Galaxies of All Kinds:
A Point-Form Summary.
This section of the course notes, and the associated PowerPoint presentation, makes the following critical points:
havingn recognized the existence of external galaxies, Hubble extended his study of their distances to as many as possible, but he also established a classification scheme to describe the range of observable properties they have. He identified elliptical, spiral, barred spiral, and irregular galaxies, and created what has come to be known as the tuning-fork diagram
the very existence of this diagram raises the question of whether there is an evolutionary relationship. Will an isolated galaxy of one type turn into another? The answer seems to be "no"
the way in which stars orbit within a galaxy allows us to determine the mass of that galaxy. If a galaxy was like the Solar System, with the dominant mass right at the centre, stars farther out would orbit much more slowly, in a predictable way. Instead, we typically discover that the outlying stars are moving as rapidly as those closer to the centre. This tells us that there is a large, distribued gravitating mass (unlike the Solar System, where the domininat gravitational influence is that of the Sun, a 'lump' right at the very centre
the inference is that galaxies are very massive, but if we assume that the light from a galaxy comes from the visible stars and gas, then there are not enough stars present to explain the large masses. There must be more material present than just the stars, and indeed galaxies may be dominated by this material, whatever it is. The problem is that it emits no light, which is why it is referred to as 'dark matter'. It may make up 90 percent of the mass of a galaxy, but we do not yet know what it is
galaxies are not very widely separated, relative to their sizes; in this way, they differ from stars in the galaxy. We now also realize that they are surrounded by considerable 'dark halos'. The net effect is that galaxies will, and do, 'rub shoulders' and interact with each other -- on very long timescales, of course. When galaxies collide or interact, there is very little chance that stars will actually hit head-on, but the orbits of the stars get scrambled as they variously feel the combined gravity of the two. Huge 'tidal tails' can be created in such interactions, and there is growing evidence that galaxies can actually merge. In this way, for example, our own galaxy and the Andromeda spiral are expected to merge completely in about a billion years
mergers are actually made more common than might otherwise have been expected by the fact that galaxies are not distributed randomly in space. Instead, we find them in groups, clusters and superclusters of various scales. Our own Milky Way is in a small cluster of ab out thirty galaxies, of which the two dominant ones are the Milky Way and the Andromeda spiral. There are also much larger cliusters, containing thousands of galaxies, and superclusters ('clusters of clusters') stretching across tens of millions of light years
large-scale surveys have also revealed enormous 'voids', regions of intergalactic space in which there are few or no galaxies to be found at all. One of the goals of modern astrophysics is to explain how all of this elaborate structure arose from the smooth conditions that prevailed in the earliest days of the 'Big Bang'
Associated Readings from the Text.Please look at: Chapter 21, pages 650-661. Chapter 22, pages 678-692.Spreading the Net.Once Hubble recognized that there were many galaxies in the universe which were comparable to the Milky Way, he realized that there were lots of questions to be addressed: Are there galaxies of various types, or are they all just like the Andromeda galaxy and the Milky Way? Are the galaxies all of the same size and mass, or are there `giant' and `dwarf' galaxies? How are the galaxies distributed? Are they generally isolated, or are there pairs, groups, and clusters of galaxies analogous to what we see for the stars themselves inside the Milky Way? Working all this out requires a number of things, the important element of which is deriving distances to the galaxies. (Without distances, it is not possible to work out total sizes and masses.) This is a long process, requiring years of observation and analysis. One thing I should emphasise is that the Cepheid variables provide direct distance estimates to only a small number of moderately nearby galaxies: in the more remote galaxies, even the bright Cepheids cannot be seen. There are other ways of working out their distances, until eventually distances for the most remote galaxies are being determined on the assumption that those galaxies are very similar to the ones we see nearby: we intercompare their total brightnesses. The distance-determination work would clearly take a long time, but there were other things one could and should do. Hubble recognized, for instance, the pressing need to establish an enormous database of galaxies so that one could address some immediate questions about their nature and tendency to cluster together. This could