Asteroids are relatively small, rocky objects that revolve around the Sun. Their name literally means "starlike bodies," but asteroids are definitely not stars. They are too small even to be classified as planets. Astronomers often refer to them as "minor planets," or sometimes "planetoids."
Asteroids differ from planets in both their orbits and their size. They generally move on quite eccentric trajectories between Mars and Jupiter, unlike the almost circular paths of the major planets. Few are larger than 300 km in diameter, and most are far smaller—as small as a tenth of a kilometer across. The largest known asteroid, Ceres, is just 1/10,000 the mass of Earth and measures only 940 km across. Taken together, the known asteroids amount to less than 1/10 the mass of the Moon, so they do not contribute significantly to the total mass of the solar system.
European astronomers discovered the first asteroids early in the nineteenth century as they searched the sky for an additional planet orbiting between Mars and Jupiter, where the Titius—Bode "law" (see Interlude 6-1) suggested one might be found. Italian astronomer Giuseppe Piazzi was the first to discover an asteroid. He detected Ceres in 1801 and measured its orbital semi-major axis to be 2.8 A.U.—exactly where the "law" predicted. Within a few years, three more asteroids—Pallas (at 2.8 A.U.), Juno (at 2.7 A.U.), and Vesta (at 3.4 A.U.)—were discovered.
By the start of the twentieth century, astronomers had cataloged several hundred asteroids with well-determined orbits. Now, at the end of that century, the list numbers over 7000. The total number of potentially visible asteroids (that is, asteroids that would be visible through present-day telescopes if we knew just where and when to look) may exceed 100,000. The vast majority of these bodies are found in a region of the solar system known as the asteroid belt, located between 2.1 and 3.3 A.U. from the Sun—roughly midway between the orbits of Mars (at 1.5 A.U.) and Jupiter (at 5.2 A.U.). All but one of the known asteroids revolve about the Sun in prograde orbits, in the same sense as the planets. The overall layout of the asteroid belt is sketched in Figure 14.1.
Figure 14.1 The asteroid belt, along with the orbits of Earth, Mars, and Jupiter (not drawn to scale). The main belt, the Trojan asteroids, and some Apollo (Earth-crossing) and Amor (Mars-crossing) orbits are shown.
Such a compact concentration of asteroids in a well-defined belt suggests that they are either the fragments of a planet broken up long ago or primal rocks that never managed to accumulate into a genuine planet. On the basis of the best evidence currently available, researchers favor the latter view. There is far too little mass in the belt to constitute a planet, and the marked chemical differences among individual asteroids strongly suggest that they could not all have originated in a single body. Instead, astronomers believe that the strong gravitational field of Jupiter continuously disturbs the motions of these chunks of primitive matter, nudging and pulling at them, preventing them from aggregating into a planet.
With few exceptions, asteroids are too small to be resolved by Earth-based telescopes, so astronomers must rely on indirect methods to find their sizes, shapes, and composition. Consequently, only a few of their physical and chemical properties are accurately known. To the extent that astronomers can determine their compositions, asteroids have been found to differ not only from the nine known planets and their many moons, but also among themselves.
Asteroids are classified in terms of their spectroscopic properties. The darkest, or least reflective, asteroids contain a large fraction of carbon in their makeup. They are known as C-type (or carbonaceous) asteroids. The more reflective S-type asteroids contain silicate, or rocky, material. Generally speaking, S-type asteroids predominate in the inner portions of the asteroid belt, and the fraction of C-type bodies steadily increases as we move outward. Overall, about 15 percent of all asteroids are S-type, 75 percent are C-type, and 10 percent are other types (mainly the M-type asteroids, containing large fractions of nickel and iron). Many planetary scientists believe that the carbonaceous asteroids consist of very primitive material representative of the earliest stages of the solar system and have not experienced significant heating or chemical evolution since they first formed 4.6 billion years ago.
In most cases, astronomers estimate the sizes of asteroids from the amount of sunlight they reflect and the amount of heat they radiate. These observations are difficult, but size measurements have been obtained in this way for more than 1000 asteroids. On rare occasions, astronomers witness an asteroid occulting a star, allowing them to determine its size and shape with great accuracy. The largest asteroids are roughly spherical, but the smaller ones can be highly irregular.
