Most of the mass in the universe is missing. Or is it merely hidden in some exotic, as yet undetectable form? No one is sure which. One thing is sure, though. The problem of the missing mass has gotten to the point where it is more than just a problem. It is an embarrassment, an obstacle to understanding such things as the structure of galaxies, the evolution of clusters of galaxies, and the ultimate fate of the universe.
A simple analogy illustrates the problem. Suppose the rockets inserting a spacecraft into an orbit around Earth were to burn too long, providing too much thrust. Then the gravitational pull of Earth would be overcome, and the spacecraft would shoot out of orbit into interplanetary space. Fortunately for astronauts, scientists can calculate quite accurately how much thrust is needed for a given orbit, so this does not happen. But suppose, through a computer error, the rockets burned too long and the spacecraft was accelerated to a speed twice as fast as the proper orbital speed-yet the spacecraft stayed in orbit! You would be forced to conclude that either Earth had more mass than you had supposed and hence a stronger gravitational pull, or that the theory you had used to make the calculation was in error.
This is about the situation astrophysicists find themselves in today. Not in trying to understand the motion of planets around the Sun-the theory works fine there-but in trying to understand the motions of stars and gas in the outer regions of galaxies, or of galaxies and gas in clusters of galaxies.
In the past few years astronomers have painstakingly measured the rate at which stars and gas clouds in the outer parts of spiral galaxies are orbiting the center of mass of those galaxies. Optical photographs show spiral galaxies to be graceful pinwheels of billions of stars, with the light falling off steadily from the central to the outer regions. Since the light is produced by stars, we naturally expect the matter and its associated gravitational force field to show a similar concentration. It follows, then, that the speed of rotation of the stars and gas should decline as one moves from the inner to the outer regions of galaxies.
Much to the surprise and consternation of astronomers, this is not what is observed. As radio and optical observations have extended the velocity measurements for the stars and gas to the outer regions of spiral galaxies, they have found that the stars and gas clouds are moving at the same speed as the ones closer in! A substantial part of the mass of the galaxy is not concentrated toward the center of the galaxy but must be  distributed in some dark, unseen halo surrounding the visible galaxy. The outer regions of galaxies, faint and inconspicuous on a photograph, may actually contain most of the matter. In the words of astronomers Margaret and Geoffrey Burbidge, it appears that "the tail wags the dog."
Just how large is this unseen halo? Why can't it be seen? No one knows the answer to either question. What is known is that the problem involves more than a few isolated galaxies. Most of the spiral galaxies in which the rotation pattern has been studied in detail, including our own Milky Way Galaxy, show evidence for an extensive halo of dark matter.
Nor is the problem confined to spiral galaxies. Perhaps the most spectacular evidence for a halo of dark matter around a galaxy comes from the giant spherically shaped galaxy, M87. X-ray observations show that M87 is enveloped in a cloud of hot, X-ray emitting gas nearly a million light years across. If this hot gas is not confined somehow, it will expand. In about 100 million years, it would disperse. Although this may seem like a long time, it is only 1 percent of the total lifetime of the galaxy. To account for the gas cloud as it is now observed, there are three possibilities: (1) some force is confining the gas to the galaxy; (2) the gas is being continuously replenished; or (3) we are observing the galaxy at a special time in its history, before the gas has had time to disperse. The third alternative is possible but improbable. The second not only requires an exorbitant amount of energy but also implies that the hot cloud should be spread out over a much larger volume of space than is observed. That leaves the first alternative, confinement by a force. The confining force could either be gravity or the pressure of an even hotter gas outside the M87 halo. Observations from the HEAOs rule out this latter possibility. That leaves gravity.
This is an important result. It means that X-ray observations can be used to measure the gravitational forces around galaxies. From the distribution of the X-ray brightness of the gas cloud, one can estimate the distribution of the gas in space. From that distribution, the mass needed for gravitational confinement can be estimated. Observations with HEAO 2 imply the presence of a halo of dark matter containing the mass of 30 trillion suns! This is several hundred times the mass observed in the disk of large spiral galaxies such as ours and the Andromeda Galaxy and about 30 times larger than the previous estimates of the mass of M87.
