On the Use of a Central Design Point in Unreplicated Split-Plot Experiments