The three largest asteroids, Ceres, Pallas, and Vesta, have diameters of 940, 580, and 540 km, respectively. Only two dozen or so asteroids are more than 200 km across, and most are much smaller. Almost assuredly, there exist hundreds of thousands more awaiting discovery. However, observers estimate that they are mostly very small. Probably 99 percent of all asteroids larger than 100 km are known and cataloged, and at least 50 percent of asteroids larger than 10 km are accounted for. Although the vast majority of asteroids are probably less than a few kilometers across, most of the mass in the asteroid belt resides in objects greater than a few tens of kilometers in diameter.
The first close-up views of asteroids were provided by the Jupiter probe Galileo which, on its rather roundabout path to the giant planet, passed twice through the asteroid belt, making close encounters with asteroid Gaspra in October 1991 and asteroid Ida in August 1993 (Figure 14.2). 
(Sec. 6.6) Both Gaspra and Ida are S-type asteroids. Technical problems limited the amount of data that could be sent back from the spacecraft during the flybys. Nevertheless, the images produced by Galileo show far more detail than any photographs made from Earth.
Figure 14.2 (a) The S-type asteroid Gaspra as seen from a distance of 1600 km by the probe Galileo on its way to Jupiter. (b) The S-type asteroid Ida, photographed by Galileo from a distance of 3400 km. (Ida's moon, Dactyl, is visible at the right of the photo.) The resolution in these photographs is on the order of 100 m. True-color images showed the surfaces of both bodies to be a fairly uniform shade of gray. Sensors onboard the spacecraft indicated that the amount of infrared radiation absorbed by these surfaces varies from place to place, probably as a result of variations in the thickness of the dust layer blanketing them.
Gaspra and Ida are irregularly shaped bodies with maximum diameters of about 20 and 60 km, respectively. They are pitted with craters ranging in size from a few hundred meters to 2 km across and are covered with a layer of dust of variable thickness. Ida is much more heavily cratered than Gaspra, in part because it resides in a denser part of the asteroid belt. Also, scientists believe that Ida has suffered more from the ravages of time. Ida is about a billion years old, far older than Gaspra, which is estimated to have an age of just 200 million years, based on the extent of cratering. Both asteroids are thought to be fragments of much larger objects that broke up into many smaller pieces following violent collisions long ago.
To the surprise of most mission scientists, closer inspection of the Ida image (Figure 14.2b) revealed the presence of a tiny moon, now named Dactyl, just 1.5 km across, orbiting the asteroid at a distance of about 90 km. A few such binary asteroids had previously been observed from Earth (for example, the second-largest asteroid, Pallas, is a binary). However, the rare binary systems known before the Ida flyby were all much larger than Ida—a moon the size of Dactyl cannot be detected from the ground. Scientists believe that, given the relative congestion of the asteroid belt, collisions between asteroids may be quite common, providing a source of both interplanetary dust and smaller asteroids and possibly deflecting one or both of the bodies involved onto eccentric, Earth-crossing orbits. The less violent collisions may be responsible for the binary systems we see.
By studying the Galileo images, astronomers were able to obtain limited information on Dactyl's orbit around Ida and hence (using Newton's law of gravity—see More Precisely 2-3) to estimate Ida's mass at about 5—10 
1016 kg. This information in turn allowed them to measure Ida's density as 2200—2900 kg/m3, a range consistent with its rocky, S-type classification.
As of 1998 only one other asteroid has been studied at such close range. In June 1997 the Near Earth Asteroid Rendezvous (NEAR) spacecraft visited the C-type asteroid Mathilde on its way to the mission's main target, the asteroid Eros. Shown in Figure 14.3, Mathilde is some 60 km across. By sensing its gravitational pull, NEAR measured Mathilde's mass to be about 1017 kg, implying a density of just 1400 kg/m3. To account for this low density, scientists speculate that the asteroid's interior must be quite porous, perhaps like styrofoam in structure. The interior's relatively soft consistency may also help explain the unexpectedly large size of many of the craters observed on Mathilde's surface.
Figure 14.3 The C-type asteroid Mathilde, imaged by the NEAR spacecraft en route to the near-earth asteroid Eros. It measures some 60 50 km, and rotates every 17.5 days. The largest craters visible in this image are about 20 km across—much larger than the craters seen on either Gaspra or Ida. The reason may be the asteroid's low-density (approx. 1400 kg/m3) and rather soft composition.
