure 22.17]

The largest of the suspected cosmic voids is one lying in the direction of the north Galactic pole at a distance between 520 and 780 Mly. The void, if real, has a volume approaching 30 million Mly3. If the evidence can be substantiated, then the existence of such immense realms of space that are apparently empty has a most profound significance for the nature of the Universe.

22.5. Large-Scale Distribution of Matter in the Universe

When we study the structure of the Universe, we mean primarily the largest scale of structure in the Universe. At this cosmological scale, we are far beyond the level of stars in galaxies or, for that matter, galaxies themselves. Even the largest galaxies are at most a few million light years in diameter and can be considered as only a first level of cosmic clumping; a second level is clusters of galaxies, in which the richest are an order of magnitude larger than galaxies, or a few tens of millions of light years in size. A third level of clumping, the superclusters, are yet another factor of 10 larger in the cosmological scheme, on the order of a few hundred million light years across. Galaxies are thus at their largest only a small percentage of the size of superclusters. As evidence accumulates, it seems that superclusters may be part of an even larger organization of interlocking chains that begins to approach a billion light years in size. If real, then this may be a fourth level of cosmic clumping which could be called clusters of superclusters.

The cosmological scale is one of billions to tens of billions of light years, where superclusters participate in a general cosmic flow known as the expansion of the Universe. Moreover, on the cosmological scale, the superclusters are themselves only a few percent, the clusters a few tenths of a percent, and the largest galaxies a few hundredths of a percent of the cosmological scale. As an analogy, if the size of the visible Universe were to shrink to that of this page, a supercluster would then be about the size of a capital letter and any galaxy would be much smaller than a period.

Before discussing the cosmological-scale of structure, we ought to examine two observed phenomena that have proven to be of immense importance in helping us to conceptualize what the large-scale distribution of matter ought to be. These phenomena are the cosmic background radiation and the X-ray background radiation.

22.5.1. Cosmic Background Radiation

A consequence of the observed expansion of the Universe by Hubble is that the Universe must have been a lot smaller and much denser in the past than it is now. Squeezing all that energy and matter together early in the Universe's history means that in the beginning the Universe should have been packed into a hot, superdense state from which the expansion initiated, that process being referred to as the big bang. One should not think of the big bang as a state in which the Universe was confined to a small point in space from which it began expanding to fill space. Space does not exist independent of, nor does it precede, the Universe. Instead, the Universe began by filling all space, and it continues to fill all space, except that after the initiation of the expansion in the big bang space grows ever larger.

Following the big bang, the Universe was filled with an overwhelming abundance of high-energy photons, sometimes referred to as the "primeval fireball." Subsequent expansion cooled this radiation so that today most of its energy lies in the microwave region and in a background sea of neutrinos. In 1934, George Gamow (1904-1968) used such a scenario to predict the existence of this low-energy (or low-temperature) cosmic background radiation that permeates all space.

In 1965, the cosmic background radiation was discovered accidentally by physicists Arno Penzias (b. 1933) and Robert Wilson (b. 1936) at the Bell Telephone Laboratory. (For their discovery, Penzias and Wilson shared the Nobel Prize in physics in 1978.) The cosmic background radiation appears to be coming from all directions in space and to have about the same intensity from all directions. The discovery of this relic of the big bang is the most significant cosmological discovery since Hubble showed that the Universe is expanding.

Following 10 to 20 billion years of expansion by the Universe, the initial high-temperature radiation of the primeval fireball has cooled so that it now corresponds to blackbody radiation at a temperature only a few degrees above absolute zero. And, if the expansion was the same in all directions, the cosmic background radiation should be isotropically distributed--the same coming from all directions. And to a very high degree it appears to indeed be isotropic. The observed spectral distribution of the cosmic background radiation (its photons have a typical wavelength of 1 mm) is compared to a 3 K blackbody radiation curve in Figure 22.18. The part of the curve at the peak and toward short wavelengths in the range of 0.13 to 0.03 cm was first derived in 1975 from measurements made from a high-altitude balloon. (The short-wavelength side of the peak must be observed from outside Earth's atmosphere because water vapor and oxygen molecules absorb microwave radiation).

[Figure 22.18]

22.5.2. New Ether Drift Experiment

The cosmic background radiation can be used as a backdrop for determining the motion of Earth or the Solar System or the Galaxy. Unlike Michelson and Morley in their celebrated experiment to detect the motion of Earth relative to the nonexistent ether, contemporary astronomers have apparently succeeded in measuring Earth's drift in the sea of cosmic background radiation photons.

In the direction of Earth's motion the radiation should be slightly hotter, or of shorter wavelength, as a result of the Doppler effect; in the opposite direction, it should be slightly cooler, or of longer wavelength. Consequently, there should be a slight departure from isotropy (anisotropy) in the cosmic background radiation resulting from Earth's motion. What has been observed is a minute anisotropy recorded as a temperature difference of approximately 0.0035 K from its average value. The maximum (hottest) and minimum (coolest) values are in the directions of the constellations Leo and Aquarius, respectively. The data in this experiment suggest that our Galaxy and the Local Group of galaxies are traveling at a velocity between 300 and 600 km/s toward a point in the constellation Hydra some 45o from the center of the Virgo cluster, the center of the Local Supercluster. These are remarkable results in that the isotropic sea of cosmic background photons can provide a backdrop against which motion in the Local Supercluster can apparently be measured.