be done in straightforward fashion by examining photographs of the sky and adopting some scheme for classifying the galaxies. Note the parallel to what Annie Cannon did in classifying the spectral types of stars! She adopted a scheme which had no physical basis in particular (no one knew what the difference was between A stars and G stars, for instance), but her work provided a careful breakdown of the numbers and types of stars in the sky. Once the new science of stellar atmospheres was understood, her catalogues provided a great list of stars of all types for careful followup and more detailed investigation.Hubble's Tuning Fork.Even to the completely untrained eye, photographs of galaxies show a great variety of forms (see the figures on pages 628-630 of your text), but there seems to be a fairly natural subdivision into three subgroups: galaxies which are smooth and regular in appearance, whether round (like a beachball) or somewhat flattened (looking like a football); galaxies which display spiral structure; and galaxies which are more irregular and unstructured in appearance. Hubble defined a classification scheme on this basis, a scheme which soon came to be known as the `Tuning Fork diagram' of galaxy types. This figure is shown schematically in the figure below, but let me comment on just a few aspects of it: The elliptical galaxies - the smooth, regularly-shaped ones - are assigned a number which is a measure of how round or flattened they look. An E0 galaxy, then, is an elliptical (E) galaxy which looks perfectly round so that it has zero (0) flattening. The most flattened E galaxies we know are E7 (there is a formula we apply in determining the numerical value). [One thing to note is that galaxies which look round may in fact be flat. For example, a particular E0 galaxy could be pancake-shaped, and we might merely be looking at it straight down; equally, it could be cigar-shaped, with us seeing it end-on. Trying to work out the actual shape of a galaxy, free of these projection effects, is not straightforward. There are, however, so many galaxies of completely round appearance that is is clear that at least some of them must really be spherical, if galaxies are oriented at random in space.] The spiral galaxies are not all simple `pinwheels.' There are some spirals which have central bars, so that the spiral arms do not originate in the very centre of the galaxy. This is why the Hubble tuning fork has the characteristic shape which gives it its name: Hubble treated the barred spirals and regular spirals as two separate but parallel sequences. The irregular galaxies are all a hodge-podge of different kinds.Evolution? What's the Evidence?Hubble created his tuning fork diagram so that he could attach a label to a particular galaxy, a label which is more than just a number but rather something which describes it in some minimal way. But the trouble with a scheme like this is that it immediately suggests that there may be some evolutionary relationship between the galaxies of various kinds. Merely by putting them down on the page in this fashion, for instance, we seem to imply that irregular galaxies are very far removed from elliptical galaxies, and that in some sense the spiral galaxies are intermediate. We encountered the same sort of thing when we first plotted a Hertzsprung-Russell diagram for stars and recognized the main sequence. We were prompted then to ask whether or not individual stars `evolved' (remember the special sense in which astronomers use that word!) up or down the main sequence. (Later we decided that they did not.) Likewise, we now ask whether or not individual galaxies change form. Does an elliptical gradually turn into a spiral? Does the reverse happen? There are plausible qualitative remarks you can make which suggest that some evolution of the galaxies might be credible. For instance, the irregular and spiral galaxies contain a fair bit of gas, some of which is going into the formation of new stars. Perhaps when this is all used up, the galaxy will look like an elliptical: the E galaxies have almost no interstellar gas in them. Conversely, perhaps the ellipticals are big and round (at least some of them) because they are only slowly contracting, just in the process of formation, and have not yet flattened out to their final (spiral?) forms. To answer this question, we clearly need to know more about the properties of the galaxies. Think again about the example provided by our study of the stars. When we discovered that the stars at the top of the main sequence were much more massive than those at the bottom, we realized that there could be no significant evolution along the main sequence. This was because no massive star could lose enough mass, even though it is converting a small fraction of it to energy, to wind up low on the main sequence; and no isolated low-mass star could ever gain enough mass to become a hot blue star. Determining the properties of galaxies is a challenging proposition, and takes a lot of time and interpretation. In the text, you will find a discussion of the relevant