—Johann Becher (1635-1682)

The bane of 17th-century astronomers was the poor optics they had to contend with. Galileo's telescopes had a restrictively narrow field of view, heavy distortion, chromatic aberration, and poor star images everywhere but in the very center. Even a century and a half later, the French comet hunter and cataloguer of nebulae, Charles Messier, was unable to resolve most of the objects he observed into more than hazy balls of light.

The bane of the modern astronomer, however, is light pollution. For all the optical obstacles they had to face, the first telescopic astronomers were able to make their discoveries because their skies were dark. Severe light pollution, as experienced today by observers in and around major cities, can turn an otherwise exquisite sky into a muddy, milky mess. Bright stars become dim ones, dim ones become invisible, and galaxies and nebulae that reveal their fine structure to viewers under country skies are unremarkable lumps of fuzz, if they are seen at all.

This is particularly infuriating because light pollution doesn't actually rear its ugly head until night falls. There is a period of time during dusk when the lights haven't yet turned on in full force, when the sky is a dramatic electric blue, foreshadowing a beautiful night under the stars—and then the lights go on.

There are actually two different kinds of light pollution. One is local glare: bright lights at night shining directly toward us. These lights destroy the eye's ability to detect dim light sources by saturating it. The saving grace is that it's easy to block out local glare by building a screen. Observatory domes minimize glare by blocking out almost the entire sky, leaving open only that section where the telescope is pointed at that moment.

The second kind of light pollution is perhaps less physically annoying, but all the more insidious for that: diffuse light pollution. Here, light doesn't shine into our eyes or into our instruments. Rather, it shines up into the air, where it reflects off particulates and back down toward the ground, and us. Most of the light never does come back down, of course, but continues straight on into space. But what does come back down cannot easily be blocked, because it covers the entire sky. The atmosphere is filled with small particles of water vapor, soot, and various other small bits thrown up by factory smokestacks. Generally speaking, you can't block out light reflected by these particles without also blocking out the very thing you wish to observe. (There are exceptions, of course, and you can well believe that astronomers take advantage of them whenever and wherever they occur.)

And it's not just astronomers who are affected. Much of the stuff in the air counts as everyday, honest-to-goodness, regular old air pollution. The skies of Mexico City, for instance, are useless for observing all but the brightest objects at night, not only because there are so many bright lights there, but also because there are so many particulates in the air to reflect all that light back down. During the day, that pollution is distinctly visible as a layer of smog that blankets the city and turns the Sun a bloody red as much as an hour before it sets.

This is all the doing of humanity, naturally. Such a thing wouldn't happen anywhere else but on the Earth.

Or would it?

In "The Color Green," I described how stars are very close to black-body emitters, so that they emit a broad spectrum of wavelengths of light, with a characteristic distribution that depends, to a first approximation, only on the effective temperature of the star. The distribution of the light of our own Sun has its peak in the yellow and green region, with somewhat less light coming from the red, and significantly less coming from the blue.

However, the peak is not quite sharp enough for the Sun to appear as anything other than a brilliant white, if you are foolish enough to look directly at it. Only near the horizon, where the atmosphere scatters a substantial portion of the shorter, blue wavelengths, does the Sun take on a orangish hue.

Most stars, like the Sun, do not have strong peaks in the distribution of their light. They may have peaks in the blue, or in the red, but because the black-body spectrum mixes in quite a bit of the other colors, the tint of the stars is muted and subtle. Even though the stars are often represented in color atlases as being deeply saturated, their colors generally are not that distinct.

That means that trying to classify stellar spectra on the basis of each star's peak presents a challenge. The spectral distributions of two stars whose temperatures are only modestly different will be so similar that you can't reliably put them in the right class just by looking at them. Some other clues must be used.

Although the stars are made predominantly of hydrogen and helium, there is a small proportion of other, more massive elements, which astronomers collectively call the "metals." (There are no exceptions. Even neon, a distinctly nonmetallic element, is considered a "metal" by this criterion.) Stars of different temperatures, and therefore different spectra, have different distributions of metals, as compared with hydrogen and helium, and these differences reveal themselves in characteristic absorption lines—dark lines puncturing the otherwise smooth progression of the spectrum. Maybe this could be used as a means for comparison.

Indeed, over the course of four years, the Italian astronomer Angelo Secchi (1818-1878) examined the spectra of about 4,000 stars, without the benefit of photography. In 1868, he was able to announce that he had arrived at a classification of stellar spectra into four different classes, which he called Types I, II, III, and IV.

Type I stars, such as Sirius in the constellation of Canis Major, had their peaks toward the blue end of the spectrum, and had very few absorption lines. Type II stars, like our Sun, had their peaks in the yellow and green, and had an almost uncountable number of thin absorption lines. Type III stars, like Betelgeuse in Orion, had their peaks in the red, and also had many lines, but these were grouped into bands. Finally, Type IV stars, none of them prominent in the night sky, were also reddish, but had very few absorption lines like the Type I stars. In fact, some of them seemed to have bright emission lines in their spectra.

