jackmell said:
Just seems like a tough call to me.
Well, it was. But the science is pretty solid here. The best evidence, I think, is in the cosmic microwave background, which was emitted back before any structures in the universe formed, and as a result its behavior is independent of the messy physics that deals with the formations of galaxies and stars and the like. The physics back before the CMB was emitted were much simpler, and therefore the detection is much cleaner than worrying about the very complicated nearby universe.
And what happened in the early universe was we had a situation where dark matter and normal matter behaved extremely differently, because normal matter experienced pressure, while dark matter did not (because dark matter doesn't interact much with anything, while normal matter at the time was a plasma, and plasmas interact very strongly with light). What we see when we look at the CMB is the imprint of pressure waves, whose effects are most clearly visualized by looking at the power spectrum of the CMB:
http://lambda.gsfc.nasa.gov/product...nyear/powspectra/images/med/dl7_f01_PPT_M.png
Here you see the power spectrum as measured by WMAP. The first peak represents pressure waves that had just enough time in the early universe to collapse inward, but not enough time to bounce back.
The second peak represents pressure waves that have a smaller wavelength, so that they could oscillate faster, and had just enough time to collapse inward then bounce back. Notice that it is vastly smaller than the first peak. Part of this is because there is an overall decreasing trend to the CMB power spectrum (because our image of it is a little bit blurry). But most of it is because of dark matter: dark matter experiences no pressure and doesn't bounce back.
We know this because of the third peak, which is almost as big as the second. If we were just seeing the overall decreasing trend, then the third peak would be much smaller than the second. But instead it's nearly the same size, because it, like the first peak, also gets a boost from the quantity of dark matter: this peak represents pressure waves with even shorter wavelengths that had enough time to collapse, bounce back, then collapse again. The dark matter just collapsed once and sat there, so its contribution remains.
The power spectrum goes on, to much shorter wavelengths, but WMAP doesn't image those shorter wavelengths very well, though other instruments have. And those other instruments show the same pattern: every even-numbered peak is suppressed compared to its odd-numbered neighbors.
It's pretty much impossible to produce this kind of pattern
without a weakly-interacting, low-temperature component of the universe that has substantial mass.