The effect on alveolar oxygen fraction (FAO2) of insufflating oxygen under a mask (or through an inflow nipple provided in the mask) during simulated mouth-to-mask ventilation was investigated using a lung model. A variety of commercially produced masks were evaluated. Two patterns of artificial ventilation were applied: 1. 500 ml tidal volume at 20 breaths per minute, and 2. 900 ml tidal volume at 12 breaths per minute. The ventilating gas mixture was oxygen 16% in nitrous oxide, and oxygen was insufflated at flow rates of 2, 4, 6, 8, 10, 12 or 14 litres per minute. The rate of rise of FAO2 and the equilibrium FAO2 attained were greatest at high oxygen inflow rates. The relationship between oxygen flow and FAO2 was not linear however, and an oxygen flow rate of 10 l/min was adequate to generate FAO2‘s around 50% with either ventilatory pattern. The equilibrium FAO2 achieved was greater with smaller tidal volumes and with larger mask deadspace. We also found that several breaths were required for equilibration of FAO2 during each trial, supporting recommendations that several breaths should be given on commencement of artificial ventilation during cardiopulmonary resuscitation.