The distinctions that may be drawn between self-defence, necessity and duress are interesting as a matter of theory, but may also be important in practice. In some jurisdictions, for example, duress and necessity are no defence to murder whereas self-defence is a defence available in principle to all crimes. In such jurisdictions, in homicide cases, the point at which one reaches the boundaries of self-defence and enters upon the terrain of necessity may thus be of crucial significance. Drawing on Suzanne Uniacke’s theory of self-defence, I would like to suggest that each defence can be distinguished by a different key issue. In necessity cases, the key issue is the moral imperative to act: what matters is whether in the circumstances it was morally imperative to act, even if this might involve the commission of wrongdoing, in order to negate or avoid some other evil. In duress cases, the key issue is the personal sacrifice D is being asked to make: should D be expected to make the personal sacrifice involved in refusing to give in to a coercive threat, rather than avoid implementation of the coercive threat by doing wrong? In self-defence cases, the key issue is D’s legal permission to act: where V unjustly represented a threat to D (normally, although not exclusively, through his—V’s—conduct), the question is whether necessary and proportionate steps were taken by D to negate or avoid the threat. For, D has a legal permission to take necessary and proportionate steps to negate or avoid an unjust threat, even if (exceptionally) these involve the use of lethal force. So baldly stated, the differences between the defences may seem obvious. Few common law jurisdictions, and few commentators, however, have appreciated the full significance of the differences, as we shall shortly see.