SummaryCarbon dioxide (CO 2 ) is the most commonly used agent for euthanasia of laboratory rodents, used on an estimated tens of millions of laboratory rodents per year worldwide, yet there is a growing body of evidence indicating that exposure to CO 2 causes more than momentary pain and distress in these and other animals. We reviewed the available literature on the use of CO 2 for euthanasia (as well as anaesthesia) and also informally canvassed laboratory animal personnel for their opinions regarding this topic. Our review addresses key issues such as CO 2 flow rate and final concentration, presence of oxygen, and prefilled chambers (the animal is added to the chamber once a predetermined concentration and flow rate have been reached) versus gradual induction (the animal is put into an empty chamber and the gas agent(s) is gradually introduced at a fixed rate). Internationally, animal research standards specify that any procedure that would cause pain or distress in humans should be assumed to do so in non-human animals as well (Public Health Service 1986, US Department of Agriculture 1997, Organization for Economic Cooperation and Development 2000). European Union guidelines, however, specify a certain threshold of pain or distress, such as 'skilled insertion of a hypodermic needle', as the starting point at which regulation of the use of animals in experimental or other scientific procedures begins (Biotechnology Regulatory Atlas n.d.). There is clear evidence in the human literature that CO 2 exposure is painful and distressful, while the non-human literature is equivocal. However, the fact that a number of studies do conclude that CO 2 causes pain and distress in animals indicates a need for careful reconsideration of its use. Finally, this review offers recommendations for alternatives to the use of CO 2 as a euthanasia agent.Keywords Carbon dioxide; euthanasia; pain; distress; anaesthesia; welfare; rodents Carbon dioxide (CO 2 ) is commonly used for euthanasia and anaesthesia of laboratory rodents, largely because of its ease of use, relative safety, and low cost, as well as its capacity to euthanize large numbers of animals in a short time span (Ambrose et al. 2000). In large institutions and those with significant rodent-breeding programmes, large numbers of rodents are often euthanized in a short time (Kline et al. 1963) and an appropriate gas agent is often the best way to approach such a challenge. However, CO 2 is not physiologically inert and the published evidence on whether or not CO 2 administration causes pain or distress in animals raises questions about its routine use. This paper reviews this published evidence and includes information on the effects of CO 2 in both humans and nonhumans. Methodological details are included when possible in order to provide a clear