l10body

1. INTRODUCTION:

This lesson supplements Ch. 24 of your text, which you should read carefully. Section 8, on gravitational lensing, is discussed in Ch. 25, pp. 555 - 557.

Our solar system belongs to the Milky Way galaxy (Lesson 9), an "island universe" containing more than 200 billion stars. But the Milky Way is only one of many billions of galaxies strewn throughout the universe.

This lesson is devoted to galaxies: how they were discovered; their variety of sizes and shapes; their masses and content; their distances and distribution throughout cosmic space; and their motions and interactions.

Galaxies are beautiful, and you can find many images on the WWW. Here are some of the best places to look:
Search results for galaxies from Astronomy Picture of the Day;
Images of Galaxies by David Malin of the Anglo-Australian Observatory;
Images of Galaxies from the Hubble Space Telescope; and
Ultraviolet Images of Spiral Galaxies from the Ultraviolet Imaging Telescope.

2. HISTORY:

The idea of a galaxy has existed since 1755, and astronomers have been observing galaxies for almost as long. But it was only in the late 1920s that astronomers finally found convincing evidence that galaxies were really other "island universes" similar to the Milky Way. This history is summarized below.

NOTES ON THE DISCOVERY OF GALAXIES
1750	Thomas Wright published "An Original Theory of the Universe" in which he explained the Milky Way by proposing that the stars were distributed in a thick disk -- the first step in the discovery that the Milky Way is a galaxy.
1755	The great philosopher Immanuel Kant, inspired by Wright's ideas, proposed that the Milky Way was only one of many galaxies scattered throughout an infinite universe.
1781	Charles Messier published Messier Catalogue of 110 "nebulae", of which 40 are galaxies. Thus, the brightest and biggest galaxies are often labeled by their number in the Messier Catalogue - e.g., M31, the Andromeda Galaxy.
1800	William Herschel published his Comprehensive Catalogue of Images of 2500 deep sky objects.
1864	John Herschel (son of William Herschel) published the General Catalogue of Nebulae and Clusters of Stars, which evolved into the New General Catalogue of more than 10,000 galaxies. Therefore, you will often see a galaxy labeled by its number in the New General Catalogue. For example, M31 is also called NGC224.
1918	From observations of globular clusters and the RR Lyrae variable stars in them, American astronomer Harlow Shapley proved that the globular clusters were distributed in a roughly spherical system, centered, not on the Sun, but at a point at a distance of some 15,000 parsecs. (Shapley overestimated the distance; we now know it is closer to 8,000 parsecs.) This was compelling evidence that the Sun is not at the center of the Milky Way.
1912-1920	American astronomer Vesto Slipher, working at the Lowell Observatory in Flagstaff, Arizona, began to measure the Doppler shifts of spectral lines from spiral galaxies. He found that the vast majority of galaxies were moving away from the Milky Way (except the Andromeda galaxy M31, which was moving toward the Milky Way).
1924	Working at at the great new 2.4-m telescope at Mt. Wilson, California, Edwin Hubble began a systematic survey to measure the distances and Doppler shifts of spiral galaxies, following Slipher's work.
1925	Harlow Shapley and Heber Curtis held a famous debate about the nature of the "spiral nebulae". Shapley argued that they were gas clouds belonging to the Milky Way; Curtis argued that they were other galaxies, as big as the Milky Way itself and at great distances. The debate did not resolve the controversy, but shortly thereafter, Edwin Hubble and others found novae and Cepheid variable stars in nearby galaxies. In 1930, Shapley conceded that the spiral nebulae must really be other galaxies. Kant's hypothesis of "island universes" was right.
1929	Hubble published his observations showing that the expansion of the universe obeyed the equation V = H₀D, now known as "Hubble's Law". He underestimated the distances of the galaxies, thus derived a value of the "Hubble Constant", H₀ = 500 km/s/Mpc, that is about ten times the modern value.
1998	Today, we know the Hubble Constant to an accuracy of about +/-15% (H₀ = 50 - 70 km s^-1 Mpc^-1). Only a few years ago, the uncertainty of was greater than a factor of two (H₀ = 50 - 100 km s^-1 Mpc^-1). Thanks to the Hubble Space Telescope, we can now see galaxies so distant that their light has been traveling to us for more than 10 billion years -- 80% of the time since the universe was formed. You can see some of these galaxies in the Hubble Deep Field.

