Abuse liability testing of opioid drugs was originally motivated by attempts to separate the analgesic effects of opioids from their likelihood for abuse. It has become apparent that the human population group likely to abuse opioids has little overlap with the population group requiring opioids to treat pain, therefore there is no longer a need to separate these two properties of opioids. This is fortunate, since, as reviewed here by Jim Woods and colleagues, the results of the plethora of studies that have attempted to distinguish these two properties in known opioids strongly indicate that they are inseparable. Evaluation of the abuse potential of novel opioids remains, however, critically important in deciding on governmental restrictions on their accessibility. In addition, opioid abuse liability testing contributes enormously to our understanding of the behavioral mechanism of action of these drugs, and in surprising and helpful ways has increased our appreciation of the various test systems used to garner information about them.