In the 1920s scientists discovered that the universe is expanding. Alexander Friedmann had proved that Einstein's general theory of relativity required space to spread out with time, and Edwin Hubble had observed that it actually did expand. Friedmann's equations, however, predicted three possible types of expansion. In two of them space becomes forever larger in slightly different ways. In the third expansion gradually slows, stops, and finally contracts. That last universe is called "closed." The parameter that makes the difference between the three situations is the amount of mass in the universe. Too much mass and the universe collapses. Too little mass and the expansion becomes runaway, virtually blowing the universe apart. When, like the Baby Bear's porridge, the mass of the universe is just right, the expansion continues forever but gradually slows down, keeping the universe much as we know it today. This is called the "flat universe."
Astronomers since World War II have been working on the problem of determining the amount of mass. Most astronomers prefer a universe that collapses or is flat. But when the amount of mass is calculated from known observable galaxies and dust, it is far too small. If these calculations are correct, the universe will undergo runaway expansion.
There is, however, mass in the universe that we cannot observe. Vera Rubin has shown that galaxies rotate too quickly near their edges. According to the laws of gravity and rotation, if the observable boundary of a galaxy were the real edge, rotational speed would be reduced in parts of the galaxy away from the center. Since Rubin's observations establish that this is not the case, there must be great clouds of unobserved mass in which each galaxy is embedded. She could even calculate how much unobserved mass there is for that galaxy.
But the newly found mass in galaxies still is not enough by a large margin to close or at least flatten the universe. The inflationary universe of Alan Guth, first conceived of in 1979 and improved on by other physicists since, made the task a little easier because their predicted universe needed slightly less mass to close or flatten it -- but still a lot more than anyone could find. In fact, a recent effort (in 2003) to determine the proportions of missing mass to observable mass found nearly six times as much unseen as seen. Astronomers and physicists have postulated many entities that could provide the missing mass. Astronomers have contributed brown dwarfs (objects in size and energy between a planet and a star) and black holes. But physicists have provided most of the new ideas, suggesting mass that cannot be observed at all (except for gravitational effects).
For example, the universe is presumed to be filled with nearly undetectable neutrinos that are now known to have a small mass each. Although the exact amount of mass has been hard to pin down, neutrinos are not a satisfactory source of the missing mass. They move at nearly the speed of light, which prevents them from clumping together. Galaxies are thought to start when a large amount of mass accidentally forms a clump, and since we have galaxies, there should also be clumps of missing mass to initiate them.
If not neutrinos, then perhaps some undiscovered particle could provide the missing mass. Particles predicted by various physicists include magnetic monopoles, photinos, axions, gravitinos, and winos. Most of these are grouped together as WIMPS -- weakly interacting massive particles -- which means that they are hard to detect if present and also hard to produce in particle accelerators (since the more massive a particle is, the more energy it takes to create it).
Yet another idea from physics is the cosmic string, first proposed in 1976. A cosmic string is not a particle, but an incredibly long, skinny bit of leftover energy from the big bang. Even a couple of kilometers of a cosmic string has a mass similar to that of Earth -- and some cosmic strings are expected to reach from one end of the universe to another. One of the nice things about cosmic strings as a way to supply the missing mass is that, unlike the various particles, they are in principle observable. So far, however, possible sightings have been few and not convincing.