That's enough to remember. But, if you are interested, you can find a more detailed summary in this History of the discovery of deep sky objects.

3. TYPES: (See text, Section 24.1)

Galaxies are classified into three basic different types: Spirals (S), Ellipticals (E), and Irregulars (Irr). Examples can be found here: Galaxy types. The largest galaxies tend to be either spirals (about 80%) or ellipticals (about 20%).

Like the Milky Way, spiral galaxies are composed of a central bulge and halo of Population II stars and globular clusters and also a disk of gas and young Population I stars. (You can review Populations I and II in Lesson 9.) Some spiral galaxies have a greater fraction of their mass in the bulge/halo population and some are mostly disk population. Just as in the Milky Way, the spiral arms contain dense clouds of gas where new stars are forming.

Elliptical galaxies have no Population I stars, no spiral arms, no cool interstellar gas, and no star formation. They are dead galaxies. If spiral galaxies are like rain forests, with cool life-giving interstellar clouds, elliptical galaxies are like deserts, with hot dry winds and little life.

Irregular is a catch-all definition for all sorts of galaxies that cannot easily be classified as spirals or ellipticals. Many irregular galaxies are the result of interactions between galaxies (see below). Many of the smaller "dwarf galaxies" are irregulars.

Edwin Hubble developed a scheme (see text, Fig. 24.9) to classify galaxies according to their shapes. He divided spiral galaxies into normal spirals (denoted S) and barred spirals (denoted SB). A third letter (0, a, b, c) denotes how tightly wound the spiral arms are, ranging from so tight that the arms can hardly be discerned (S0 and SB0) to a relatively open pattern (Sc and SBc). Likewise, Hubble classified elliptical galaxies according to their shapes, ranging from spherical (denoted E0) to very flattened (denoted E7). You can see plenty of examples of galaxies classified according to Hubble's scheme here: Hubble Types.

Like the Milky Way, galaxies look very different depending on the wavelength band of the electromagnetic spectrum in which they are observed, as illustrated here: Multiwavelength atlas of galaxies.

4. DISTANCES: (See text, Sec. 24.2)

Measuring the distances of galaxies is one of the most challenging problems of observational astronomy. We have already mentioned this problem in Lesson 9 in the context of measuring the size of the Milky Way. In the Milky Way, we measure distances to about 1000 parsecs by measuring stellar parallaxes. But, to measure distances beyond 1000 parsecs, astronomers must rely mainly on the inverse square law of light. (There are a few important exceptions, such as SN1987A, where one can infer the distance from the observed expansion rate, as you did in homework 4.) I repeat the idea here, since it is so important. If absorption by intervening dust clouds is unimportant, the brightness (B), is related to luminosity (L) and distance (D) by the inverse square law:

B = L/(4p D²)

Thus, if we knew the luminosity of a star (or some other astronomical source) and we measured its brightness, we could solve this equation to determine its distance. The problem is: how can we know L? It turns out that there are certain types of sources for which we can infer L from some other observed property. We call these sources standard candles.

We have already discussed two examples of this method. One is "main sequence fitting" (see Lesson 4). If we know that a star is a main sequence star, one can infer its luminosity from its spectral type (surface temperature). Astronomers can use this method to measure distances in the Milky Way and the distances of the nearest galaxies -- out to a distance of about 200,000 light years. But beyond that, astronomers can only see the most luminous red and blue giants, and such stars are not likely to be main sequence stars.

Fortunately, nature has provided a more luminous kind of standard candle: Cepheid variable stars. Cepheid variables are not main sequence stars -- they are probably helium core-burning stars. But they pulsate regularly, and their luminosities are related to their pulsation periods by the famous period-luminosity relationship, originally discovered by Henreitta Leavitt. We can infer the luminosities of such stars by measuring their pulsation periods (see Lesson 9). With the Hubble Space Telescope we can observe Cepheid variables in galaxies at distances of about 100 million light years. See Cepheid Variables in M100.

