The amount of arsenic in the urine, faeces and in duplicate diets of two couples who had eaten customary Japanese meals were monitored for 7 days by arsine-generator atomic absorption spectrophotometry. For the four volunteers, the mean daily intake of arsenic from their diets was 182 micrograms (range 27 to 376 micrograms). The dietary arsenic was composed of 5.7% inorganic arsenic, 3.6% methylarsonic acid, 27.4% dimethylarsinic acid and 47.9% trimethylarsenic compounds. The mean amounts of arsenic eliminated daily in urine and faeces were 148 micrograms (50-416 micrograms) and 46 micrograms (0-138 micrograms), respectively. The urinary arsenic was composed of 1.4% inorganic arsenic, 3.5% methylarsonic acid, 33.6% dimethylarsinic acid and 61.4% trimethylarsenic compounds. The daily intake of arsenic influenced the total amount of arsenic excreted in the urine (r = 0.7302, P less than 0.01) and the amount eliminated in the faeces (r = 0.5900, P less than 0.01) the next day. Specifically, there was also a significant correlation between the daily intakes of trimethylarsenic compounds and dimethylarsinic acid and the amounts of these compounds found in the urine the following day (r = 0.6833, P less than 0.01 and r = 0.6630, P less than 0.01, respectively). Considering the amounts of arsenic compounds present in seafood and in other components of the diet together with the urinary elimination patterns of arsenic compounds, it seemed probable that the trimethylarsenic compounds in the urine originated largely from fish and shellfish, which contain mainly arsenobetaine. Trimethylarsenic compounds in the urine should therefore be the preferred indicator of arsenic arising from the ingestion of seafood, especially fish and shellfish. In this study, the mean daily intake of inorganic arsenic from the diet (0.18 micrograms/kg) did not exceed the FAO/WHO JECFA Tolerable Daily Intake of 2 micrograms inorganic arsenic kg.