To clearly visualize the following discussion, it would be immensely helpful to have a chart of
the whole sky, ideally in the galactic plane of reference. Even better would be an "overhead"
schematic, centred on the Sun, showing at least the cardinal galactic longitudes with the names
of the constellations along the Milky Way appropriately placed. Such a diagram of the Sun's
neighbourhood would look something like this, where the broad arrow shows galactic rotation,
while the small arrow at the centre show's the Sun's motion relative to nearby stars. The galactic
centre lies off the diagram in the direction of Sagittarius ("Sgr").
Scientists like to classify things and astronomers are no exception. The study of stars resulted in one of the most important "rubber chicken graphs" in astronomy -- the Hertzsprung-Russell diagram. Its foundation is the relationship between a star's spectral type (closely related to colour and temperature) and its luminosity. Theory and observation went hand in hand to flesh out the diagram over time, and now we have a pretty complete picture of stellar variety, and how stars evolve.
And the range is incredible -- from dim red stars of less than 1/1000th of the Sun's output to blazing blue beacons of nearly 1,000,000 solar candlepower, with respective estimated lifetimes of hundreds of billions to as short as a few million years. For hydrogen burning, or main sequence stars the spectral types from hot/blue/bright to cool/red/dim are: O, B, A, F, G, K and M. Outside this scheme are extremes like white dwarfs and supergiants with different colour-luminosity relationships, but all are described by the grand design of the H-R diagram.
When we observe spiral galaxies, we see that the arms contain many luminous, blue, type O and B stars and glowing nebulae (red in photographs). Our sky is similarly populated, being almost without exception confined within or near the Milky Way band. Spiral galaxies also show a more or less prominent central bulge of yellower stars, and we see a similar feature in the constellation Sagittarius, although much obscured. Lastly, spirals, particularly when edge-on or nearly so, show dark clouds liberally sprinkled throughout the disk system. Our sky is similarly "polluted," the most prominent such material being the Great Rift running from Cygnus south to Centaurus. These are the most obvious clues to our galaxy's type, but could really only be inferred once astronomers had photographs of other galaxies to work with.
Star motions and streaming
While the ancients believed the stars to be fixed and unchanging, modern astronomy has changed all that. A high dispersion spectrogram of a star will instantly show if it has any significant component of motion along the line of sight from the shift in its dark absorbption lines. This is called the radial velocity, and a large body of such data can be fairly quickly amassed. With nothing more than this, we can infer certain elements of our motion with respect to the stars, as we shall shortly see. To see where the stars are really going we need another component to derive the motion in 3-D space. This is the part that takes real patience, because it means having to see the stars move across the sky as seen against the background of more distant stars (quasars nowadays). This proper motion, as it is called, relies on accurate measurements of position over time, usually at least decades for the nearby stars within a few hundred light years. For the more distant stars, the radial velocity is all we will have to work with for a long time.
In the early days of stellar spectroscopy, a relationship was observed whereby some stars of later spectral type (cooler) appeared to move more quickly toward the half of the sky in the direction of galactic longitude 270 degrees. The cooler the star, the faster the observed motion, the exception being the red giants. Many stars of type M, along with RR Lyrae variables, were found to be zipping along at 130 km/s relative to the Sun. These stars were called high velocity stars, and today a speed of greater than 60 km/s earns a star that title. The significance of all this was realized when stellar evolution theory caught up. But before then, a study of the radial velocities of more than 2,000 naked eye stars revealed the Sun to be heading toward the star 99 Herculis, not far from Vega, at 20 km/s relative to the stars' mean motion. Another study of the proper motions of more than 700 fifth-magnitude type A stars showed them appearing to diverge out of Hercules/Lyra and converge on the opposite side of the sky in Vela, in excellent agreement with the radial velocity measurements.