Beyond 100 million light years, Cepheid variables are too faint to observe, even with the Hubble Space Telescope. But again, nature has kindly provided a luminous source that appears to be a reliable standard candle: Type Ia supernovae (see Lesson 7). In the past few years, astronomers have discovered that the maximum luminosity of a Type Ia supernova is correlated with the rate at which its light dims: the slower the dimming rate, the more luminous the supernova is at maximum. So, if astronomers can find a supernova in time to see its maximum luminosity, observe its spectrum (to make sure it is Type Ia), and watch its decay rate, they know its maximum luminosity. Since they observe its maximum brightness, they can use the inverse square law to infer the distance of the galaxy to which it belongs.

Two modern developments in astronomical instrumentation have made it possible for astronomers to do this. One is the development of wide-field CCD cameras on large ground-based telescopes. With such cameras, astronomers can obtain images of an area of sky containing many galaxies. Comparing such images of the same region of the sky, taken a week or so apart, they can find a dozen or so supernovae. Then, with very large telescopes such as the Keck telescope (the second development), they can obtain spectra of all of these supernovae and find those of Type Ia.

At maximum light, type Ia supernovae are so luminous that they can be seen at distances of 10 billion light years or more -- almost to the edge of the observable universe. They appear to be the most accurate standard candles by which we can map the distant universe. Because this technique is so promising, two international teams of astronomers are devoting a considerable fraction of observing time on the world's largest telescopes to observing them. Some of the results are described here: Hubble Pinpoints Distant Supernovae.

Only a few years ago, the distances to distant galaxies were uncertain to about a factor of 2 (e.g., we could only say that a certain galaxy must be at a distance between 1.5 and 3 billion light years). But with the techniques described above, enabled by the Hubble Space Telescope and large ground-based telescopes, we can determine the distance of a galaxy with an error of about +/- 15% (e.g., between 1.7 and 2.3 billion light years). That's a big improvement.

It's important to realize that the yardsticks by which we determine cosmic distance all depend on each other. For example, the period-luminosity relationship of the Cepheid variable stars must be calibrated by observing Cepheid variables in nearby galaxies (such as the Large Magellanic Cloud), whose distance can be determined by main sequence fitting. Likewise, the maximum luminosities of Type Ia supernovae must be calibrated by observing such supernovae in galaxies whose distances can be determined by observing Cepheid variables in them -- see Measuring the Expansion Rate of the Universe. Each of the techniques by which we develop reliable yardsticks to measure cosmic distances is a step on the cosmic distance ladder, illustrated in Fig. 24.10 of your text. (Important note: the limits of useful range given in Fig. 24.10 are out of date, and way too small in some cases. The numbers I have given here are more accurate.)

I have described four steps to measure cosmic distances: (1) parallax; (2) main sequence fitting; (3) Cepheid variable stars; (4) Type Ia supernovae. There are several other techniques for inferring cosmic distances, some fairly good and others rather unreliable and outdated. Your text (p. 521) describes another good one -- (5) the Tully-Fisher relation -- by which astronomers can infer the infrared luminosities of spiral galaxies from their rotation speeds. With redundant techniques, astronomers can cross check the various techniques against each other to see which ones are most accurate and reliable. That's important, because the uncertainty in the cosmic distance scale depends on the accumulated uncertainties from every step in the ladder.

Prof. Ned Wright of UCLA gives a fairly complete summary of all the methods to measure cosmic distances in The ABC's of Distances. You don't need to know all these methods, but you may wish to review the five that I have mentioned above.

5. HUBBLE'S LAW: (See text, Section 24.4)

In 1912, Vesto Slipher was observing the spectra of galaxies at the Lowell Observatory in Flagstaff, Arizona. He made the remarkable discovery that the spectral lines of most galaxies are shifted to the red, indicating that the galaxies are moving away from the Milky Way. (One notable exception is the nearby giant spiral galaxy M31 in Andromeda, which is actually moving toward the Milky Way.) Then, in 1929, Edwin Hubble, working with the new 2.4 meter telescope on Mt. Wilson, near Los Angeles, made an even more remarkable discovery. He plotted the recession velocities of galaxies (V, inferred from their redshifts) versus their estimated distances (D), in what we now call a Hubble diagram (see below).