issues. In the end, though, the most important piece of evidence -- in nice analogy to what we learned about the stars -- is that the galaxies are, at least in some measure, distinguished by mass. Although there are a range of masses, the very biggest ellipticals tend to be more massive than the spirals, and a typical irregular galaxy is of considerably lower mass still. Right away this seems to tell us that a truly isolated galaxy almost certainly does not evolve from one kind into another on its own. On the other hand, it may be that smaller galaxies merge to form larger ones (and change structure in the process). We will return to this point later. [Important digression: It is true that eventually the gas inside a typical spiral galaxy will be all used up. This will take a surprisingly long time, and the Milky Way (for instance) will not soon become devoid of gas. At present there are new stars being born in places like Orion in the Milky Way, but there are also dying stars which eject gas as they becme supernovae or planetary nebulae. Still, there is inevitably a slow net decrease in the amount of gas in the interstellar medium because some of the material winds up being locked up in the form of planets, white dwarfs, neutron stars, or black holes. In the very long run, the gas supply will be exhausted. But calculations show that a spiral can remain pretty much as it is, with a continuing processing and recycling of gas, for many billions of years.]The Masses of Galaxies: Dark Matter.How did we first figure out the mass of the sun? The asnwer is straightforward: we noted that the Earth takes exactly a year to go around it, and we know how far we are from the sun, so we can calculate how much matter the sun must contain to provide the pull of gravitation which forces the Earth to move in exactly this path. We didn't have to use the Earth in this calculation. We could instead have used Mercury, which is closer to the Sun and which moves faster. Had we done so, we would have gotten exactly the same answer for the sun's mass - as we would if we were to consider any of the planets, or Halley's comet, or an asteroid. This, I am sure, does not surprise you: it reflects the fact that the dominant mass in the Solar System is that of the Sun. It might have been otherwise! Remember that whenever there is a spherically-symmetric distribution of material, the gravity you feel depends only on how much matter lies closer to the center than you are yourself. Suppose, for instance, that you could put the entire solar system out to and including Mars inside a huge perfectly spherical shell, with steel walls a million miles thick. Within the shell, all the terrestrial planets would orbit the sun exactly as before! But Jupiter, being outside the ball, would feel the combined gravitational effects of not just the sun (and the nearly-negligible inner planets), but also the enormous extra contribution of the steel shell itself. This would require Jupiter to move very rapidly in its orbit, to avoid being pulled inward; and an external observer would be able to deduce the appropriate total mass, just as we now use the Earth or any other planet to infer the mass of the Sun. The other outer planets would likewise move at the rapid speeds which would be determined by the gravitational effect of the total mass interior to their orbits. The notion of a steel shell is unrealistic, of course. Suppose, more sensibly, that the Solar System was filled with a lot of distributed matter, like small rocky chunks which added up to a large total mass. In that case, Pluto would orbit in a way which implies that there is a lot of mass interior to its orbit, while Mercury would give us a much smaller value. Indeed, if there was a lot of distributed junk of this sort, it could be that Pluto would feel such a strong net gravitational pull that it would have to be moving just as fast as Mercury to avoid falling in towards the sun. Now let us turn our attention to the galaxies. Knowing the distance of a galaxy of interest, we can calculate its size and work out how far the various constituent stars are from the galaxy centre. We then study the starlight from the different sides of the galaxy, and use the observed Doppler shifts to work out how fast the stars are moving as the galaxy spins. (See page 680-681.) These speeds, coupled with the distances, then tell us how much mass lies closer to the centre of the galaxy than the particular star whose light we are examining. If essentially all of the mass of the galaxy were in some huge central lump, the stars farther from the center would be moving more slowly, just as Pluto moves more slowly than the Earth. What we find instead is that stars far from the center move at a speed which is often not much less than those closer in: the 'rotation curves' are said to be flat. This tells us that the matter in a galaxy is spread out, not merely housed in some big central object. But what form is this matter in? If the distributed matter were in the form of stars, you would expect to see lots of distibuted light rather than just