An arsenic speciation survey was carried out in water samples from the Tinto and Odiel Rivers (southwest of Spain), as well as their common estuary. Both rivers are affected by acid mine drainage (AMD) and represent an input of heavy metals into the estuary, which also suffers from industrial water discharges. Samples were taken in December 2000 and July 2001. The arsenic species considered were arsenite (As(III)), arsenate (As(V)), monomethylarsonic (MMA) and dimethylarsinic (DMA) ions using coupled high-performance liquid chromatography-hydride generation-inductively coupled plasma-mass spectrometry (HPLC-HG-ICP-MS) for their determination. Parameters such as pH, salinity, redox potential and dissolved O2 were also measured. The results revealed that the acid mine drainage originating mainly during winter along the upper part of the Tinto River course causes high inorganic concentrations of dissolved arsenic, up to 600 microg l(-1) of As(III) and 200 microg l(-1) of As(V). In summer, As(III) levels decreased due to the diminution of the input from acid mine drainage and also because of oxidation, with a corresponding increase of As(V) level. Furthermore, the extreme acidic conditions of this river (pH 2.3-2-6) do not allow biological activity sufficient to produce significant concentrations of methylated arsenic species. The arsenic concentrations in the nearby Odiel River were always 5-10 times lower than in the Tinto River, with arsenic levels usually below 100 microg l(-1), dominated by As(V), indicating that it is less affected by acid mine drainage. The highest inorganic arsenic species concentrations were found where the river crosses a mining site, which corresponds to the highest As(III) values. Significant biological activity in this river produced methylated species that were detected along the water-course, with the highest concentrations at the lower course of the river, accounting for up to 53-61% of the total dissolved arsenic. At the common estuary formed by both rivers, only arsenate was detected in most samples at lower concentrations than in the riverine water samples. The tidal cycle showed a similar pattern of dilution of the arsenate when seawater comes into the estuary. Methylated species were not found either in summer or winter, at least at the 0.1 microg l(-1) level, possibly because of the high turbidity of the waters, producing an inhibition of the phytoplankton activity. In addition to the riverine inputs into the common estuary, industrial activity also represents an important source of arsenic as the discharge from a Cu smelter produced the highest arsenate level of all samples in estuary and also the only sample with significant arsenite concentration. Furthermore, the underlying iron-oxide-rich sediments represent an importance source of arsenic into the water column. In three nearby estuaries not affected by industrial activity or acid mine drainage, arsenic levels remained below detection limits.