The two-stage gold amalgamation technique with elemental mercury vapour detection was compared with a technique employing a single gold trap. When peak area was measured, and special attention paid to the gold trap orientation, the one-stage amalgamation procedure provided the same precision, accuracy and detection limits as the benchmark two-stage technique (Fitzgerald and Gill, 1979). The overall analysis time, however, was reduced from about 10 to 2 min per sample. An absolute detection limit (2σ of trap blanks) of less than 1 pg of Hg⁰ was attained using an atomic fluorescence detector, while relative standard deviations of 2.8, 1.5 and 2.2% were obtained for 500, 1000 and 2000 pg of Hg, respectively. Recoveries of close to 100% with RSDs of <3% were obtained in the determination of certified reference materials, intercalibration hair samples and lakewater using this more rapid procedure.