Hubble's Data	Modern Data
Hubble Diagrams (from Ned Wright's cosmology tutorial)

When Hubble did this, he found that the recession velocity of the galaxies tended to increase with their distance. Fitting a straight line to the data, he wrote down a formula that we now call Hubble's Law:

V = H₀D.

This simple equation is one of the most important formulas in all astronomy (indeed, in all science). The number H₀ is called Hubble's constant. H₀ is usually expressed in units km/sec/Mpc, so that the formula gives the recession velocity V of the galaxy in km/s if the distance D is expressed in units of Mpc (1 Mpc = 1 million parsecs = 3 million light years).

When Hubble discovered this relationship, he had no reliable way of measuring the distances of galaxies. He only guessed their distances from their sizes and brightness -- and he was way off! He underestimated their actual distances by almost a factor of 10. From such estimates, he inferred a value H₀ = 500 km/s/Mpc.

To find the Hubble constant, astronomers must measure both the recession velocities and distances of many distant galaxies. It's easy to measure the recession velocities to a precision better than 1% from the redshifts of their spectral lines. The hard part is to measure the distances of galaxies, as we have discussed in the previous section. A few years ago, the distances of galaxies were uncertain to a factor of about two; but today, we believe that we can measure distances to an accuracy of +/- 15%. As a result of such measurements, we now believe that H₀ lies in the range 50 - 70 km/s/Mpc.

Hubble's Law tells us something profound: the entire universe is expanding, as if it began in a huge explosion! Moreover, the value of the Hubble constant tells us the age of the universe since this explosion. We'll have a lot more to say about the implications of Hubble's Law in Lesson 12. But for now, we'll merely take it as an empirical fact and consider how it can be used to determine the distances of any galaxy. All we need to do is to take a spectrum of the galaxy and measure the wavelengths, l(apparent), of its spectral lines. Then we can infer its recession velocity V from the Doppler formula:

l(apparent)/l(true) = [(1 + V/c)/(1-V/c)]^1/2,

and calculate its distance from V and Hubble's Law. (This equation is exact. The simpler equation on p. 69 of your text is approximate, and not accurate when V is close to c.)

If we trust that Hubble's Law is a universal law of nature, we can use it to measure the distances of galaxies even if we can find no standard candles (such as Type II supernovae or Cepheid Variable stars) in them. Thus Hubble's Law becomes a powerful tool for mapping the distribution of galaxies in the universe.

6. CLUSTERS OF GALAXIES:

Galaxies are not distributed uniformly throughout space -- most are clumped together in clusters and superclusters that are evidently held together by gravity.

In fact, our own Milky Way galaxy belongs to a small cluster of galaxies, called The Local Group (text, p. 520), which contains two large spiral galaxies (the Milky Way and M31) and more than 30 other galaxies, most of them much smaller than the Milky Way. You can find a diagram here and a complete listing of its galaxies here.

The nearest giant cluster of galaxies is the Virgo Cluster of galaxies. It contains more than 2,000 galaxies, many of them larger than the Milky Way. The distance of the Virgo Cluster from the Milky Way is still somewhat uncertain, but a good estimate is about 50 million light years. Beyond the Virgo Cluster is an even bigger cluster of galaxies, the Coma Cluster of galaxies, at a distance of approximately 300 million light years. It contains some 6,700 galaxies.

Astronomers have found thousands of clusters of galaxies. You can find images (both optical and X-ray) of several in Galaxy cluster mug shots from the University of Arizona.

Astronomers have discovered a very important fact about the rich clusters of galaxies: they are powerful X-ray sources. The X-rays are emitted by hot (temperature of a few millions of degrees) gas that fills the space between the galaxies. Compare the optical and X-ray images of clusters in Galaxy cluster mug shots. Also have a look at Hot Gas and Dark Matter. The mass of the X-ray emitting gas in a rich cluster is comparable to the net mass in all the galaxies. But, as we shall see in the next section, the clusters are held together by an even greater mass of "dark matter".