a very bright central object. And of course we do see exactly this: the light of the galaxies is spread out, so there are clearly many stars there, spread out over a large volume. But careful analysis shows that there must be lots more material present in addition to the stars: on their own, the stars simply cannot add up to as much mass as we know must be present to explain the dynamics (internal motions and rotation) of the galaxies. The problem is that this material, whatever it is, gives off very little (if any) light. For this reason, it has come to be known as dark matter. No one knows yet what it is. It could be made up of myriads of small chunks comparable in size to a planet, or supermassive black holes, or exotic subatomic particles. The discouraging fact is, however, that up to ninety percent of the matter in the universe may be in this dark form! We know it is present because it influences other material through its gravitational effects; but it cannot be seen. Needless to say, this makes it hard for astronomers to study it.Galaxy Interactions.What would happen if two galaxies were to collide? Suppose, for instance, our galaxy and the Andromeda spiral were to run into each other, face-to-face. You might expect that there would be colossal collisions between various stars, but in fact the chances of that happening are very slim. The stars are so widely separated, and the spaces between them so vast, that essentially every star would slip throught the gaps and go on unaffected. This does not mean that the net effect would be negligible. You would not merely see two galaxies pass through each other in ghostly fashion and go on their way intact! There are two important things that would take place: Any amounts of distributed gas in the galaxies would interact quite vigorously. (This is analogous to the sort of consideration we discussed when we talked about how the globular star clusters give us information about the formation of the Milky Way.) This might lead to some spectacular astrophysical phenomena, new bursts of star formation, and so on. Individual stars, although untouched, would move off in new directions because they now feel the extra gravitational tug of a passing galaxy of very large mass. A star which, with its companions, once orbited a galaxy in a simple predictable way may move off along a great sweeping arc in some new direction. In short, the gravitational influence of the other galaxy in the collision can lead to the formation of tidal tails. The importance of galaxy-galaxy collisions was first realized when strong radio sources in the sky were first identified with objects that `looked like' colliding galaxies. Of course, we see only a brief snapshot of such an event! It will take tens or hundreds of millions of years for one galaxy to pass through another, and in a single human lifetime we will see no change in the appearance of such a system. But we can do better than merely speculate qualitatively about the likely appearance of such collisions and the ways in which they will change. Thanks to the advent of modern computers, we can now simulate collisions and work out how the individual stars will move under the complex and competing gravitational influence of the participating galaxies. This sort of experiment first attracted serious attention from the astronomical community in the 1970s thanks mainly to the work of two brothers, Juri and Alar Toomre. In class, I presented a few figures which showed some of their results, figures which show how the individual particles (stars) would be expected to move as the collision takes place. I also showed some photographs of real galaxies in which we see what look like perfect examples of collisions of the sort simulated. The close similarity of the images gives one some confidence that these are real collisions in the universe, a conclusion which is strongly reinforced by the discovery that the stars are moving (as shown by the Doppler shifts in the spectrum of the light) at just about the speed and in the directions which the computer simulations say they should be if a collision is to explain the features we see. One of the wonderful results of the computer revolution is that we now have enough programming power to carry out even more detailed simulations than those the Toomre brothers could so -- and create wonderful movies from them! In an associated PowerPoint presentation, you will find links to such simulations.The Cross-Sections of Galaxies.You may be surprised to learn that galaxies can collide, because our experience with individual stars within the Milky Way was that they almost never run into one another. But there is an important difference. A typical star may be a few million miles in diameter, and they are separated from their nearest neighbours by something like a few light years - some tens of trillions of miles. This means that they are separated by billions of times their own diameters, and can effectively be treated like small points that never `rub shoulders' as they pass. They can be said to have small cross-sections as they