Secchi's classification was refined a decade or two later by the American astronomer Henry Draper (1837-1882) and his staff at the Harvard College Observatory. In place of Secchi's four classes, Draper instead used the sixteen letters A through P. By this time, the photocell, essentially a light-sensitive battery, had been developed to measure the brightness of a star's light at different colors with better precision than was possible with the eye.

With the aid of the new photocell, Draper's staff was able to reorganize his original classification, changing the order of some and removing others that were the result of duplication or errors in measurement, arriving at the sequence of classes: O, B, A, F, G, K, M.

So fine were the new classifications, in fact, that each of these seven classes was further subdivided into ten subclasses, so that within the broad class of G stars, one would classify stars as being anywhere between G0 and G9. A G0 star would be hotter than the other G stars, being almost of class F; a G9 star would be cooler, edging toward the class K stars. The Sun has spectral class G2, indicating that it's toward the hotter end of class G.

These seven classes basically covered the same ground as Secchi's first three types. At the time, late in the 19th century, it wasn't well understood what powered the stars. It was generally thought that the stars formed hot in some way, and then gradually cooled down throughout their lives. For that reason, the hotter O and B classes were called the "early" classes, and the cooler K and M classes were the "late" ones. Even after it was understood in the late 1920s that atomic fusion powered the stars, and that their temperature and therefore their spectral class depend more on mass than age, astronomers persisted (and persist to this day) in describing the various spectral classes as "early" and "late."

What about the other stars, Secchi's Type IV stars? They were not particularly common in the night sky, with none of them being particularly well-known or prominent. But when viewed through a telescope, they had such a distinctly red color that Secchi said of them that they "gleam like rubies among the other stars." As far as their temperatures were concerned, they seemed to cover some of the same ground as the old Type III stars—perhaps even a bit redder—but they were chiefly distinguished from the other stars by their general lack of absorption lines. When constructing his new classification, Draper didn't know what to make of Secchi's Type IV stars, so he lumped them in one additional letter classification, class N.

Spectroscopic examination of stars allowed astronomers to do more than simply classify the stars; it also allowed them to analyze their composition. As I explained in "How to Cook a Star," the light emitted by compounds in the laboratory showed a remarkable correspondence with light absorbed by the outer atmospheres of the stars. In the case of the class N stars, the bands in the spectra suggested absorption by a variety of carbon-containing ionic species, notably CH and CN.

Toward the end of the 19th century, stellar spectroscopy felt the impact of photography. Astrophotography began in a limited way in 1850 with a picture of Vega, a prominent star in the constellation of Lyra the Lyre, and the fifth-brightest in the night sky. However, taking a spectrum of a star is considerably more difficult than simply taking a picture of the star. When recording the star itself, all of the star's light, of whatever color, piles up on the same point on the photographic plate.

In order to analyze the star's spectrum, that spectrum has to be spread out on the plate, and that means that the star's light is distributed over a relatively broad portion of film, rather than concentrated at a point. That requires either a larger telescope to gather more light, or longer time exposures. It took more than two decades before Draper took the first spectral image of a star—again Vega. (Draper was one of the leading proponents of astrophotography. He also took the first photograph of a nebula—the Orion Nebula—in 1880.)

Once the method was refined, however, it permitted astronomers to examine the spectrum at leisure, without concern for atmospheric conditions, and in much greater detail than was possible at the eyepiece. The relative strength and breadth of spectral lines could also be measured, instead of relying on the subjective and imperfect eye of the visual observer.

By recording spectra of the class N stars, Harvard astronomer Edward Pickering (1846-1919) was able to determine that these stars showed signs not only of CH and CN (which break down on the Earth, because they are rapidly oxidized), but also C2, molecular carbon. Because of this, class N stars (as well as class R stars, which Pickering distinguished on the strength of markings in the blue end of the spectrum which N stars lack) are known as carbon stars.

But what is all that carbon doing in the atmospheres of these stars? Carbon is a solid at ordinary temperatures, and remains solid long after most other substances have melted. Finally, at a temperature of about 4100 K, carbon sublimates—that is, turns directly from a solid to a gas.

But even that high temperature is fairly cool by stellar standards. The atmospheres of most stars are too hot to allow carbon to remain in molecular form; the Sun's surface is at about 5800 K, and gets hotter, rather than cooler, with increasing altitude. That means that carbon stars must be fairly cool—consistent with their red color.

What distinguishes carbon stars from their non-carbonaceous class M counterparts is that they have an atmosphere rich in carbon dust—soot, essentially. This carbon comes from the core of the star, which for most of the star's life has been fusing hydrogen into helium. When the star runs out of hydrogen and becomes a red giant, however, conditions in the core become hot enough to fuse helium into carbon. In stars no more than about twice the mass of the Sun, the carbon remains in the core, which is revealed when the outer layers of hydrogen and helium blow off and the star becomes a white dwarf.

