Influence of controlled consumption of marine fish on the urinary excretion of arsenite, arsenate, dimethylarsinic and monomethylarsonic acid (DMA, MMA) was investigated in two experiments. Arsenic species were separated by anion-exchange chromatography and detected with hydride-technique atomic absorption spectrometry (detection limit 1, 10, 2, 2 μg/l). Firstly, 13 probands ate different types of seafood after having refrained from any seafood for 1 week. DMA levels rose from 3.4±1.3 μg/g creatinine ( n=12; a day before seafood) to a mean peak level of 28.2±20.6 μg/g ( n=13; 10–23 h after; P<0.001; max. 77.7 μg/g). No other species were excreted before the meal, but small amounts of arsenite (8.5% positive; max. 1.7 μg/g) and MMA (1.2%; 1.6 μg/g) within 2 days after it ( n=82). Consumption of white herring caused the highest DMA levels. Secondly, eight probands ingested white herring (dose 3.5 g/kg; DMA content 32.1±15.3 ng/g wet weight; n=36). No arsenite, arsenate and MMA was found in the urine or in the herring tissues. The mean DMA mass excreted after the meal (65.3±22.0 μg/24 h) was about 6-fold higher than the sum of base DMA excretion (3.0±1.7 μg/24 h) and the ingested DMA mass (7.9±2.7 μg). This indicates that the elevated DMA excretion after herring consumption is not caused by the metabolism of inorganic arsenic but of other arsenic species present in the fish tissue, e.g. arsenobetaine or fat-soluble arsenic species.