The same principles can be used to measure the gravitational field on a much larger scale. X-ray observations of clusters of galaxies show that the mass needed to confine the hot gas observed in clusters of galaxies is about 5 or 10 times greater than the mass that can be detected in these clusters through observation in any wavelength band, from radio through X-ray. This is in agreement with optical observations. They show that the motions of galaxies orbiting around the center of mass of the cluster can be understood only if the gravitational field is much stronger than would be deduced from the amount of detectable matter. That is, they imply that about 80 to 90 percent of the mass of the cluster has escaped detection.
On an even larger scale, studies of the motion of the Local Group of galaxies that includes our Milky Way Galaxy indicate that we are part of a supercluster of galaxies. An analysis of the motion of the Local Group suggests that a large amount of hidden matter is necessary to provide the gravitational force needed to keep the supercluster from flying apart. The amount of missing mass is about 10 times the amount of visible mass.
In summary then, radio, optical, and X-ray observations of galaxies, clusters of galaxies, and superclusters of galaxies indicate that 80 to 90 percent of the matter is either missing or hidden from view. If this ratio holds throughout the universe, then our ideas as to the ultimate fate of the universe may be profoundly affected. In a large measure, the fate of the universe is determined by the mass density of the universe, that is, the amount of mass in a unit volume. If the mass density is larger than a certain critical value, the expansion of the universe that began with the initial "big bang" will not continue forever but will slow down, and the universe will collapse. The endpoint of such a collapse is unknown. The universe could collapse forever into a universe-sized black hole, or it could go through an unending cycle of expansion, collapse, and reexpansion. On the other hand, if the mass density is too low, the universe will expand forever; it will be "open." Current estimates indicate that the mass density of the universe falls short of the critical density by a factor of 10 or more, implying that the  universe will expand forever. However, if the mass density is 10 times greater than it appears to be, as suggested by the missing mass mystery, then the universe may be closed after all. Seen in this light, the hidden mass problem becomes a very big problem indeed.
What is the answer to the problem? Is something wrong with our understanding of gravity? Is there some additional force that comes into play over these very large scales, a force that is missing from our calculations of the orbits or of the confinement of hot gas? Or is the universe full of dark matter that has escaped detection? Although attempts have been made to modify gravitational theory in the required way, most of the effort has been concentrated into ways that the matter could be hidden from view.
Astronomers have searched long and hard for this matter. They have used radio, infrared, optical, ultraviolet, and X-ray telescopes to scan the outer regions of galaxies and the intergalactic spaces for enough cool gas, hot gas, or dust. They have found some of each, but not enough to solve the problem of the missing mass.
A large population of white dwarfs, neutron stars, or black holes could remain hidden from the view of optical telescopes, but they would have to be 50 to 100 times more abundant on the outer edges of galaxies than in the regions of our galaxy that have been carefully observed so far. No plausible explanation as to why this should be has been advanced. Furthermore, if the population of collapsed stars were in fact 50 or more times larger in the outer regions, we might expect to find far more X-ray sources in the outer regions of galaxies than are observed. In addition, heavy elements ejected from these stars prior to their collapse should be 50 or more times more abundant in the outer regions of galaxies than in the inner regions. This is just the opposite of what is observed. Thus collapsed stars are unlikely candidates to explain the missing mass.
Another durable suggestion has been that a major part of the missing mass in galaxies and clusters of galaxies is made up of very low mass stars. These stars, which would have masses of only a few percent that of the Sun, are red, brown, and black dwarf stars. These stars are very dim because of their small size and low surface temperature. The red dwarfs, which have a mass of 10 to 50 percent that of the Sun, are known to be very common in the solar neighborhood. Of the 90 nearest stars to the Sun that have been classified, 62 of them are red dwarfs. Red dwarf stars produce intense radio, optical, and X-ray flares. This property should make it possible for advanced X-ray telescopes, working in concert with the Space Telescope, to attack the question as to whether 90 percent of the matter on the edges of galaxies is in the form of red dwarf stars.
