STUDY OBJECTIVE: To compare the performance of five pulse oximeters during hypoperfusion, probe motion, and exposure to ambient light interference. DESIGN: Prospective study. SETTING: Laboratory facility at a university medical center. PATIENTS: 8 unanesthetized, ASA physical status I volunteers. INTERVENTIONS: We evaluated five common pulse oximeters with respect to three scenarios: (1) an operating room light was shone on oximeter probes, (2) a motion generator was used to generate 2 Hz and 4 Hz hand motion, and (3) a pneumatic compression device overlying the brachial artery was used to simulate hypoperfusion. Electrocardiographic (ECG) and arterial blood gas values were considered gold standards for heart rate (HR) and oxygen saturation (SpO2) respectively. SpO2 nondisplay and values greater than 4% from simultaneous arterial SaO2-oximeter values were defined as errors. Nondisplay of HR, or HR greater than 5% from ECG values, were also considered errors. MEASUREMENTS AND MAIN RESULTS: The Ohmeda and Nellcor N200 with finger probe had the highest total failure rates with respect to both SpO2 and HR due to ambient light interference (p < 0.05). The Nellcor N200 with finger probe and N200 with C lock were the most accurate with regard to SpO2 during 2 Hz and 4 Hz motion (p < 0.05). However, all oximeters failed dramatically during 4 Hz motion when measuring HR. In the hypoperfusion model, the Nellcor N200 with finger probe and the Nellcor C Lock oximeters performed significantly better than all others in terms of both HR and SpO2 (P < 0.05), while the Criticare oximeter failed 100% of the time. CONCLUSION: There are significant differences in the accuracy of commercially available pulse oximeters during nonideal circumstances, with failure rates varying from approximately 5% to 50% depending on the oximeter and source of interference. Furthermore, no single oximeter performed the best under all conditions.