Although about 80% of all large galaxies are spirals and 20% are ellipticals, the proportions are reversed in rich clusters. We know two reasons why this should be so. First, the galaxies in rich clusters are all moving through the hot intergalactic gas in the cluster at speeds of hundreds of kilometers per second. This hot gas acts as a powerful wind blowing through the galaxy, which will blow away any gas produced by stars and supernovae. This hot wind prevents the recycling of gas from stars to the disk of the galaxy that replenishes the interstellar medium of spiral galaxies. The second reason is that galaxies in rich clusters collide with each other, and galaxies that have suffered many collisions tend to become elliptical in shape.

Superclusters: Most galaxies do not reside in rich clusters like Virgo or Coma -- they reside in small groups like the local group. But these groups are not uniformly distributed throughout the universe. Instead, individual galaxies, small groups, and rich clusters are distributed in a kind of "foamy" texture consisting of huge sheets and filaments of galaxies separated by large "voids" where very few galaxies are found. The typical diameters of the voids are 50 - 100 million light years. You can see some of this texture in this all-sky map of The Nearest 15,000 Galaxies. But this map gives no information on the relative distances of these galaxies.

What astronomers really need is a 3-dimensional map of the universe that shows not only the locations of the galaxies in the sky, but also their distances. They can do that by taking spectra of thousands of galaxies and then inferring their distances from the redshifts of their spectral lines and Hubble's Law. Many groups of astronomers are conducting such redshift surveys. They are huge efforts. Up to now, only a few strips of the sky have been mapped in this way, and not too deeply; but we can already see that they are telling us something very important about the texture of the universe. Be sure to study Galaxy Clusters and Large Scale Structure, where you can see one of the best examples, the CfA redshift survey.

The most ambitious such project is the Sloan Digital Sky Survey. A new 2.5 meter telescope, nearing completion at Apache Point, New Mexico, is especially designed to observe more than 100 million galaxies and measure spectra of more than 1 million of them. The project will take about a decade to complete, using this telescope full-time.

7. MASSES OF GALAXIES AND CLUSTERS:

Astronomers have four basic methods to measure the masses of galaxies and clusters: rotation curves, random velocities, X-ray emission, and gravitational lensing.

Rotation curves: In astronomy, we often infer masses from orbits. Thus, we infer the mass of the Earth by applying Kepler's 3rd Law (as understood by Newton) to observations of the orbit of the Moon. Likewise, we infer the mass of Jupiter from the orbits of its moons; the mass of the Sun from the orbit of the Earth; the mass of stars from the orbits of binary stars about each other.

By exactly the same logic, astronomers can infer the mass of the Milky Way galaxy from the (more-or-less) circular orbits of its stars and gas, as we described in Lesson 9. But when they did, they found that the Milky Way has about five times as much mass as the mass in stars and gas that astronomers can detect through radiation. Astronomers call this unseen matter dark matter, and they call the ratio of dark matter to visible matter the Mass-to-Light Ratio (M/L). (More precisely, M/L is the mass of the system in solar units divided by its light in solar units. Thus, the Sun has M/L = 1.) The Milky Way has M/L = 5. Likewise, other spiral galaxies have values of M/L ranging from 5 to 50.

In fact, by measuring the rotation curves of a spiral galaxies we can not only infer its total mass, we can infer how this mass is distributed throughout the galaxy. This concept is demonstrated very nicely in the page Stronger Evidence: Rotation Curves of Galaxies. Be sure to play with the Java experiment there, and notice how the motions of the stars change when you change the distribution of mass within the galaxy. If the mass of the galaxy is concentrated toward its center, the interior stars move faster. (Just as the inner planets of the solar system move faster than the outer planets.) But if the mass of the galaxy is spread out, the orbital velocities of the stars can actually increase with radius.

The most accurate method to measure the rotation curves of galaxies is to observe the Doppler shift of the 21-cm line emitted by the interstellar hydrogen gas with radio telescopes.

We find that rotation velocities of spiral galaxies at large distances from the center of the galaxy are much greater than they would be if the mass were distributed in the same fashion as the luminous stars. Indeed, we find that the dark matter in spiral galaxies must be distributed in a more-or-less spherical shape that extends far beyond the distribution of stars, as illustrated here.

In homework 5, you will use this method to measure the mass distribution of a spiral galaxy.