On reaching Eros in February 1999, NEAR will go into orbit around it, coming as close as 24 km to the surface. For a period of 1 year, the spacecraft will make detailed measurements, at many different electromagnetic wavelengths, of the S-type asteroid's size, shape, gravitational field, composition, surface and internal structure, and magnetic field. Scientists back on Earth are eagerly awaiting the most detailed view yet of one of Earth's smallest neighbors.
Apart from those of Ida, Mathilde, and (soon) Eros, most asteroid masses are unknown. However, a few of the largest asteroids do have strong enough gravitational fields that their effects on their neighbors can be measured and their masses thereby determined to reasonable accuracy. Their computed densities are generally compatible with the rocky or carbonaceous compositions just described.
The orbits of most asteroids have eccentricities lying in the range 0.05—0.3, ensuring that they always remain between the orbits of Mars and Jupiter. Very few asteroids have eccentricities greater than 0.4. Those that do are of particular interest to us, however, as their orbits may intersect Earth's orbit, leading to the possibility of a collision with our planet. They are collectively known as Earth-crossing asteroids. Those stray asteroids having very elliptical orbits or orbits well inside the main asteroid belt have probably been influenced by the gravitational fields of nearby Mars and especially Jupiter. These planets can disturb normal asteroid orbits, deflecting them into the inner solar system. Asteroids whose paths cross Earth's orbit are termed Apollo asteroids (after the first known Earth-crossing asteroid, Apollo) if their orbital semi-major axes exceed 1 A.U., and Aten asteroids otherwise. Those crossing only the orbit of Mars are known as Amor asteroids. (See again Figure 14.1.)
Although we are currently aware of only a few dozen Earth-crossing asteroids, they are among the best known because of their occasional close encounters with Earth. For example, the perihelion of the Aten asteroid Icarus (Figure 14.4) is within 0.2 A.U. of the Sun. On its way past Earth in 1968 it missed our planet by "only" 6 million km—a very close call by cosmic standards. More recently, in 1991 an unnamed asteroid (designated 1991 BA) came much closer, passing only 170,000 km from Earth, less than half the distance to the Moon. In December 1994 the Apollo asteroid 1994 XM1 missed us by a mere 105,000 km.
Figure 14.4 The asteroid Icarus has an orbit that passes within 0.2 A.U. of the Sun, well within the orbit of Earth. Icarus occasionally comes close to our own planet, making it one of the better-studied asteroids in the solar system. The asteroid's motion relative to the stars makes it appear as a streak (marked) in this long-exposure photograph, taken in the 1970s.
The potential for collision with Earth is real. Calculations imply that most Earth-crossing asteroids will eventually collide with Earth. On average during any given million-year period, our planet is struck by about three asteroids. Because Earth is largely covered with water, on average two of those impacts should occur in the ocean and only one on land. Several dozen large land basins and eroded craters on our planet are suspected to be sites of ancient asteroid collisions (see, for example, Figure 14.17 later in this chapter). The many large impact craters on the Moon, Venus, and Mars are direct evidence of similar events on other worlds.
Most known Earth-crossing asteroids are relatively small—about 1 km in diameter (although one 10 km in diameter has been identified). Even so, a visit of even a kilometer-sized asteroid to Earth could be catastrophic by human standards. Such an object packs enough energy to devastate an area some 100 km in diameter. The explosive power would be equivalent to about a million 1-megaton nuclear bombs, a hundred times more than all the nuclear weapons currently in existence on Earth. A fatal blast wave would doubtless affect a much larger area still. Should an asteroid hit our planet hard enough, it might even cause the extinction of entire species—indeed, many scientists think that the extinction of the dinosaurs was the result of just such an impact (see Interlude 14-1). Some astronomers take the prospect of an asteroid impact seriously enough to advocate an "asteroid watch"—an effort to catalog and monitor all Earth-crossing asteroids in order to maximize our warning time of any impending collision.
