Least significant differences for combined analysis of experiments with two or three factor treatment designs See original record