Some anesthetic gases interfere with the determination of blood O2 values. We evaluated, for its potential for such interference, a gas mixture containing (v/v) approximately 70% ethane, 20% sulfur hexafluoride, and 10% cyclopropane, as is currently used in trace amounts to determine ventilation-perfusion (v/Q) ratios. Normal human blood samples were first tonometered with control gas mixtures containing (v/v): (1) 20.9% O2, 0.04% CO2; (2) 16.0% O2, 4.0% CO2; and (3) 9.9% O2, 7.8% CO2. A large quantity (20.9%) of the v/Q mixture was blended experimentally into the control mixture and the tonometry repeated. The entire experiment was then repeated substituting pure N2 for the v/Q mixture as a dilution control. O2 values were determined by three methods: (1) a polarographic electrode, ABL-300 (ABL); (2) a spectrophotometric method, Co-Oximeter (COOX); and (3) a galvanic cell, Lex-O2-Con (LEX). The v/Q gas mixture lowered significantly all measured LEX values by 2.5-3.6 saturation percent (sat%), but showed no effect, dilution or otherwise, on the O2 values determined by the COOX and ABL methods. The N2 dilution lowered the LEX values by an average of only 0.9 sat%; the ABL and the COOX were approximately 0.6 sat% lower. We therefore suggest that, if any of these O2 measurement methods are used in the presence of the v/Q mixture, baseline O2 values should be determined both before and after injection of the mixture into the bloodstream, prior to performing other experimental manipulations. The difference between the two values, if any, can then be used to interpret subsequent results.