Comet Overhead
© Walter Pacholka, Astropics/SPLHale-Bopp, seen here from Joshua Tree National Park, California, was one of the brightest comets of the 20th century.
Few cosmic apparitions have inspired such awe and fear as comets. The particularly eye-catching Halley's comet, which last appeared in the inner solar system in 1986, pops up in the Talmud as "a star which appears once in seventy years that makes the captains of the ships err". In 1066, the comet's appearance was seen as a portent of doom before the Battle of Hastings; in 1456, Pope Callixtus III is said to have excommunicated it.

Modern science takes a more measured view. Comets such as Halley's are agglomerations of dust and ice that orbit the sun on highly elliptical paths, acquiring their spectacular tails in the headwind of charged particles streaming from the sun. We even know their source: they are Kuiper belt objects (KBOs) tugged from their regular orbits by Neptune and Uranus.

But there's a problem. Certain comets, such as Hale-Bopp, which flashed past Earth in 1997, appear simply too infrequently in our skies. Their orbits must be very long, far too long to have an origin in the Kuiper belt. The conclusion of many astronomers is that the known solar system is surrounded in all directions by a tenuous halo of icy outcasts, thrown from the sun's immediate vicinity billions of years ago by the gravity of the giant planets.

This celestial Siberia is known as the Oort cloud, after the Dutch astronomer Jan Oort, who proposed its existence in 1950. This diffuse sphere of material encircling the solar system has never been seen, but if the longest-period comets are anything to go by, it must be vast, reaching out about 1000 times further than the outer edge of the Kuiper belt. At such huge distances, it would not be passing planets that throw the comets sunwards - it would be the tug of the Milky Way and nearby stars. The Oort cloud would truly be where our solar system meets the void.

Unfortunately, if looking for Planet X is difficult, spotting the Oort cloud is a nightmare. It is simply too dim and distant, and its pieces too small, for telescopes to spy. That is unfortunate, as counting and estimating the size of such objects could help in reconstructing a picture of the sun's birthplace, and perhaps provide us with a glimpse of the unadulterated material from which the giant planets were pieced together.

So far, the only information about this primordial rubble comes from stray comets and the largest KBOs, which should have a similar composition. "That's like trying to figure out what a whale looks like from the exposed blowhole on one side and the tip of the fin on the other," says Hal Levison, a planetary scientist at the Southwest Research Institute in Boulder, Colorado.

Even so, mapping the rest of the whale might be only a few decades off. Oort cloud objects should dim and diffract the light coming from distant stars. These occultations last just a fraction of a second, but astronomers can use them to measure the size and distance of the intervening body, a technique already being put to work on KBOs. Flickers induced by turbulence in Earth's atmosphere make the subtle detections of the more distant Oort objects impossible from ground-based detectors, but future space-telescope surveys should be able to detect them in great numbers.

Other mysteries remain. The numbers and trajectories of the long-period comets seen so far suggest that the Oort cloud contains trillions of objects a kilometre across or larger, with a combined mass several times that of Earth. That is more material than our current ideas about the solar system's formation can explain - which means that our models might need a fundamental overhaul, says Levison.