Atomic absorption spectrometry of lead was studied by using a carbon tube flameless atomzer. The variables affecting sensitivity and reproducibility such as the inside dlameter of the tube, the inert gas and its flow rate, and sample inlection volume, as well as interference of diverse ions were investigated.Each carbon tube having the inside diameter ranged from 2.5 to 5. Omm was made of the carbon rod with a length of 70mm and with a diameter of 6.35mm: Smaller bores gave higher sensitivity, but larger bores gave higher reproducibility and lower back grounds. In the present study, the tube with the inside diameter of 4.0-4, 5 mm was used.The effect on the peak absorbance of an inert gas, which protected the carbon tube from atmospheric oxidation at high temperature, was examined. The ratioof the peak absorbance in argon, nitrogen and helium was 1 1 0.9, at the flow rate of each gas of about 0.4l/min. The peak absorbance of lead decreased as argon flow rate increased.A fairly linear calibration curve was obtained in the range of(5-50)10-10g of lead with injection volume of 5μ l and argon flow rate of O.4l/min. The detection limit was 6.5 10-11g and the relative standard deviation, 2.8 per cent. Most cations, at 100 times as much as lead, scarcely interfered in the. determination of Iead, but several cations at 1000 times, seriously interfered. Lead content in sea water was determined by extracting it into diisobutyl ketone as diethyldithiocarbamate, injecting 10μ l of the organic phase and measuring the peak absorbance at thff wave length of 283.3 nm. With this: method, 1.7 8. Oppbof lead were found in the sea waters of Suruga Bay, Shizuoka Pref.