T maze reversal learning after several different overtraining procedures.