22.5.3. X-ray Background Radiation

The second observed phenomena suggesting that the Univserse is isotropic is the X-ray background radiation. With the advent of rocket-borne X-ray telescopes, it was found that a diffuse background of X-rays are coming from every direction in space. This radiation is in addition to the many discrete sources of X-rays in our Galaxy, other galaxies, and clusters of galaxies. The uniformity of this radiation over the sky is a strong argument that it could not be from our Galaxy but must be coming from far beyond. Some astronomers have suggested that this background radiation is telling us about an all-pervading gas that contains more mass than all the galaxies put together. The thermal spectrum of background X-rays corresponds to temperatures of several hundred million degrees for the supposed gas.

There is an opposing view that insists that the X-ray background radiation is just the sum of very many faint, unresolved, discrete sources, such as quasars. Deep sky surveys, such as that by the Einstein Orbiting Observatory, seem to find increasingly larger numbers of discrete X-ray sources. As a result, it seems possible to account for at least a third of the diffuse X-ray background by supposing the radiation to be due to a large number of quasars at immense cosmological distances. With better technology coming in the near future it may be possible to show that the Universe is not pervaded by hot gas but that all the X-ray background is the product of a vast number of quasars spread through distant space that once dominated the early Universe. The important point for us is that this phenomenon confirms the conclusion from the cosmic background radiation that on the largest scale the Universe is surprisingly isotropic, that is, it is the same in every direction spanning a scale of approximately 30 billion ly.

22.5.4. How Large Are Superclusters?

Redshift velocities for the Perseus, Coma, and Hercules superclusters tend to group around 5000, 7000, and 11,000 km/s. If, as most astronomers do, one assumes that galaxies, small groups, or clusters having redshift velocities in the neighborhood of these three values are physically related, then these three superclusters become only part of even larger structures, which approach a substantial fraction of a billion light years in length. And rather than being spherical in shape, they are more like long filaments, with the Coma and Perseus superclusters as dense knots along interlocking chains and sheets, with vast cosmic voids separating the various strands and sheets.

There are three analogies we might use to derive a mental picture of the structure of the Universe on scales less than a billion ly: islands of matter isolated in a great void, a sea sponge in which matter and voids form one interconnected structure, and finally swiss cheese in which the voids are isolated pockets in an otherwise continuous structure. Which of this models most nearly fits what astronomers think they are finding? At present, the structure seems clearly to the more like a sponge than either islands or swiss cheese (Figure 22.19). Having this model in mind one wonders whether such structure is more local or is there reason to believe that throughout all space and over all cosmic time such structure has existed.

[Figure 22.19]

Since the superclusters in Perseus, Coma, and Hercules are within 1 billion ly of our Galaxy (cosmological redshifts less than 0.05), they could be a relatively recent feature in the Universe. What evidence exists that superclusters were part of the early Universe? It has recently been pointed out that several pairs of quasars with nearly equal redshifts lie very close to each other on the sky. This suggests that they are physically associated, and for their observed redshifts they are within 15 to 100 Mly of each other in space. Such separations suggest that the physical association is provided by the fact that they are both members of a supercluster. Thus all quasars could be members of superclusters. Those pairs with redshifts of about 4 are seen as they existed some 15 billion years ago, or part of the early Universe. Although this is a very indirect argument that superclustering has been a common feature throughout the time span of the Universe, it is still an important consideration in understanding the nature of the Universe.

Until more redshifts have been measured for galaxies, it will be difficult to resolve many of the questions surrounding the cosmic design.

22.5.5. Large-Scale Distribution of Matter

In 1692, Isaac Newton describes in a letter to a friend what can be expected of gravity in shaping the Universe. Although there are now additional considerations to the cosmological question of which Newton was not aware, his statement is an argument for an infinite, centerless, and edgeless Universe that can be unstable on a local scale but must be stable and homogeneous on the cosmological scale. Gravity tends to pull matter into clumps, making any irregularities in the Universe more irregular. Thus like a box within a box within a box, gravity could work on the smallest scales, clumping up matter, and then advance to the next largest scale, and the next, and so on. Finally, if given long enough and nothing opposes it, gravity could produce the superclusters and the large-scale sponge-like patterns of cosmic space (Figure 22.19). But has gravity actual done such a thing or is it more likely that the sponge structure is a remnant of the formation of the Universe. At this point, we only have vague suspicisons of the answer to that question which we will discuss in the next two chapters.

Let us summarize again what we know about the structure of the Universe. The distribution of matter on sizes of several hundred million light years resembles a sea sponge. This is apparently how matter is distributed within 1 billion or so light years of us, and it is apparently the same regardless of the direction in which we look. But on the largest scales in the Universe, the cosmic background radiation and the X-ray background radiation both suggest the Universe is surprisingly isotropic and homogeneous and that no structure exists on the cosmological scale of billions to tens of billions of light years. Finally if the Universe is indeed as uniform on the largest scales as it appears, then any large scale curvature of space as a consequence of general relativity must be uniform throughout space and also Hubble's velocity-distance law is applicable everywhere in the Universe.

Copyright 1995 J. C. Evans
Physics & Astronomy Department, George Mason University
Maintained by J. C. Evans;