move about. Galaxies present a different picture. The Andromeda spiral, for instance, is about two million light years away, but is about one hundred thousand light years in diameter. In other words, it is only about twenty `galaxy diameters' from its nearest neighbour, and it is correspondingly more likely to suffer a real physical encounter if it moves closer to the Milky Way. This tendency is very much increased now that we realize that the galaxies may also contain large amounts of dark matter in an extended halo which surrounds the visible parts. A simple analogy is to imagine letting a bunch of mice into a large enclosed area, like a baseball ground. Of course they might seek each other out, but if they ran around completely at random the chances of any two bumping into each other would be slim. The same would not be true for elephants, which are so much bigger that they present larger targets.Mergers.Astronomers now recognize that collisions and interactions between galaxies are very common. There is yet another reason for this: galaxies do not just sit all alone in the emptiness of space. As we will see, they tend to be found in groups and clusters, and this enforced proximity between numbers of galaxies also enhances the probability that collisions will take place. Indeed, we now believe that some of the most massive galaxies (the very big ellipticals) may themselves have been formed by smaller galaxies which not only collided but later merged so that all the stars within them became inextricably mixed together. There are ways of testing such theories - indeed, much of my present astronomical research concerns itself with such matters. So although we now believe that the Hubble tuning fork diagram was right to suggest that isolated galaxies do not transform themselves spontaneously from one type into another, the probability is very good that pairs or groups of galaxies can mix and merge, changing their appearances and nature as they do so.The Distribution of Galaxies: Conurbations.Once Hubble identified the galaxies as such, and started to derive distances to a reasonable number of them, it became possible to make a plot which shows where they are distributed in space. The first thing astronomers realized was that the galaxies are not distributed at random in the emptiness of space. Instead, they are grouped and clustered on various scales. In describing the sort of structure we see, I will ask you to visualise the sort of arrangement we see when we look at human habitations on the Earth. North of Kingston we find the occasional house in complete isolation from all others, but more commonly they are grouped into small clusters in what we call hamlets or towns. Kingston itself is large enough to be classified as a city, and its presence as a centre of employment and supply means that there are small outlying communities clustered around it. When we look at Toronto, we see an enormous city, around which we find a pronounced clustering of mid-sized communities like Mississauga and Oakville, Oshawa and Ajax. On even larger scales, we find what urban planners call conurbations: huge complexes of whole cities. The best example of this in North America is probably provided by the northeastern seaboard of the United States, an area in which the metropolitan areas of New York, Washington, Baltimore, and Pennsylvania and their outlying subsidiary communities nearly merge (and may eventually do so if growth continues at its present pace). There are also contrasting areas of low urban density, such as the sparsely populated interior of Australia (all the big cities - Melbourne, Sydney, Adelaide - are on the coast) and the central states in the USA. These areas are never completely devoid of people and habitation, but will be seen to correspond to the near voids we see in the large-scale distribution of galaxies. With this analogy firmly in mind, then, let us take a look at how galaxies are distributed.One Big Difference.Cities spring up for some social or logistical reason - a good harbour, excellent land for agriculture, an easily-defended region - and then grow and prosper from there. But cities have the great virtue that they do not move around (although of course they may weaken in influence and actually eventually decrease in population). Galaxies are a little different: if you were to put a galaxy down at rest in the universe, it would feel the gravitational influence of surrounding galaxies and matter (including the dark matter, however it may be distributed) and would wind up moving in response. In short, the distribution of galaxies we see right now is not exactly as it will be forever or as it always was. On the other hand, there is a limit to how far a given galaxy can have moved in the time it has been around, and we may be able to identify groups of galaxies which are close together now and which must always have been so, given the modest speeds with which they are moving (as determined from the Doppler shift in the spectra of the light we receive from them). Just as we did when studying the stars in the galaxy, then, we