ABSTRACTThose who wish to refute the view that it is worse to kill than to let die sometimes produce examples of cases in which an agent lets someone die but would be generally agreed to be no less reprehensible than if he had killed. It is argued that the examples produced typically possess a feature which makes their use in this context illegitimate, and that when modified to remove this feature, they provide support for the view which they were designed to undermine.