Thousands of physicists, astrophysicists, and astronomers are searching for dark matter, mysterious stuff whose gravity seems to hold the galaxies together. However, an old and highly controversial theory that simply changes the law of gravity can explain a key property of galaxies better than the standard dark matter theory, one astronomer reports. That claim isn't likely to win over many skeptics, but even some theorists who favor the standard theory say the analysis hands them a homework problem they should solve.
"The standard theory should explain this, and it doesn't yet. That's fair to say," says Simon White, a cosmologist at the Max Planck Institute for Astrophysics in Garching, Germany, who was not involved in the current analysis.
In 1933, Swiss astronomer Fritz Zwicky suggested the existence of dark matter when he found that the galaxies in a particular cluster swirl about each other too fast to be bound by their gravity alone. In the 1970s, American astronomer Vera Rubin and others discovered that the stars at the edges of individual galaxies also appear to move too fast to be held by the gravity of the stars in the center. Those outer stars ought to move more slowly than the ones circling closer in—just as Jupiter orbits the sun more slowly than Earth. Instead, the speed of the stars generally increases with the distance from the galactic center, eventually flattening out at a maximum value. That observation seemed to clinch the case for some sort of dark matter.
Or did it? In 1983, Mordehai Milgrom a physicist at the Weizmann Institute of Science in Rehovot, Israel, found that he could explain the so-called galaxy rotation curves without dark matter if he simply assumed that on the galactic scale, dynamics and gravity worked a bit differently from what Isaac Newton postulated. Specifically, Milgrom assumed that for very small accelerations, the square of the acceleration, not just the acceleration, is proportional to the gravitational force.
For the past 28 years, Milgrom's idea, known as Modified Newtonian Dynamics (MOND) has generated a long-simmering debate. Many researchers argue that ever more evidence from clusters of galaxies, the largest scale structure of the universe, and the afterglow of the big bang points to the existence of dark matter . Still, a few researchers counter that when they look at the details, MOND does a better job—at least on the galactic scale.
Now, in the latest shot from the MOND side, Stacy McGaugh, an astronomer at the University of Maryland, College Park, reports that MOND can explain an observed correlation between the mass and the rotation speed of galaxies—that is, the speed of those outer stars—called the baryonic Tully-Fisher relation. MOND researchers had tried to do this before, but for their models to work, they had to make an untested assumption about the relationship between a star's mass and the amount of light it puts out. That assumption introduces a large uncertainty, weakening the argument.
To avoid that problem, McGaugh gathered data from various sources on 47 galaxies that contain more hydrogen gas than stars. The mass of the gas can then be estimated directly. McGaugh made a plot of visible mass versus rotation speed for the galaxies. He then plotted the prediction that comes straight out of MOND in a few lines of algebra. The MOND line went right through the data. "You draw the line and the data fall right on it," McGaugh says. "No muss, no fuss." He reports the result in a paper  in press at Physical Review Letters.
Crucially, McGaugh finds very little scatter in the data—just what would be expected if the mass of gas and stars was directly determining the rotation speed. It's not clear exactly what dark matter models would predict, McGaugh says. However, such models make no strong connection between the amount of visible matter and the rotation speed. Indeed, galaxies with the same mass of dark matter can have different numbers of stars. So it would be surprising if dark matter models yielded such a tight correlation.
"I think the data are good, and the fact that MOND fits is striking," says White, who has worked extensively on simulating the evolution of the universe. "I think Stacy is right in holding this up and saying [to dark matter modelers], 'Look at this [correlation]. Go see if you can explain it.' " Still, White says, dark matter can explain the variations in the afterglow of the big bang and other cosmological data with which MOND struggles.
But whether MOND is right may be beside the point, says Jerry Sellwood, a theoretical astrophysicist at Rutgers University in New Brunswick, New Jersey. "The real strength of Stacy's paper is that it points to something that can't be explained in cold dark matter, irrespective of whether MOND is right." At the least, Sellwood says, McGaugh deserves credit for keeping others honest about what their models can do.