A method of inducing euthanasia by carbon dioxide (CO2) inhalation in the home cage of an animal is described and tested for distress by behavioural as well as by hormonal measures. The animals were maintained in their home cage while CO2 was induced at a flow of 6 l/min. The behaviour of the animals was measured continuously as were the serum concentrations of glucose, ACTH and corticosterone 30, 75 and 120 s after the CO2 was introduced into the cage. In order to test for distress, two groups of rats were pre-treated with acepromazine (orally) and pentobarbiturate (i.p. injection) respectively, in order to reduce possible distress caused by CO2 euthanasia, and were compared with control groups. There were no signs of distress by behavioural or by hormonal changes. All changes seen could be attributed to experimental effects and, especially as there was no difference between the pre-treated and the control groups of rats, it must be assumed that the described method of euthanasia is in concordance with animal welfare, it leads to rapid death without severe distress or pain, and it seems therefore to be 'humane'.