The nearby Andromeda Galaxy is a wonderful laboratory. Here is a spiral galaxy we can see from outside, and is near enough to study individual stars. Velocity measurements from spectra showed the rotation of the disk, with an expected decrease in angular speed with increase in distance from the centre. From within our own galaxy the situation is not so straightforward. Let's assume all stars went around the galaxy in circular orbits in one plane, with angular velocity decreasing with increasing galactic distance. Within, say, 10,000 light years stellar radial velocities would be zero toward the four cardinal longitudes (0/Sgr, 90/Cyg, 180/Aur, 270/Vel). Stars would be in approach in quadrants centred on longitudes 135/Cas and 315/Cen, and in recession toward longitudes 45/Aql and 225/Mon. Making allowance for the scatter in observed radial velocities, this is borne out; evidence of the nearly circular orbits of the Sun and nearby stars. Looking in the directions of greatest radial velocities, the farther the star the higher the speed. For our corner of the galaxy this works out to a rate of change of 4 km/s/1,000 light years. The next step was to find out how fast we and the stars orbit the galaxy.
In the 1920s an ingenious method was devised to determine the solar velocity about the galactic centre. The distribution of the globular clusters was clearly centred about the nucleus of the galaxy, and appeared roughly spherical. Such a distribution implied randomness in the clusters' orbits, the average of which would provide a standard of rest. Because the globulars are mostly found toward Sagittarius, and few lie near the ideal 90 degrees from the galactic centre, the vectors derived from radial velocity values will necessarily be more uncertain. The result obtained was 180 km/s, but the realization that the system of globulars is likely somewhat flattened, hence exhibiting a systematic rotation, combined with other evidence, has had that value increased to 220 km/s.
The globulars were again pressed into service to determine the distance to the galactic centre. Spectroscopy of member stars and associated Cepheid variables enabled the centre of the globular system, and hence the galaxy's nucleus, to be pegged at 22,000 light years. Later studies of the area of the nucleus itself from optical to radio wavelengths have refined this to 25,000 light years. This means that it takes the Sun, at its accepted distance from the galactic centre, some 240 million years to complete one orbit -- the galactic "year."
Modern studies using brilliant O and B stars' and hydrogen cloud radial velocities show our galaxy's disk material rotation speed to actually increase with distance from the centre (however, the angular rate of rotation does decrease). Between 5,000 and 50,000 light years. there is a general increase from roughly 220 to 290 km/s, apparently with some large- and small-scale variations, likely due to streaming effects of the spiral arms. This increase in speed is explained by the distribution of mass throughout the galaxy, unlike the case with our solar system where nearly all the mass is in the Sun.
It was a study of stars in the Andromeda Galaxy that first pointed to the marked difference between those in the central bulge and those in the disk system. It was soon shown that the same applied to the Milky Way. The first stars studied, our Sun and those stars in our neighbourhood, belong to Population I. They are characterized by relative youth and high metal (elements heavier than helium) content. The brightest, blue members can exceed absolute magnitude -9, and supernovae still regularly pop off. These stars were born well after the earliest epoch of star formation, when ancient supernovae had seeded space with the metals that formed in the massive progenitor stars. Indeed, the Population I stars are only found in or among the spiral arms, associated with the molecular clouds and nebulae that are the future birthplaces of other stars.
By contrast, the central bulge is composed of Population II stars and is rather gas poor. The brightest stars are about absolute magnitude -3, and are pretty much all evolved cool main sequence stars from the earliest periods of star formation. Metal content is much less than in Population I stars, and any members massive enough to go supernova have long since exploded. Globular cluster stars also belong to Population II, as do the faster high velocity stars.
This is something of an oversimplification, as the variety of stellar morphology, composition and location implies a more graded continuum.
To see where we are in the Milky Way we need a way to map our surroundings. Parallax measurements, using the swing of the Earth in its annual orbit around the Sun, works for only the nearest stars out to maybe 200 light years (soon to change with the Hipparcos satellite). To reach farther we turn to the H-R diagram. Knowing a star's apparent brightness, we work out its distance from the assumed inherent luminosity. For a number of reasons this is often not very accurate, but it will get us in the ballpark. Because of their brightness and the fact that they are found mainly within the arms of other spiral galaxies, O and B type stars are the obvious choice. An added benefit is a way to find the distance to many emission nebulae, as only O and B stars emit enough UV radiation to ionize these clouds of mainly hydrogen gas.
