The gold concentration in serum, synovial fluid and urine has been determined using a HITACHI MODEL 207 atomic absorption spectrophotometer. Flow rate of acetylene was setted at 2.0l/min. and lamp current was setted at 8mA. The samples were pretreated by the determination of gold. Serum was diluted with deionized water by two to ten-fold. To O.2ml. of synoival fluid 50units/ml of hyaluronidase was added and then incubated for 20 minutes at 37℃. This sample was then treated the procedure for serum given previously. Urine was treated using the modification method of Christions procedure. The recovery rates with gold sodium thiomalate in serum, synovial fluid and urine were 99.9%, 102.5% and 93.6% respectively. When gold sodium thiomalate was injected in rabbits, the maximum serum gold level was attained 1 hour after injection and then gradually decreased at about 50% of the maximum level 24 hours after injection. In the observations of changes in the gold levels in the serum of patients with rheumatoid arthritis after injection of gold sodium thiomalate and gold thioglucose, its peak level after gold thioglucose injection tended to be slightly lower as compared with the same amount of gold sodium thiomalate injection. In those patients who had been given 300 mg. or more of gold salt-preparation, one additional injection of 25 mg. of gold salt could attain the gold level of around 200μg/100ml, in serum a week later, and this level was very nearly the same with that after 3-4 days of one additional injection of 1Omg. of gold salt. The amount of gold excreted in the urine varied from patient to patient, and the mean excretion rates of gold sodium thiomalate and gold thioglucose in the urine within 24 hours after intramuscular administralion were 9.8% and 6.5% respectively. The simple, rapid and accurate method for analysis in biological specimens using atomic absorption spectrophotometer would offer one of the excellent measures in the clinical management of patients with rheumatoid arthritis.