The selective retention of the Sb(III) chelate with ammonium pyrrolidine dithiocarbamate (APDQ on a column of Chromosorb 102 resin from a buffered sample solution including Sb(V) was used for the determination of Sb(III). The retained antimony was eluted with acetone. The retention of the Sb(III)-iodide compounds with sodium iodide on the Chromosorb 102 resin column from the same solution after reducing Sb(V) to Sb(III) by iodide in acidic solution was used to preconcentrate the total antimony. The retained antimony was eluted with 0.25 mol 1(-1) HNO3. The antimony in the effluent was determined by flame atomic-absorption spectrometry. Also, the total antimony was determined directly by graphite-furnace atomic absorption spectrometry. The Sb(V) concentration could be calculated by the difference. The recoveries, were greater than or equal to95%. The detection limits of a combination of the column procedure and flame AAS for antimony were 6 - 61 mug 1(-1) and comparable to 4 mug 1(-1) for a direct GFAAS measurement. The relative standard deviations were <6%. The procedure was applied to the determination of Sb(III) and Sb(V) in spiked tap water, waste-water samples and a certified copper metal with the satisfactory results.