The first map using this technique showed rather clearly three spiral features within 15,000 light years, with the Sun near the inner edge of the middle one. Based on the direction as seen in our sky these were named, in order of decreasing distance from the galactic centre, the Perseus, Cygnus-Orion (our location) and Sagittarius arms. This was an important first step, but it was clear that our view out into the galaxy wasn't. Dark clouds like the Coal Sack and Great Rift obviously hindered our view, but it was found that even the clearer areas must cause some amount of obscuration. The scarcity of galaxies in and near the Milky Way band couldn't be a coincidence. In addition, stars at great distances often appeared to be too red for their type, much as the setting Sun gets redder (and dimmer) as it is seen through more atmosphere. Subsequent investigation revealed an average dimming in the galactic plane of a factor of two per 3,000 light years, with much higher local absorption within denser clouds. The galactic centre is dimmed by about 27 magnitudes, which means 0.0000000015 percent of visible light makes the journey to our telescopes.
The Emperor's new eyes
Radio astronomy began quite by accident, but was quickly pressed into service. The technology matured rapidly, and ever larger radio telescopes were built to obtain better resolution. At the same time new wavelengths were monitored, revealing emission from different atoms and molecules in interstellar space as well as from stars and galaxies. To the much longer wavelengths of radio, even the densest of interstellar clouds are virtually transparent. Now we can "see" everywhere, but how do we interpret the data? For the particular radio feature under investigation, if we know exactly the rest wavelength (the 21-cm emission line of neutral hydrogen, for example) we can use doppler shifting to good effect. Along a given line of sight through the galactic plane, a number of discrete features may reveal themselves by their different radial velocities, i.e. the component of relative velocity toward or away from us. By fitting the data to a model (ideally supplemented by observed stellar motions) of galactic rotation, we can map out the locations of these features. This will likely only give a crude approximation, as the method doesn't take into account local variations in velocity. Even so, such maps do show something of the spiral nature of our galaxy.
While being more akin to optical astronomy, the technology to peer into the infra-red arrived after radio. The shorter wavelengths of IR don't have the penetrating ability of radio, but it's still pretty darn impressive; for each magnitude of loss at 9,700 nm, blue light at 440 nm is attenuated by 50 magnitudes! Of course, shorter IR wavelengths are more heavily absorbed than this.
Space is a veritable chemical lab. Over the decades ever more numerous and complex molecules have been found floating among the stars. Optical absorption and radio emission features have allowed us to identify well over 50 molecules, most of them organic, some not found on Earth nor reproduced in terrestrial labs. With stuff out there like ammonia, methanol and formaldehyde, one can imagine space stuff being cleaned, pickled and preserved! But you'd have to sweep up a lot of space to fill a jar -- average interstellar density is one atom per cubic centimetre. The highest concentrations of interstellar molecules are within so-called molecular clouds, where we generally also find high concentrations of dust and hydrogen molecules. A wide range of size and mass is to be found among the clouds; a typical large cloud is 25 light years. across with 2,000 solar masses of gas and dust, while a small globule would be 10,000 AU across, containing lass than half of a solar mass of material (but with a probable hydrogen density 100,000 times greater than the interstellar average).
The visible discrete clouds, relatively nearby on the galactic scale, can be probed for distance by counting stars seen against the cloud and comparing with the count for a nearby unobscured area. Because of the wide range of absolute magnitudes among stars, the uncertainty of this technique can be as bad as a factor of two. Here are some derived distances in light years: the great Rift in Cygnus, 2,000 to 3,000; the Rift from Aquila south to Lupus, 650 and less; the relatively thin clouds from Orion south to Vela, 2,000; the Taurus cloud, 450; the Coalsack, 550. These are useful numbers as they allow us to further flesh out the spiral arm we reside in. How is this so? In most spirals the dark nebulae are concentrated on the inside edge of the arms, and we may reasonably assume the same for the Milky Way. At radio wavelengths, carbon monoxide (CO) emission was found to be the best tracer of these clouds, allowing the otherwise distant and invisible clouds to be detected and mapped, the optical studies making a baseline for comparison.
Density wave theory
Many of the galaxies we observe are spirals; so many so (even in the very distant past of the Hubble Deep Field) that we must infer that spiral arms aren't just a brief phase a galaxy goes through. And assuming that, we must conclude that the arms are either stationary, or that they rotate as a "solid" unit. Why? If the arms were tied to the orbiting stars/gas within them, we would find many of them wound up tightly after a few rotations about the galaxy because of differential rotation of the disk material. This hasn't been observed. Of course, we have no way of directly detecting any rotation of spiral arms, only the stuff in them. So we must resort to theory, and stationary arms aren't in vogue.
