Abstract: Controlled-pore glass (22.6 nm pore diameter, 80 to 100 mesh) was silylated and treated with CS2 (details given) to form the NN-dialkyldithiocarbamate. The treated glass was then packed into a glass tube (4 cm x 2.5 mm) that formed part of a flow injection system. Optimum conditions were established for the extraction of metal ions and for their elution with HNO3, which was followed by direct nebulizer injection for determination by AAS. The capacity of the collector varied from 0.9 mmol g-1 for Rh(III) to ~4 mmol g-1 for Co2+, Cu2+ and Hg2+. The calibration graphs for Rh(III), Co2+, Cu2+, Hg2+ and Hg22+ were rectilinear up to 0.6, 1, 0.14, 2.5 and 5 µg mL-1, respectively, and the corresponding detection limits were 1, 2, 0.05, 50 and 25 ng mL-1 for 5- or 10 mL samples; the coefficient of variation were 2.2 to 2.7%. Little or no interference occurred with various other ions up to 50 µg mL-1. Flow injection analysis with online pre-concentration using a minicolumn loaded with dialkyldithiocarbamate immobilized on controlled pore glass is described for the determination of Rh(III), Co2+, Cu2+, Hg2+, and Hg22+. The detection limits range from 0.05 ng mL-1 for Cu2+ to 50 ng mL-1 for Hg2+ for 5- or 10 mL samples, improvement of 2-3 orders of magnitude compared with direct injection. The operating conditions are optimized and the effects of interferents are studied. The capacity of the collector varied from 0.9 mmol g-1 for Rh(III) to ~4 mmol g-1 (Co2+, Cu2+, Hg2+).