Embolic complications of indwelling arterial lines are well documented. We evaluated a method of blood sampling from indwelling arterial lines that minimizes blood loss and eliminates the embolic risks associated with retrograde flushing. The values for PaO2, PaCO2, and pH obtained by the conventional sampling technique were compared to those obtained by a technique termed the “three drop” method. Thirty-five paired samples were obtained from patients in the pediatric ICU. Meaningful statistical and clinical correlations were observed for PaO2(r = .97,p< 10-6), PaCO2(r = .97,p< 10-6), and pH (r = .98,p< 10-6). Evaluation for the slope of the regression line for each pair of variables was also significant (p< 10-6). The means for each variable were also assessed. Only the PaCO2values were different (t= 2.49.p< .002). As the absolute value was 2 torr, we feel that there is no clinical significance to this finding. These data confirm that the three drop technique of sampling for arterial blood gas analysis is reliable. It also removes the risks of retrograde flushing and minimizes blood loss.