Calculations show that a natural consequence of a stable, highly flattened system like our galaxy is a spiral-shaped pattern of density variation brought about by the gravitational potential in and near the galactic plane. The most stable configuration is one with two trailing arms originating on opposite sides of the nucleus, all parts rotating at a constant angular rate like a two-armed boomerang. Theory predicts features that are in close agreement with what is observed; a spiral zone between 10,000 and 50,000 light years.; pitch angle of the arms at six degrees; arm spacing of nearly 10,000 light years.; and large-scale streaming of gas and young stars on the order of six to 10 km/s. What the theory predicts that we can't confirm is that the arms revolve more slowly than the disk material, except at the outer ends of the arms where the speeds are about the same.
According to this model, then, here's what happens. Stars and gas in the disk catch up to and linger in the spiral troughs of low gravitational potential. This causes a pile-up (on the inside edge of the arm) and increase in density, causing some clouds to collapse and form stars or clusters, often with associated nebulae. Many of the more massive new stars will quickly cook off as supernovae while still in the arm or shortly after emerging (massive stars live for as short as one to two galactic days), seeding their environs with future star stuff. Meanwhile the inexorable pull of the galaxy drags everything along through the following ridge of high gravitational potential -- faster here, with the necessary drop in overall density -- by which time all the brightest stars are no more. The arms are only about ten percent denser than the interarm areas; it's primarily the blue supergiants that make the arms so visibly prominent. And then, a good fraction of a galactic year later, the cycle repeats.
The big picture
Let's complete our picture of the Milky Way system, and put a scale to all the components. The oldest, largest and dimmest component is the nearly spherical halo about 300,000 light years. in diameter. A billion very metal-poor red stars up to 15 billion years or so old reside here. The most prominent members are the globular clusters. Orbits of the stars are random and fairly elliptical; the halo is the source of the faster high velocity stars (really the effect of our faster motion leaving them behind as they saunter through the disk).
The central bulge, also composed of relatively old, red stars, albeit with a higher metal content, merges gradually with the halo. It's about 15,000 light years. across and 6,000 thick. Recent evidence suggests it may actually be a bar with an aspect ratio of 2:1 as seen from above. The central 1,500 light years. contains gas more concentrated than anywhere else in the galaxy; enough to make 100 million Suns.
The disk, at 100,000 light years. across, is where we reside. The thicker component is about 2,000 light years. thick and is the home of older, higher velocity, lower metal content disk stars. The thin disk, at 500 light years. thick, is where the dense gas clouds congregate, from which the most massive, high metal content, rapidly evolving stars form.
For the inner part of the Milky Way closer than the Sun, rotation speeds imply a mass of 100 billion Suns -- a mass-to-light ratio of 4:1. But in considering halo star motions and other factors, the mass of the galaxy jumps to a staggering 1,000 billion Suns -- a mass-to-light ratio of 20-30:1! This seems to indicate a lot of unseen dark matter, mainly in the outer galaxy.
We get around
We are near the inside edge of the Cygnus spiral arm. However, during the Sun's lifetime of 20 galactic years it must have passed in and out of spiral arms as many times. Recall the Sun's motion toward the star 99 Herculis. This is 30 degrees toward the galactic centre from circular motion, and 23 degrees upward out of the galactic plane. We will be at perigalacticon (closest to the nucleus) in 15 million years -- less than one galactic month. At apogalacticon the Sun is seven percent farther out. About two million years ago we crossed the galactic plane going northward, and are currently 50 light years. above it. Coincidentally, at about the time of perigalacticon the Sun will have climbed to its highest point of about 250 light years.; could be an interesting view! About 32 million years after that we'll have passed again through the galactic plane and lie 250 light years. below. This vertical oscillation cycles at 3.5 times per galactic year, kind of like the horse on a merry-go-round. The cause of this bobbing up and down is the gravitational pull of the denser disk material.
In preparing this article, much was gleaned from the fifth edition of Bart and Priscilla Bok's The Milky Way, by Harvard University Press, 1981. Also helpful in nailing down a few facts was the more recent and somewhat more general The Guide to the Galaxy, by Nigel Henbest and Heather Couper, Cambridge University Press, 1994. Both are highly recommended.