The best way to measure them is either to use a four-point system (where a known constant current is passed through the resistor, and the voltage drop is measured) or to put a bunch of them in a series string, run a voltage through the string, then measure the voltage drop across each resistor. The latter method is most suitable for matching, rather than getting an absolute value.
you can also use a Wheatstone bridge which you'll find in most high school physics labs --
since EBay is now taken up more by dealers than garage sales it is getting trickier to score a bargain on an HP3468 or HP3478 -- but good 5.5 digit meters from Fluke, Racal Dana, Keithley still abound. RLC bridges tend to be "clunky" but can be very, very acurate and comparatively inexpensive.
I've used a Kelvin Bridge (a Tinsley, but much older model than the one shown) for very low resistance measurement (microhms or less). Ran off a small lead-acid battery, so I guess current was fairly high compared to modern ones.