You should read the original publication for complete details. See also this forum thread and elsewhere on the forum where this has been discussed.
Yes, there should be more than two samples to test adequately. You can compare two groups of samples, or multiple groups.
Good luck!