Scientific American, March 2000


by Philip and Phyllis Morrison



Check out the prediction of the third transit for a star; once it tells you three times, it's true..
From a cloudless blue sky the resplendent solar disk dazzles the eye. But during the tempered glow of crimson sunsets or through a veil of thin cloud, the eye can sometimes spy areas of darkness. In 1607 Johannes Kepler himself, brilliant theorist and no mean observer, glimpsed a "little daub" on a projected image of the sun. Two years later he published a tentative explanation, positing the "Extraordinary Phenomenon, or Mercury in the Sun," an infrequent transit of Mercury silhouetted against the glaring disk. He could not be quite sure, for his planetary tables were still not reliable enough to predict a transit, and he was in fact wrong.

There is a second distinct kind of sun-stain, no shadow but a true blemish, physically rooted in the sun itself. Each one is a kind of long-lived cyclone called a sunspot, a feature of the changeable magnetic weather in the hot solar atmosphere. The year 2000 marks a newsworthy maximum in the 11-year cycle of sunspots as they appear, grow, often merging, then finally die away.

In 1613 Galileo used his telescope to map and vividly describe spots he observed, although the Chinese had first noted them 1,000 years or more earlier. Both Kepler and Galileo wrote of two kinds of dark sun-stain: the neat circular planet dot that moved off in a few hours, and the large, complex sunspot areas that may endure to return after the four weeks of a solar rotation. The annals of Charlemagne, emperor of the West, as reported by Galileo from his history of the Franks, record that eight centuries earlier in France many people had noticed a dark mark on the sun lasting eight days. The emperor assembled his savants, who agreed that the spot must have been Mercury. Galileo is stern; the shame of those old astronomers who did not know that no transit of Mercury could last even eight hours! (One may suspect that they had simply kept the puzzle's secret to themselves, for a predictable dot already gone is less disturbing than spots of unknown nature.)

Image: Dusan Petricic
By 1629 Kepler's best tables allowed him to predict transits. Attention, astronomers! Two transits for 1631: Mercury in November, Venus weeks later. Mercury's dot (too small to fit Kepler's 1607 daub) was seen by telescope in Paris just before it left the disk about five hours early. "The cunning god ... I found him out," wrote the cunning astronomer Pierre Gassendi to a colleague. The latest transit of Mercury came in the afternoon of November 15, 1999, watched as a pleasure by an astronomer friend of ours in Denver, right on time by present tables.

The 18th century tried hard to use precise timing of Venus transits from different places for scaling the solar system, but the method has long since been supplanted. Mercury transits a dozen or so times each century, next in 2003. Venus trespasses rarely, a century or so passing between successive closely spaced pairs of crossings. The last Venus transit was in 1882; the next will be in 2004.

Transits arise from fortunate remote alignments, but sunspots are physical events, associated with a variety of solar eruptions. Great plasma clouds fly out past our planet and disturb our magnetic weather. At the last spot maximum in 1989, electrical power failed for six hours over most of Quebec as the consequence of strong ground currents. Many Earth satellites risk electronics damage from similar effects in orbit. This year the northern lights are likely to appear once or twice even way down south in Dixie. Spots2K may be more evident than Year2K! Try not to miss the bright colorful auroras, the dark sunspots themselves (safely watched by projection onto paper) or the surprising behavior of short-wave radio.

More than two dozen faint planets around distant sunlike stars have been found, without a glimpse of even one of them. Such finds use slow to-and-fro movements that any close-in big planet enforces on its massive star by their reciprocal gravitational pull. One unseen planet appeared to promise a fair chance of a transit visible here every few days. In so remote a transit no planet dot is to be seen, for distance has shrunk the whole disk of the star to a bright point of light. The transit came on time in mid-November; the star's light dimmed by almost 2 percent when the big planet crossed in front of its star 150 light-years out there. The telltale dimming in starlight (and a repeat on time at the next orbit) wonderfully confirmed the indirect measurements. Certainly a mere dimming is no substitute for detailed images. It now looks as though transits can identify planets as small as Earth, whose meager pulls on their own suns are undetectable.

To catch the rapid dimming requires looking at the right time from the right direction at an unexamined star. How could we hope to win a lottery like that? There is one way: buy plenty of tickets! Translated: your instruments need to stare for years at many, many stars of the sun's kind just as they stand in the sky. Measure brightness star by star, each time ready to catch the rare diagnostic dimming of the light by only one part in 10,000 or 12,000. Miss the few hours of transit, and you have wasted time on that star. Many stars will vary in myriad other ways, but seek one pattern, a sudden minor dimming that lasts for hours and then strictly repeats at much longer intervals. No other star variation is like that. (Don't fail to check out the prediction of the third transit for a star; once it tells you three times, it's true.) Repeat, repeat, repeat, measuring brightness for some years.

By then you can expect to have located up to a few hundred Earth-class planets orbiting their suns among some 150,000 random stars, plus shelves full of other star variations. What is proposed is a space probe, its overall weight under a ton, bearing a telescope one meter in aperture and a few meters long, that looks into a deep pool of stars as the craft orbits the sun, a little outside Earth's own path. The focal plane is filled by 21 of the latest CCD photodetector arrays, in all some 90 million pixels; the probe is precisely guided to hold each star image on its own small patch of pixels. The design is well advanced. If all lab tests go well, launch into solar orbit a few years from now. Kepler--the proposed probe is named after the first predictor of transits--is the work of a team led by William Borucki and David Koch at the NASA Ames Research Center.

Seining a whole sea of stars for the telltale transit dimming pattern looks like the easiest scheme so far for finding Earth-like planets (unless we hear first from remote astronomers in residence).