The rate of appearance of unlabelled glucose was calculated from changes in plasma glucose specific radioactivity after a single intravenous injection of labelled glucose and compared with the actual constant infusion rate of unlabelled glucose into an anaesthetized dog with all sources of endogenous glucose production surgically removed. The mean steady-state rate of appearance of unlabelled glucose calculated from the area under the specific radioactivity versus time curve was 7% higher than the actual infusion rate (n = 4), but the difference was not statistically significant. The variability in the rate calculated in this manner was, however, greater than the variability we have reported with rates determined from a primed constant infusion of tracer. Using 15- to 60- or 60- to 120-min specific radioactivity data the mean rate of appearance of glucose, calculated on the assumption of a one-pool model for glucose turnover in vivo, was approximately 60% higher than the actual infusion rate. The results also indicate that it is possible to construct multi-pool models, but it is difficult to equate specific physiological events with the individual terms of the multi-experimental equation which describes the changes in plasma glucose specific radioactivity.