High concentrations of CO2 have been used for stunning poultry, but have been considered inhumane because the birds appear to find it aversive. However, Ar or a mixture of Ar and CO2 have been suggested to be more humane. The responses of hens to different gas atmospheres were tested during a series of approach-avoidance tests. Twelve hens were trained in the test apparatus, which consisted of a upper and lower chamber, which were connected by a chute. During the actual testing phase, gas would be injected into the lower chamber. The hens did not have access to feed the night before the training session. During the actual training session, hens were placed in the upper chamber and feed was placed in the lower chamber. The feed was used to motivate the hens to enter the lower chamber. Once the birds were trained to enter the lower chamber, the tests began. During each of six test sessions, a different gas treatment was injected into the lower chamber: air, 30% CO2 -in-air, 45% CO2 -in-air, 60% CO2 -in-air, 70% Ar / 30% CO2, and Ar. All hens were randomly exposed to each treatment. Training continued throughout the experiment, and alternated with the testing days. The results indicated that birds took longer to leave the upper chamber during the training session the day after they have been exposed to 45% CO2 -in-air and Ar. However, there were no differences among the treatments in the latency to leave the upper chamber. There were no differences in the number of stops and retreats when the hens were approaching the lower chamber that was enriched with CO2. The lowest number of stops and retreats were found in the control and Ar treatments. Only one hen was not stunned by the CO2-enriched and Ar-enriched atmospheres. There were no differences in the percentages of hens stunned among the stunning gas treatments. Hens in the 30% CO2 -in-air treatment entered the lower chamber at close to the same frequency as in the Ar treatment. When exposed to 60% CO2 -in-air, most of the hens became stunned in the chute on the way to the lower chamber. Hens that entered the lower chamber before being stunned were intermediate for 45% CO2 -in-air and 70% Ar / 30% CO2. The authors concluded that when stunning/killing chickens, there does not seem to be a welfare problem when using mixtures of up to 60% CO2 -in-air compared to 70% Ar / 30% CO2. Using Ar appears to be slightly more welfare-friendly, but not enough to warrant its use over the other options.