Arsenic (As) mobility and transport in the environment are strongly influenced by arsenic's associations with solid phases in soil and sediment. We have tested a sequential extraction procedure intended to differentiate the following pools of solid phase arsenic: loosely and strongly adsorbed As; As coprecipitated with metal oxides or amorphous monosulfides; As coprecipitated with crystalline iron (oxyhydr)oxides; As oxides; As coprecipitated with pyrite; and As sulfides. Additions of As-bearing phases to wetland and riverbed sediment subsamples were quantitatively recovered by the following extractants of the sequential extraction procedure: As adsorbed on goethite, 1 M NaH2PO4; arsenic trioxide (As2O3), 10 M HF; arsenopyrite (FeAsS), 16 N HNO3; amorphous As sulfide, 1 N HCI, 50 mM Ti-citrate-EDTA, and 16 N HNO3; and orpiment (As2S3), hot concentrated HNO3/H2O2. Wet sediment subsamples from both highly contaminated wetland peat and less As-rich sandy riverbed sediment were used to test the extraction procedure for intra-method reproducibility. The proportional distribution of As among extractant pools was consistent for subsamples of the wetland and for subsamples of the riverbed sediments. In addition, intermethod variability between the sequential extraction procedure and a single-step hot concentrated HNO3/H2O2 acid digestion was investigated. The sum of the As recovered in the different extractant pools was not significantly different than results for the acid digestion.