A Matter of Fact?
Alternative Galaxy Glues
A basic principle of sound science is that if a simple and familiar cause is capable of explaining what's observed, there's no justification in complicating things further or indulging in exotic, speculative inventions. Going with the simplest explanation won't guarantee that you're right; but it's the way to bet.
Observations show that galaxies don't rotate the way they should if nothing but gravity were holding them together. Whereas the familiar Newtonian law says that the velocity of stars in their paths around the galactic center should increase steadily as the distance from the center increases, measurements show that their velocities increase sharply in the central region to a value that then remains practically constant over the remaining distance to the rim. Besides posing the problem of a prediction that isn't fulfilled, a further implication is that the amount of observed mass making up the galaxies isn't enough to hold them together, and they ought to be flying apart.
The standard response strives to preserve accepted theory and posits the existence of "dark matter" as an invisible presence, detectable only by its ability to induce the effect that it was contrived to explain. Invoking dark matter--and more recently, "dark energy"--to account for observations that regular theory has trouble dealing with, from violently energetic cosmic events to complex structures spanning light-years, has become something of a fad, and at the latest count it came in seven different varieties. You put it wherever you want and assign it the properties it needs to produce the desired result. It has been dubbed "cosmic duct tape"--capable of fixing anything.
However, two papers have come to my attention which suggest that the rush into realms of the unseeable and untestable might be a bit premature. The first is "From Dark Matter to MOND," by R.H. Sanders of the Kapteyn Astronomical Institute, University of Groningen, the Netherlands, online at http://arxiv.org/abs/0806.2585. MOND stands for MOdified Newtonian Dynamics, and takes the approach that the simple and satisfying inverse square law that works fine for things like putting up satellites and calculating trajectories of bodies in the Solar System (although the anomalies measured in the progress of space probes reaching its outer regions might be calling that into question) could be in need of modification when applied over greater distances. This seems reasonable to me, since the only observational data we have of how thing move on larger scales is that from such measurements as the motions of stars in galaxies, and the observational data says that the simplest form of the law doesn't work. Sanders's paper shows how modifying Newton's value for gravitational acceleration in a way that becomes significant only at higher values (reminiscent of they way in which Relativity modifies Newtonian dynamics at high velocities) produces a stellar velocity vs. distance curve that matches observations extremely closely without need of dark matter.
The other paper is "Galaxy Rotation Curves Without Non-Baryonic Dark Matter and Modifications to Gravity: Effect of the Ampere Force," by David Tsiklauri of the University of Salford, Greater Manchester, UK, online at http://arxiv.org/abs/0806.1513v1. Tsiklauri shows that it's not even necessary to modify the law of gravity. Results matching observations can be obtained based on the amount and distribution of observed matter, by taking account of the effect of the ampere force term of electrodynamics, known for over two hundred years.
I'm sure that William of Occam would have approved.