Mercury is not going to be attracted to a magnet. Magnetic colloids (ferro fluids), as taylaron mentioned, are likely your best bet. Since the particles are nanoscale, they probably wouldn't settle or separate too much due to Brownian motion and such. Then again, I've never worked with the stuff.
I'm also curious - wouldn't each particle act as a dipole by itself? And how would that relate to the affect of a magnet? I think I recall seeing "spikey shapes" emerge from the fluid when exposed to a magnet.