It is not yet possible to image any of these newly discovered planets. The techniques used to find them are indirect, based on analysis of light from the parent star, not light from the unseen planet. As a planet orbits a star, gravitationally pulling first one way and then the other, the star wobbles slightly. The more massive the planet or the less massive the star or the closer the planet to the star, the greater the star's movement. If the wobble happens to occur along our line of sight to the star, then we see small fluctuations in the star's velocity, which can be measured using the Doppler effect. (Sec. 3.5) Alternatively, if the wobble is predominantly perpendicular to our line of sight, then the star's position in the sky changes slightly from night to night. The star's motion is very small, which is why unambiguous measurements have been so hard to obtain. However, both types of wobble have now been seen and confirmed.

The first figure shows two sets of data that betray the presence of planets. The top graph shows the line-of-sight velocity of the star 51 Pegasi, a near-twin to our Sun lying some 40 light years away. These data were acquired in 1994 by Swiss astronomers using the 1.9-m telescope at Haute-Provence Observatory in France. The regular 50-m/s fluctuations in the star's velocity have been confirmed by several groups of astronomers and imply that a planet at least half the mass of Jupiter orbits 51 Pegasi with a period of just 4.2 days. The bottom declination (see More Precisely 1-2) of the star Lalande 21185, corrected for the motion of Lalande through the Milky Way Galaxy and the motion of Earth around the Sun over the last half century. A 30-year-period wobble at the level of about 0.01" can be seen—about the amount of

Currently, these techniques can be used only to detect planets that are at least as large as Jupiter. Thus far, none of the reports of nearby extrasolar planets has claimed an object of Earth dimensions. In fact, many of the findings to date indicate "hot Jupiters"—gigantic planets surprisingly close (within 1 A.U.) to their parent star. Furthermore, only one such unseen planet has been found (and confirmed) in each system, although tentative analysis of the Lalande data suggests that there may be as many as three Jupiter-sized planets, each with a period of several years. So do not think that the newly discovered planetary systems closely resemble our own solar system. At least so far, they don't—and that has theorists puzzled.

The second figure represents the latest challenge to theory. The HST image shows what may be our first-ever view of an extrasolar planet—the tiny dot at bottom left (indicated by the arrow) is thought be some researchers to be a Jupiter-sized object escaping from the double-star system (top center) where it apparently formed. Theorists speculate that the planet might have been ejected following a close encounter with one of its two parent stars; the long filament of gas may mark the planet's path. The problem is that that stars are estimated to be just 300,000 years old, which is far too little time for a jovian world to form by accretion. However, if the planet formed direclty, through instabilities in the prestellar nebula (Sec. 15.2), then both its unstable orbit and its rapid formation may be explained.

The below chart summarizes all the confirmed extrasolar planets found as of mid-1998. Most appear to fall into the hot-Jupiter category, although some may well turn out to be brown dwarfs—"failed stars" having insufficient mass to become true stars (see Interlude 19-1). The dividing line between genuine Jupiter-like planets and brown dwarfs is unclear, but it is thought to be around five Jupiter masses.