can look at any apparent pairs or groups of galaxies in the sky and ask ourselves the following questions: Are the galaxies truly `side-by-side', or is it a chance coincidence in which merely see one galaxy in roughly the same direction as one which is much farther away? If a cluster of galaxies can be identified, is it gravitationally `bound' or will the galaxies gradually drift away in independent directions into the emptiness of space?The Local Group.Our own Milky Way galaxy, and the Andromeda spiral which is found a couple of million light years away, are the two dominant galaxies in what astronomers refer to as the Local Group. (Only for an astronomer would a distance of two million light years be referred to as `local'!) This is in the figure below, in a three-dimensional representation. Measurements of the masses of the galaxies, coupled with an analysis of their motions, shows that the Local Group is indeed bound. That is, the combined gravity of the whole ensemble of galaxies is large enough that our own Milky Way will never `escape' it, and the member galaxies will continue their dance forever.Nearby Clusters.Astronomers draw a technical distinction between small `groups' of galaxies and somewhat larger `clusters' of galaxies in somewhat the same way that urban planners distinguish towns from cities. (I will not explain the distinctions here.) When we look farther out into space, we find that the Local Group is situated about fifty million light years from a large cluster of galaxies, a gathering which contains some thousands of galaxies. This cluster lies in the direction of the constellation Virgo. For this reason, it is referred to as the Virgo cluster. (Of course, the cluster of galaxies is very much farther away than any of the visible stars in the constellation, all of which are inside our own Milky Way galaxy.) There are other clusters of galaxies of comparable size and richness in other directions, but close inspection of them soon reveals an interesting fact: the clusters of galaxies are themselves clustered, just as mid-sized cities like Oshawa and Oakville are found in some numbers near large centers like Toronto. This higher-order clustering has earned the name `Superclustering,' and we are said to be in an outlying group in the Local (or Virgo) Supercluster.Larger-Scale Structure.With the passage of decades and the development of new instruments and telescopes, astronomers have been able to determine the distances to literally hundreds of thousands of galaxies. When the positions of these galaxies are plotted, a remarkable thing is found. Superclusters of extremely large scale, hundreds of millions of light years across, are arranged in ordered patterns, with long filaments or `great walls' of galaxies separated by immense voids, also hundreds of millions of light years across, in which very little is to be seen at all. The figures on page 691 of your textbook show this beautifully. It is not much exaggeration to say that this discovery represents one of the biggest challenges to modern observational and theoretical astrophysics. It is moderately straightforward to concoct a theory in which we explain how a single massive gas cloud can contract under the influence of its own gravity to form a single galaxy. (Many of the details will be wrong and rather messy, but the general outline is fairly well understood!) But it is much more difficult to understand how the universe, which was much more uniform and homogeneous in the early days (as we will learn in a subsequent lecture), could form such immense structures and develop such enormous order and coherence in the limited time that everything has been around. Single galaxies may form fairly readily here and there, but how do they `know' how to line up in big walls and sheets hundreds of millions of light years across? Astronomers all over the world are stretching their imaginations and observational capabilities to the limits in trying to understand this. Previous chapter:Next chapter0: Physics 016: The Course Notes, spring 2005. 1: The Properties of the Sun: 2: What Is The Sun Doing? 3: An Introduction to Thermonuclear Fusion. 4: Probing the Deep Interior of the Sun. 5: The Sun in More Detail. 6: An Introduction to the Stars. 7: Stars and Their Distances: 8: The HR Diagram: 9: Questions Arising from the HR Diagram: 10: The Importance of Binary Stars: 11: Implications from Stellar Masses: 12: Late in the Life of the Sun: 13: The Importance of Star Clusters in Understanding Stellar Evolution: 14: The Chandrasekhar Limit: 15: Supernovae: The Deaths of Massive Stars, 16: Pulsars: 17: Novae: 18: An Introduction to Black Holes: 19: Gravity as Geometry: 20: Finishing Off Black Holes: 21: Star Formation: 22: Dust in the Interstellar Medium: 23: Gas in the ISM: 24: The Size and Shape of Our Galaxy: 25: The Discovery of External Galaxies: 26: Galaxies of All Kinds: 27: The Expanding Universe: 28: Quasars and Active Galaxies: 29: The Hot Big Bang: 30: The Geometry of the Universe: 31: Closing Thoughts: Part 1:Part 2:Part 3: |
(Wednesday, 22 April, 2026.)
