See also Angus et al. [334], who have performed the same analysis with the opposite conclusion.

McGaugh:
My own opinion is that it is virtually impossible to use satellite motions to test MOND. One really needs to know the orbits; especially since there can be hysteresis in the case that MOND is a modification of inertia. The real issue for statistical tests (that average over individual orbits) is identifying and correcting for contamination by interlopers (galaxies that are not true satellites). I have no particular objection to the interloper correction made by Klypin & Prada, but neither am I convinced that it is unique. If you look at the raw input data, σ(r) is flat or slowly rising before interloper correction. Which says to me you can drive any model you want through the data by a judicious choice of interloper correction. Even if we happen to get this correction right, how would we be able to tell?

More generally, this is a good case study of how hard it is to remain objective about this issue.

Return to Literature page