Arsenic speciation was determined in the solution extracted from a ferrihydrite binding gel layer in a DGT unit (FB-DGT) deployed in water using coupled high-performance liquid chromatography and hydride generation atomic fluorescence spectrometry (HPLC-HG-AFS). However, the extent of change in arsenic speciation during the extraction process is unknown. By identifying the arsenic species in the FB-DGT gel directly, using X-ray absorption near edge spectroscopy (XANES) fitting with model arsenic compounds, we obtain a better understanding of the ability of FB-DGT to measure labile arsenic species in solution. The results presented herein confirm that FB-DGT accumulated labile inorganic and methylated arsenic species. Arsenic species bound to the FB-DGT gel were stable for up to 2 weeks following deployment. However, caution should be applied when interpreting the proportion of As species measured by HPLC-HG-AFS in solution extracted from FB-DGT because the distribution of arsenic species in extracted solutions was found to be modified by the extraction process. Some (~20 %) of arsenate was converted into arsenite, and a significant amount (~25 %) of dimethylarsinic acid (DMAsV) was converted into monomethylarsonic acid (MMAsV) or arsenate (AsV). Only inorganic arsenite (iAsIII) was stable during the extraction process. These findings suggested that, based on the XANES measurement, although As species were quantitatively accumulated in the FB-DGT gel, the FB-DGT technique was still unsuitable for studying As speciation. This is because changes in arsenic speciation were observed to occur during gel extraction, prior to HPLC-HG-AFS measurement, and because the XANES technique is available for routine measurements.