Although most asteroids orbit in the main belt, between about 2 and 3 A.U. from the Sun, an additional class of asteroids, called as the Trojan asteroids, orbit at the distance of Jupiter. Several hundred such asteroids are now known. They are locked into a 1:1 orbital resonance with Jupiter by that planet's strong gravity, just as some of the small moons of Saturn share orbits with the medium-sized moons Tethys and Dione, as described in Chapter 12. (Sec. 12.5)
Calculations first performed by the French mathematician Joseph Louis Lagrange in 1772 show that there are exactly five places in the solar system where a small body can orbit the Sun in synchrony with Jupiter (or, for that matter, with any other planet) subject to the combined gravitational influence of both large bodies. These places are known as the Lagrange points of Jupiter's orbit. As illustrated in Figure 14.5, three of these points (referred to as L1, L2, and L3) lie on the line joining Jupiter and the Sun (or its extension in either direction). The other two—L4 and L5—are located on Jupiter's orbit, exactly 60° ahead of and behind the planet. All five Lagrange points revolve around the Sun at the same rate as Jupiter.
Figure 14.5 The Lagrange points of the Jupiter—Sun system, where a third body could orbit in synchrony with Jupiter on a circular trajectory. Only the L4 and L5 points are stable. They are the locations of the Trojan asteroids.
In principle, an asteroid placed at any of the Lagrange points will circle the Sun in lockstep with Jupiter, always maintaining the same position relative to the planet. However, the three Lagrange points in line with Jupiter and the Sun are known to be unstable—a body displaced, however slightly, from any of those points will tend to drift slowly away from it, not back toward it. Since matter in the solar system is constantly subjected to small perturbations—by the planets, the asteroids, and even the solar wind—matter does not accumulate in these regions. No asteroids orbit near the L1, L2, or L3 points of Jupiter's orbit.
This is not the case for the other two Lagrange points, L4 and L5. They are both stable—matter placed near them tends to remain in the vicinity. Consequently, asteroids tend to accumulate near these points. For unknown reasons, Trojan asteroids tend to be found near Jupiter's leading (L4) Lagrange point rather than the trailing (L5) point. Recently, a few small asteroids have been found similarly trapped in the Lagrange points of Venus, Earth, and Mars.
The main asteroid belt also has structure—not so obvious as the Trojan orbits or the prominent gaps and ringlets in Saturn's ring system, but nevertheless of great dynamic significance. A graph of the number of asteroids having various orbital semi-major axes (Figure 14.6a) shows that there are several prominently underpopulated regions in the distribution. These "holes" are known as the Kirkwood gaps, after their discoverer, the nineteenth-century American astronomer Daniel Kirkwood.
Figure 14.6 (a) The distribution of asteroid semi-major axes shows some prominent gaps caused by resonances with Jupiter's orbital motion. Note, for example, the prominent gap at 3.3 A.U., which corresponds to the 2:1 resonance—the orbital period is 5.9 years, exactly half that of Jupiter. (b) An asteroid in a 2:1 resonance with Jupiter receives a strong gravitational tug from the planet each time they are closest together (as in panels 1 and 3). Because the asteroid's period is precisely half that of Jupiter, the tugs come at exactly the same point in every other orbit, and their effects reinforce each other.
The Trojan asteroids share an orbit with Jupiter—they orbit in 1:1 resonance with that planet. The Kirkwood gaps result from other, more complex, orbital resonances with Jupiter. For example, an asteroid with a semi-major axis of 3.3 A.U. would (by Kepler's third law) orbit the Sun in exactly half the time taken by Jupiter. (Sec. 2.4) The gap at 3.3 A.U., then, corresponds to a 2:1 resonance. An asteroid at that particular resonance feels a regular, periodic tug from Jupiter at the same point in every other orbit (Figure 14.6b). The cumulative effect of those tugs is to deflect the asteroid into an elongated orbit—one that crosses the orbit of Mars or Earth. Eventually, the asteroid collides with one of those two planets or comes close enough that it is pushed onto an entirely different trajectory. In this way, Jupiter's gravity creates the Kirkwood gaps, and some of the cleared-out asteroids become Apollo or Amor asteroids.
Notice that although there are many similarities between this mechanism and the resonances that produce the gaps in Saturn's rings (see Chapter 12), there are differences too. (Sec. 12.4) Unlike Saturn's rings, where eccentric orbits are rapidly circularized by collisions among ring particles, there are no physical gaps in the asteroid belt. The in-and-out motion of the belt asteroids as they travel in their eccentric orbits around the Sun means that no part of the belt is actually empty. Only when we look at semi-major axes (or, equivalently, at orbital energies) do the gaps become apparent.