In more massive stars, the extra heat produced by helium fusion creates convection between the outer, cooler layers and the carbon-rich hot core. Carbon wells up from the bottom and condenses as molecular carbon in the outer envelope of the star. As it happens, the carbon in the stellar atmosphere absorbs more blue light than it does red, just as it does over polluted cities on the Earth. The star becomes a deep, saturated red—much redder than you would expect based strictly on its temperature. It is almost as if you were observing an ordinary red giant through a red filter—except that the red filter is on the star, not in front of your eye. Moreover, the carbon continues to circulate in the atmosphere of the star, so that most, if not all, carbon stars are variable.

Secchi listed no well-known stars amongst his Type IV stars, the ones that became Draper's classes R and N. (Classes R and N have since been combined into a new class C, for carbon.) However, there are some prominent stars that were later determined to be carbon stars. One of the best known is mu Cephei, in the northern constellation of Cepheus, also known as Herschel's Garnet Star, for its deep red-orange color. Another is R Leporis, or Hind's Crimson Star, in Lepus the Hare, at the foot of Orion the Hunter. There is even some evidence that Betelgeuse, the distinctly orange star in the shoulder of Orion, is also a carbon star. However, its carbon bands are faint, its variability is low, and it is generally classified as a class M star.

Carbon particles can do more than redden a star. Sometimes, they can blacken it to the point of invisibility.

In 1795, the English amateur astronomer Edward Pigott (1753-1825) reported his observations of an obscure star in the constellation of Corona Borealis, the Northern Crown. Pigott had discovered the first of many different kinds of variable stars; he had, in 1784, discovered eta Aquilae, the first Cepheid to be identified (even before delta Cephei itself, discovered by his friend John Goodricke (1764-1786)). Variable stars not bright enough to be given a Greek letter designation are assigned capital letters beginning with the letter R, and since Pigott's star was the first such star to be identified in Corona Borealis, it was named R Coronae Borealis (or R CrB for short).

Pigott found that the variability of R CrB was "upside-down" when compared with other variables. Some variable stars turn out not to be variable at all. Instead, they are double stars that usually shine with the brightness of both of the individual stars put together. However, as the stars rotate around a common center of gravity, one of the stars periodically gets in the way of the other. When this happens, we see only one of the stars. The double star is too far away for us to see either of the individual stars, even through a telescope, but since we now see the light of only one of the stars, the star appears to dim. Then the star in front moves on, unblocking the second star, and the pair returns to its original brightness.

Of those stars that are intrinsically variable, some vary regularly, like the Cepheids, or at least semi-regularly. The rest are called irregular variables, and the most outstanding example of these is the eruptive variable, which can brighten by many magnitudes over a brief period of time, then return to its original brightnesses. Such stars should have been observed from time to time throughout history, and must have presented a problem for astronomers who felt that the heavens had to be changeless. That axiom had to be right; clearly, it was the star that had to be wrong, so such stars were, by and large, not reported by western astronomers. (There are reports of novae—stars that seem to appear, as if new, in the sky—but, generally speaking, only outside Europe.)

Pigott found that R CrB also erupted, but "in reverse," so to speak. Instead of brightening suddenly, it would dim by several magnitudes, usually within a few weeks. Ordinarily at the threshold of naked-eye visibility, the star's light would diminish by a factor of perhaps a few thousand, putting it beyond many amateur telescopes even today. (Probably especially today, with modern light pollution.) What's more, it went on erupting, according to no discernible schedule, and in fact, it still does so, having last produced a reverse eruption in 2003, and before that in 2000, 1996, and 1984.

Spectroscopic analysis of R CrB reveals considerable amounts of carbon. The most immediate speculation was that the carbon came directly from the star itself. However, it might be that R CrB is merely orbited by dust clouds, which occasionally obscure the star.

There are problems with this idea, though. Any dust cloud in orbit around R CrB is probably not massive enough to accrete gravitationally—certainly not in the 200 years that the star has been actively observed. However, in the absence of an atmosphere, dust grains will stick electrostatically, and that should have a measurable effect on the dimmings. That doesn't seem to be the case.

The other possibility—that the carbon does come directly from the star—is bolstered by the fact that R CrB and other related specimens, unlike most carbon stars, are hotter class F and G stars. Their surfaces and immediate surroundings are too hot for carbon to condense right away into its molecular form. Instead, it remains gaseous until the stellar wind has carried it far enough away for it to cool to carbon's condensation point. Then the carbon forms dust and abruptly obscures the surface of the star.

But this theory may have its own problems. For instance, R CrB ought to be so hot that carbon won't form until the carbon has been carried out a distance equal to 10 times the diameter of the star itself. At such a great distance, shouldn't the carbon be so dissipated that it won't significantly obscure the star? Perhaps electrostatic attraction keeps the carbon atoms close together until they can cool down enough to form molecules. Stay tuned.

Adapted from Astronomical Games, March 2004.