Random velocities: Unlike spiral galaxies, elliptical galaxies don't rotate rapidly. The stars move in random orbits, not circular orbits. Even so, astronomers can measure the masses of elliptical galaxies by a technique that is similar to measuring rotation curves of spiral galaxies. They measure the Doppler shifts of the spectral lines of many stars to calculate average random velocities and distances from the center of the galaxy. Then they can infer the mass of the galaxy from an equation very similar to Kepler's Third Law. The result is similar to that for spiral galaxies: elliptical galaxies have M/L = 5 to 50.

By exactly the same logic, astronomers can infer the masses of clusters of galaxies by measuring average random velocities and orbital distances of many galaxies from the center of the cluster. In fact, the first evidence for dark matter in the universe was found by this technique in 1933 by the astronomer Fritz Zwicky. This method is described in Dark matter by UC Berkeley and in The Evidence for Dark Matter by Jonathon Dursi of Queens University (with another nice Java demonstration). Rich clusters have even higher mass-to-light ratios than galaxies -- in some cases as high as M/L = 300. That fact implies that there is more dark matter between the galaxies in the clusters than there is in the galaxies themselves.

X-ray emission: Another way to infer the masses of elliptical galaxies is to observe the X-ray emitting gas in the galaxy. Since this gas is very hot, it would naturally tend to expand and escape from the galaxy. But it doesn't because it is bound by the galaxy's gravity, which depends on the galaxy's mass. Therefore, by measuring the temperature and radial distribution of the X-ray emitting gas, we can infer the mass. Actually, the method depends on the same physical principle as the method of measuring random velocities of stars; the main difference is that we are measuring the random velocities of the atoms in the hot gas when we measure the X-ray temperature.

You can see a fine example of X-ray emission from a giant elliptical galaxy in this optical and X-ray image of M87. The image on the right is an optical picture of the Virgo cluster of galaxies, with the giant elliptical M87 in the center (the caption says the optical image is on the left, but that's obviously a mistake). The image on the left is an X-ray picture of the same cluster, showing that there is a huge cloud of X-ray emitting gas centered on M87 and much larger than the optical galaxy. M87 itself is about 10 times as massive as the Milky Way. By analyzing the X-ray image, astronomers can see that the halo of dark matter around M87 has five times as much mass as M87 itself.

In the few cases where we can compare the mass of dark matter in clusters of galaxies as inferred from the random velocities of the galaxies with that inferred from the X-ray emission, the results agree. That fact provides a good check that the results are reliable, at least within the error margins of the measurements (about a factor of 2).

Up to now, the technique of measuring the masses of clusters of galaxies through their X-ray emission has been limited by the inability of X-ray telescopes to measure their images and spectra simultaneously. But that situation will improve very soon, with the launches of three powerful new X-ray telescopes: NASA's AXAF (November 1998); the European Space Agency's XMM (August 1999); and the NASA/Japan ASTRO-E (early 2000). These X-ray telescopes are far more powerful than any before, so we can expect to know a lot more about the distribution of dark matter in clusters of galaxies very soon.

Gravitational lensing: Astronomers have found yet a third way of measuring the distribution of dark matter. That is by measuring the deflection of light rays by gravity. This phenomenon is described below in section 8.

The most important conclusions from these studies are: (1) galaxies and clusters of galaxies -- indeed, the universe itself -- contain approximately 10 times as much dark matter as visible matter; and (2) the dark matter in galaxies and clusters extends over much greater distances than the luminous matter.

We don't know what this dark matter is. We discussed some possibilities in Lesson 9. But here, we see that the dark matter is everywhere, not just in the Milky Way. As you will see, the amount and distribution of dark matter controls the evolution of the universe -- not only the development of galaxies in the universe, but also its ultimate fate. Therefore, understanding the nature of the dark matter is one of the most important questions in all science.

8. GRAVITATIONAL LENSING:

We have already introduced this idea in lesson 8, where we discussed the distortion of light around a black hole, and in lesson 9, where we discussed the search for MACHOs. You should check those links. But the clearest evidence for gravitational lensing comes from images of galaxies lying behind clusters of galaxies. First, review gravitational lensing. It explains why a galaxy lying directly behind another galaxy will have its image distorted by gravity into the shape of a ring, called an Einstein ring. This actually happens: here's the best example seen so far: A Bull's Eye for Merlin and the Hubble.

