We compared three commercially available kits for use in determining triglycerides (triacylglycerols) in serum, by different methods, for accuracy and precision, with use of pooled sera of low, normal, and high triglyceride concentrations. Accuracy varied from -17% to +16% of the independently determined value, depending on the method and triglyceride concentration. In general, the Calbiochem method was the most nearly accurate, and only it met Center for Disease Control criteria for precision at all three triglyceride concentrations. The relationships of mean values to the suggested normal range for each method varied, some significant differences being found. Our data demonstrate the importance of a correlation between the independently determined value and the normal range for the method used and stresses the need for normal-range values to accompany all patients' reports.