The separation method of inorganic arsenic and methylarsenic in urine or aqueous solution, and the method of determination using an atomic absorption spectrophotometer were examined. As regards the separation of As (III) and As (V) as well as monomethylarsonic acid (MMAA) and dimethylarsine acid (DMAA) and the volatilization procedure of arsine and respective methylarsines, Braman's method was used. Next, atomization of these arsine, monomethylarsine and dimethylarsine was performed by introducing them into a heated quartz cell for the measurement of arsenic using an atomic absorption spectrophotometer. As a result of examining the above-mentioned method, the following improved items were found. 1) For the atomization of arsine, monomethylarsine and dimethylarsine, it was necessary to heat the quartz cell to 950-1, 000°C. 2) As a reducing reagent of arsenic in the test material, 2 ml of 10% NaBH4 was needed. 3) If Ag+, Cu2+, Ni2+ ions existed in the test material solution, the generation of arsine, monomethylarsine and dimethylarsine was suppressed. When the concentrations of these metal ions were not more than 5 ppm, their influence can be neglected. Comparing the result of this method with that of wet ashing method on urine at various arsenic concentrations, correlation between data obtained by these two methods was significantly high (r=0.990, P<0.01). An observation on the excretion patterns of As (III), As (V), MMAA and DMAA in the urine after the subjects having been ingested seafood containing much arsenic was made as time passed. The number of subjects was three, and they restricted their ingestion of seafood for 48 hours before the experiment. An arsenate-rich liquid used in this experiment was prepared by extraction from a kind of seaweed (Hijikia fusiformis). Total arsenic in this liquid was 2.88ppm, which was shared by each arsenic in the amount 7% for As (III), 86% for As (V) and 7% for DMAA. Each subject was orally ingested once an amount of this liquid corresponding to 10μg/kg of arsenic, and was let urinate at fixed time intervals for the measurement of arsenic. First, excretion peak of As (V) was noted within 2 hours after the ingestion. It was thought that this fact meant that As (V) in the ingested arsenic liquid was excreted into the urine, while maintaing its form as it was. In the urine after that time, As (V) always showed lower values than any of As (III), MMAA and DMAA. Excreted As (III) maintained higher values than As (V) from the 2nd hour to the 12th after the ingestion. This fact suggests that As (V) was reduced to produce As (III) in the living body. MMAA gradually increased from the 4th hour to the 12th after the ingestion, and this increase continued until the 24th hour. The excretions of As (III) and MMAA in the urine were reversed after the 10th hour, and it is thought that this fact was caused by the generation of MMAA due to methylation of As (III) in the body. As to DMAA, although about the same quantities of arsenic were contained as As (III) and DMAA in the arsenic liquid, more than twofold larger quantity of DMAA than As (III) was excreted in the urine within two hours after the ingestion. Next, in the urine after the 4th hour, about twofold larger quantity of DMAA than MMAA was excreted at any time in the urine after the 4th hour. Furthermore, the excretion pattern of DMAA ran almost parallel with that of MMAA. The facts abovementioned indicates that As (V) was metabolized to As (III), MMAA and DMAA step by step in the body. At the 48th hour after the ingestion of As (V) liquid, total arsenic concentration in the urine decreased to the order of values before the experiment. The quantity of arsenic excreted during this period corresponded to 36% of the ingested arsenic quantity. And the percentage of each arsenic in the total excretion for 48 hours after the ingestion was 47.4% for DMAA, 25.3% for MMAA, 17.5% for As (III) and 9.7% for As (V).