But if the galaxy does not lie directly behind the gravitational lens, you will see two images, one smaller and one larger. The point is illustrated very well in this amusing film clip, The Real Einstein Ring, where you see two images of the lensed galaxy (Einstein), one bigger and one smaller, unless Einstein lies directly behind the lens, in which case the two images merge into a ring.

If the lensing object is not spherical, it can bend the light in more complicated ways, forming multiple images of the background object. Here's an example of an Einstein Cross seen by the Hubble Space Telescope, in which the background galaxy is broken into four images.

With the Hubble Space Telescope, astronomers have found several examples of distant galaxies seen through the gravitational lens due to an intervening cluster of galaxies. In that case, the bending of light due to the complicated distribution of mass in the cluster can create several images of each galaxy. The most spectacular example is shown below. (Click on the images to see larger versions.)


Cluster CL0024 and multiple lensed images of background galaxy (inset on lower left)	Mass distribution of CL0024 (courtesy of Tony Tyson, Lucent Technologies)

In the image on the left, taken by the Hubble Space Telescope, we see a cluster of reddish galaxies at a distance of about 5 billion light years. We also see several arc-shaped images of blue galaxies. Eight of them are loop-shaped: they are all different images of the same galaxy! It is a ring shaped galaxy lying behind the cluster, at a distance of about 10 billion light years. Each image is distorted in a different way because the light rays have been bent by the mass from different galaxies in the intervening cluster. The inset in the lower left is a (magnified) mathematical reconstruction of what the actual galaxy would look like without the distortions. (There are several more blue images from other background galaxies.)

The distortions of these multiple images carry a tremendous amount of detailed information about the gravitational fields -- hence the mass distribution -- in the intervening cluster of galaxies. The image above on the right is a mathematical reconstruction of the mass distribution in the cluster that is required to cause the observed multiple images. Observations of lensed galaxies may be the most powerful method that astronomers have for mapping the distribution of dark matter in clusters of galaxies.

Figure 25.24 of your text shows two other examples of multiple images of background galaxies caused by gravitational lensing by a cluster of galaxies.

Besides providing a method for measuring dark matter, gravitational lenses have another big advantage: they can make astronomical objects appear much more luminous than they really are. For example, in July 1997, a press release by the Space Telescope Science Institute announced the discovery of the Farthest galaxy in the Universe . This galaxy was discovered behind a cluster of galaxies by the Hubble Space Telescope, and the 10-m Keck Telescope in Hawaii found that it had a redshift of 4.92, making it the most distant object seen in the universe at that time. That galaxy would have been too faint to find but for the fact that the lensing action of the cluster made it much brighter than it would have been otherwise.

But records are made to be broken, as you can (should) see here. Astronomers are pushing the frontiers of the universe fast!

9. COLLIDING GALAXIES:

Galaxies in clusters are close enough and big enough that they often pass nearby each other, as illustrated here. In that case, the gravity due to a passing galaxy can perturb the orbits of the stars about the center of the other galaxy, causing great tidal distortions of the shape of each galaxy. In fact, we now know that these tides can easily cause new spiral arms to form in a disk galaxy. Hundreds of millions of years thereafter, the spiral arms may wind up more and more tightly, as the galaxy changes its morphology from type Sc to Sb to Sa.

For example, the spiral arms of the Milky Way may be a result of gravitational tides due to its satellite, the Large Magellanic Cloud. A famous example of a galaxy in which spiral arms are caused by a passage of a satellite galaxy is the "Whirlpool Galaxy" M51.

In fact, sometimes galaxies will "collide," passing right through each other. I put "collide" in quotes because it is unlikely that any stars will actually hit each other during such collisions, since interstellar distances are so great. But the collisions will greatly disturb the orbits of the stars, leaving highly deformed galaxies.

