why a weight reading is considered constant when the last two consecutive weights are within 0.0003g?

This sounds like "rule of thumb" that is either based upon the precision capablity of a particular weight measuring scale, or the accuracy requirements of a particular experiment.