In the method for lithium (Li) analysis, flame emission photometry and atomic absorption spectrophotometry (AAS) have been used most frequently. In addition, lithium can be analyzed by ion-selective electrode (ISE) or fluorscence polarization immunoassay. We evaluated the comparison between AAS method based on the principle of absorption of light at 670.8 nm by Li and ISE method based on the principle of voltage difference generated by Li in contact with lithium ionophore. We compared with those obtained by AAS (AA/AE Spectrophotometer 551, Instrumentation Laboratory Co.) and ISE(CSYNCHRON EL-ISE, Beckman Co.) in the serum and urine of 6 patients and evaluated time-related changes of serum lithium concentration after dosing in both methods. The results are summarized as follows : 1. In within-run precision study for lithium concentration, coefficient variations (CVs, %) ranged from 1.34 to 2.17 for AAS and from 0.34 to 0.85 for ISE method. In between-run precision study for lithium concentration, CVs ranged from 1.23 to 1.72 for AAS and from 0.61 to 1.38 for ISE method. 2. The correlation study between AAS and ISE method resulted in Y=0.946X+0.137 (N=32, r=0.933, X=AAS, Y=ISE) for serum lithium and Y=1.092X+0.977 (N=28, r=0.943, X=AAS, Y=ISE) for urine lithium. 3. Time-related changes of serum lithium concentration in both AAS and ISE method resulted in peak serum levels about 2 hours after dosing and then rapidly decreased after the peak serum level and finally arrived at nearly initial levels about 9 hours after dosing. 4. The reference range of serum lithium was found as undetectable level for both AAS and ISE method and the reference range of urine lithium to the urine creatinine was 0-0.00014 mmol/mg(mean 0.00002 mmol/mg) for AAS method.