Brown and black dwarfs are a much more difficult proposition. These objects, which are essentially freely wandering Jupiter-like objects, are so dim that it may be impossible to ever detect them. Although there are no sound theoretical reasons for believing that they exist in the required  numbers, it is possible that such objects were produced in large numbers by the star formation process in globular clusters long ago, when galaxies were just beginning to form. The black and brown dwarfs may then have diffused out of their star clusters and formed very large halos around galaxies. Because of their low luminosities, they would be extremely difficult to detect, even if there were quadrillions of them around every galaxy.
One argument against the missing mass being in the form of normal matter of any type comes from cosmologic considerations. According to the big-bang model, the deuterium (heavy hydrogen) that is observed to exist in interstellar space was created about three minutes after the "beginning" in a billion degree broth of neutrons, protons, photons, and neutrinos. But if the broth were too thick, that is, if the mass density were too high, the deuterium would have all been processed into helium. The greater the mass density, the greater the fusion of deuterium nuclei into helium nuclei, and the less deuterium remains. By observing the amount of deuterium in interstellar space, we can get an idea as to the mass density of normal matter in the universe. The observations suggest that the mass density of normal matter is at most 10 percent of the value needed to turn around the present expansion.
This result lends support to yet another hypothesis for the missing mass, namely, that it is in the form of neutrinos. Neutrinos are elusive subatomic particles that are produced in certain nuclear reactions. Nuclear reactions of the type that produce neutrinos are thought to have been so common in the early universe that many cosmologists have believed for some time that we are literally awash in a sea of neutrinos.
Until recently, however, it did not seem to matter much, because neutrinos were thought to be particles with some energy but no mass, in the same way that photons have energy but no mass. Since it was thought that the energy of the neutrinos was by now quite low, the great abundance of neutrinos was of no practical consequence, or so it seemed.
Then a recent experiment suggested that the neutrino might have a very small mass. The mass of an individual neutrino might be very small, 100 million times smaller than that of a hydrogen atom. Yet, because there are so many neutrinos in the universe, their combined mass could dominate the universe! Thus, the solution to astronomy's greatest riddles, that of the missing mass, might have been found, not by studying distant galaxies, but in a series of experiments right here on Earth.
Serious questions about the neutrino hypothesis must still be answered. For one thing, further experiments have clouded the issue as to whether neutrinos really have mass, and if so how much. There is also a problem in understanding how it is possible for matter to form into galaxy-sized clumps in a universe dominated by fast-moving neutrinos. An analysis of this question suggests that clumps the size of superclusters would form first and that galaxies and clusters of galaxies would condense from these clumps. Yet the  HEAO observations of clusters of galaxies indicate that just the opposite happened. The neutrino hypothesis also suggests that the fraction of missing mass around galaxies should be much less than in clusters of galaxies. This is apparently not observed. These problems have led some astrophysicists to postulate that the existence of yet another particle, the gravitino, is responsible for the missing mass. Gravitinos would have been formed in the very early universe, less than about one millionth of a second after the expansion began, when the temperature was around 100 billion degrees. These particles, which are expected to be more massive than neutrinos, would condense into galaxy-sized clumps. The theory therefore predicts that the fraction of missing mass around galaxies is about the same as in clusters of galaxies. This is what the data now available suggest-a point in favor of the gravitino hypothesis. However, the data are sparse, and no one believes that the final answer is in. More data and calculations are needed.
Thus, the plot thickens, and the number of suspects multiplies in the mystery of the missing mass. And why not? That's the way a good mystery should read, and this is one of the best around.