Recent Improvements in the graphite furnace have greatly reduced analytical interferences. Problems remain that appear to be related to the poroslty of graphite and to the fraglllty of pyrolytlc graphlte protectlve coatings. We compare graphite tubes from several sources which have been coated in several different ways for their differences In analytlcal sensitlvlty for Cu, AI, Sr, TI, and Mo. A separate set of experiments using the determlnatlon of AI In a chloride matrlx shows even greater variation between graphlte materials prepared differently. Experiments uslng the graphlte furnace at steady-state temperature, in the platform and probe variatlons, showed less than 10% signal suppression from 50 pg of CaCI, or CuCI, on 2 ng of AI.Our recent papers (1-6) have attempted to improve the analytical conditions for trace metal determination with the graphite furnace and atomic absorption spectroscopy. When the furnace is operated at a steady-state temperature (2, 3, 5-9) during the time the analyte is vaporized and the absorbance signal is integrated