Currently accepted methods for serum triglycerides, namely, the manual modified VAN HANDEL-KAWADE method and the semi-automated LOFLAND method, were critically examined. Since the absorbents, the quality of sulfuric acid, the processes of shaking, extraction, saponification, removal of ethanol after saponification, and the stability of colored solution appear to be essential factors governing precision and accuracy of the determination, the effects of these factors were studied. It was found that the process of removal of ethanol was most important. To obtain the reproducible absorption, it is most advantageous to evaporate off ethanol from the solution of pH⋍6 on an oil bath at 101℃. As a result of re-examination on LOFLAND'S semi-automated method using a standard AutoAnalyzer,
it is shown that the mixing coils can be omitted,
thereby making the analysis possible in 18 -19 minutes (In the original LOFLAND'S method it takes more than 22 minutes). In addition, this modification results in smaller amount of chromotropoc acid reagent necessary (2.9 ml/min. as compared to 5 ml/min. in the original method). The determination rate in this modified method is 25 samples per hour. The average values, standard deviations, and coefficients of variance for the same serum sample for ten analyses according to the suggested manual method were 55.0 mg/dl, 1.83 mg/dl, and 4.69%, respectively, and those according to the suggested semi-automated method were 53.6 mg/dl, 1.78mg/dl and 4.68%, respectively.