If one galaxy is much more massive than the second one, the smaller galaxy may merge into the larger one. When that happens, the larger galaxy will swell up somewhat because it is "heated" as its stars redistribute the energy of the high speed stars from the other galaxy. The absorption of smaller galaxies by larger galaxies is called cannibalism. Such cannibalism occurs commonly in large clusters of galaxies and is responsible for the existence of very massive galaxies often found at the center of such clusters. A good example of a big fat cannibal that has obviously eaten many smaller galaxies is the giant elliptical M87 in the Virgo Cluster of galaxies, which we already discussed above in the context of its X-ray emission.

Although the stars in galaxies miss each other when galaxies collide, the interstellar gas in the two galaxies, which fills the space between the stars, must crash. When it does, the gas is compressed and this compression can trigger a huge outburst of star formation called a starburst. Here's a spectacular picture of starburst activity in a famous pair of colliding galaxies called the "Antennae": Hubble Reveals Stellar Fireworks Accompanying Galaxy Collisions. Be sure to look at the animations there.

Another interesting example of star formation caused by collisions of galaxies is the Cartwheel Galaxy. In this case, a smaller galaxy passed almost straight through the center of the larger one, sending an expanding circular compression wave through its interstellar gas. New stars are being formed in this big "splash".

Because star-forming regions contain many luminous blue and red supergiant stars, starburst galaxies are exceptionally luminous, especially at infrared wavelengths. They are more common in the early universe, when galaxies were colliding more often than they are today.

More than 300 years ago, Isaac Newton wrote down the equations we need to describe the motions of stars and planets due to their mutual gravity, and his solutions describe the orbits of the planets and comets very well. We can use these same equations to describe the collisions of galaxies. But it becomes much more difficult to solve Newton's equations when we must describe the motions of billions of stars all tugging on each other instead of just one planet tugged by the gravity of the Sun. But today, theoretical astrophysicists have developed techniques to simulate collisions of galaxies by solving such equations on the world's largest computers. You have already seen some examples of these computer simulations in the above page on the Antennae galaxies.

Watch for an opportunity to see the IMAX film "Cosmic Voyage". It has the best computer simulations ever made of galaxy collisions, as well as spectacular photography of the real sky.

One of the few galaxies that is moving toward the Milky Way is M31 in Andromeda, the other giant spiral galaxy in the Local Group. Perhaps in a few billion years M31 will crash into the Milky Way. Here's a computer simulation of the Impact of the Milky Way and M31 from the Canadian Institute for Theoretical Astrophysics.

If you are interested, you can find several more computer simulations in Interacting galaxies by Joshua Barnes of the University of Hawaii.

10. SUMMARY:

The idea of a galaxy -- an "island universe" of billions of stars -- was first proposed more than two centuries ago, but was not confirmed until 70 years ago.

Galaxies are fundamental building blocks of the universe. They come in many sizes and shapes, ranging from dwarf galaxies containing only a few million stars to giant galaxies many times more massive than the Milky Way. In spiral galaxies, new stars are being formed at a high rate; in elliptical galaxies, star formation has ceased long ago.

One of the most challenging problems of astronomy is to estimate the distances of galaxies. Just 10 years ago, the uncertainty in measuring distances to galaxies was about a factor of two. But today, astronomers have developed a variety of techniques to measure distances with an uncertainty of about +/- 15%.

Most galaxies are moving away from the Milky Way: the more distant the galaxy, the faster they are moving. This fact is expressed by a simple equation called Hubble's Law. Astronomers can use Hubble's Law to estimate the distances of galaxies from their spectra, and hence to map the distribution of galaxies in the universe.

Galaxies are not distributed uniformly through space. They are concentrated in small groups, clusters and superclusters, separated by large voids of intergalactic space containing few galaxies. The quest to understand why galaxies are distributed in this manner is a major research area of modern astronomy.

The masses of galaxies and clusters of galaxies can be measured by several methods. Like the Milky Way, most of the mass of galaxies and clusters is some mysterious dark matter, which we have not yet identified.

The gravity due to the mass in clusters of galaxies can bend the light of galaxies behind the cluster. This gravitational lensing can magnify, distort, and produce multiple images of the background galaxy. By analyzing the distorted images, astronomers

In clusters, the distances between galaxies are not great compared to their sizes. Disturbances due to passing galaxies can cause the formation of spiral arms. Galaxies can also "collide", disturbing their shapes greatly and often triggering bursts of star formation.

(Return to course home page)