In no other class of cases does "luck" play so large a part in the final disposition of the prisoner. A jury is quite as likely to send an insane man to the electric chair as to acquit a defendant who is fully responsible for his crime. To recapitulate from the writer's experience: (1) The ordinary juror tends to be sceptical as to the good faith of the defence of insanity. (2) When once this distrust is removed by honest evidence on the part of the defence, he usually declines to follow the legal test as laid down by the court on the general theory that any one but an idiot or a maniac has some knowledge of what he is doing and whether it is right or wrong. (3) He applies the strict legal test only in cases of extreme brutality. (4) In all other cases he follows the medical rather than the legal test, but instead of acquitting the accused on account of his medical irresponsibility, merely convicts in a lower degree. The following deductions may also fairly be made from observation: (1) That the present legal test for criminal responsibility is admittedly vague and inadequate, affording great opportunity for divergent expert testimony and a readily availed of excuse for the arbitrary and sentimental actions of juries, to which is largely due the distrust prevailing of the claim of insanity when interposed as a defence to crime. (2) That expert medical testimony in such cases is largely discounted by the layman. (3) That in no class of cases are the verdicts of jurors so apt to be influenced solely by emotion and prejudice, or to be guided less by the law as laid down by the court. (4) That a new definition of criminal responsibility is necessary, based upon present knowledge of mental disease and its causes. (5) Lastly, that, as whatever definition may be adopted will inevitably be difficult of application by an untutored lay jury, our procedure should be so amended that they may be relieved wherever possible of a task sufficiently difficult for even the most experienced and expert alienists.