Until late 1994 this discussion would have been very short indeed, as there was no firm observational evidence for the existence of any brown dwarfs at all! Since then, however, continuing advances in observational hardware and image-processing techniques have begun to bear fruit, and several likely brown dwarf candidates have now been identified.

Detecting a brown dwarf is no easy task. These objects are so faint that they are hard—although not impossible—to detect directly, and most progress to date has been made in finding them when they happen to orbit another star. (Recall from Chapter 17 that most stars are found in binary systems; the same may very well be true of brown dwarfs.) In searching for brown dwarfs in binaries, astronomers have used many of the same techniques employed in the search for extrasolar planets, as discussed more fully in Chapter 15. (Interlude 15-1) By measuring the tiny "wobbles" in a star's motion, either by carefully tracking its position in the sky or by measuring small variations in its radial velocity, researchers can infer not just the presence but also some of the physical properties of the star's unseen companion. The size of the wobbles is truly minuscule—only a few thousandths of an arc second in position and a few tens of meters per second in radial velocity. Nevertheless, such precise measurements are now becoming possible.

The Hubble image below shows Gliese 623, a binary system containing a brown dwarf candidate (marked by an arrow) originally identified by radial velocity measurements. Astronomers hope that continued observations of this system will allow the low-mass companion's mass to be measured with

Actually, one might reasonably ask where the dividing line between brown dwarfs and Jupiter-sized planets really lies. The answer is not clear-cut and may depend on circumstances. Theoretical estimates indicate that objects less than about 10-20 times the mass of Jupiter cannot form by the usual contraction/fragmentation star-formation process, but the upper limit on the mass of a planet (which forms within a disk surrounding a protostar) is not well known. On the other hand, the theory of planet formation suggests that planets are likely to form in roughly circular orbits. (Sec. 15.2) Thus, the best we can say at present is that a low-mass (less than 10 Jupiter masses, say) object in a roughly circular orbit is most likely a planet, whereas a higher-mass object in an eccentric orbit is probably a brown dwarf. Whether the other two possible combinations (low mass/high eccentricity and high mass/low eccentricity) can occur, and how they will be classified if they are observed, remains to be seen.

Recent orbital measurements like those just described have so far produced two very strong brown dwarf candidates, both of them orbiting low-mass M-class stars, and at least three more possible brown dwarfs. (At present, the possibility that the latter three objects are planets cannot be completely ruled out.) In addition, in the past year, infrared and spectroscopic studies of stars have begun to find strong indications of isolated brown dwarfs. Infrared observations are a particularly effective way to search for brown dwarfs because these objects emit most of their radiation in the infrared part of the spectrum, while true stars tend to be brightest at optical or ultraviolet wavelengths.

The images below show the binary star system Gliese 229, first identified as a possible brown dwarf by ground-based infrared observations (left), and recently imaged from space (right). The two objects are only 7" apart; the fainter "star" has a luminosity only a few millionths that of the Sun and an estimated mass about 50 times that of Jupiter. (The bright diagonal streak in the latter image is caused by a hardware problem in the CCD chip used to record it.)

The presence of lithium, an element that is present in interstellar gas but that is rapidly depleted in a star once nuclear fusion begins, is another telltale indicator of brown dwarfs. Researchers using a spectrometer on the giant Keck telescope in Hawaii have so far identified two brown dwarf candidates by this means. Finally, several more candidates have been identified by combining multiple long-exposure images to reveal very faint, very red objects that had hitherto gone undetected.

The next few years may well see a flood of new observations of very-low-mass stars—perhaps even a revolution in our understanding of the entire star-formation process.