The spirit of the Roman law partook of the same care and caution in the condemnation of those charged with crime.  The maxim was:—­

“Satius est impunitum relinqui facinus nocentis, quam innocentem damnare.”

That there may be no mistake concerning the fact that this commission is bound as a jury by these rules, the same as juries in civil courts, we again quote from Benet:—­

“It is in the province of the court (court-martial) to decide all questions on the admissibility of evidence.  Whether there is any evidence is a question for the court as judges, but whether the evidence is sufficient is a question for the court as jury to determine, and this rule applies to the admissibility of every kind of evidence, written as well as oral.” (Benet, pp. 225, 226.)

These citations may be indefinitely multiplied, for this principle is as true in the law as any physical fact in the exact sciences.  It is not contended, indeed, that any degree of doubt must be of a reasonable nature, so as to overset the moral evidence of guilt.  A mere possibility of innocence will not suffice, for, upon human testimony, no case is free from possible innocence.  Even the more direct evidence of crime may be possibly mistaken.  But the doubt required by the law must be consonant with reason and of such a nature that in analogous circumstances it would affect the action of a reasonable creature concerning his own affairs.  We may make the nature of such a doubt clearer to the court by alluding to a very common rule in the application of the general principle in certain cases, and the rule will readily appeal to the judgment of the court as a remarkable and singularly beautiful example of the inexorable logic with which the law applies its own unfailing reason.

Thus, in case of conspiracy, and some others, where many persons are charged with joint crime, and where the evidence against most of them must, of necessity, be circumstantial, the plea of “reasonable doubt” becomes peculiarly valuable to the separate accused, and the mode in which it is held it can best be applied is the test whether the facts as proved, circumstantial, as supposed, can be made to consist just as reasonably with a theory that is essentially different from the theory of guilt.

If, therefore, in the developments of the whole facts of a conspiracy, all the particular facts against a particular person can be taken apart and shown to support a reasonable theory that excludes the theory of guilt, it cannot be denied that the moral proof of the latter is so shaken as to admit the rule concerning the presumption of innocence.  For surely no man should be made to suffer because certain facts are proved against him, which are consistent with guilt, when it can be shown that they are also, and more reasonably, consistent with innocence.  And, as touching the conspiracy here charged, we suppose there are hundreds of innocent persons, acquaintances of the actual assassin, against whom, on the social rule of noscitur a sociis, mercifully set aside in law, many facts might be elicited that would corroborate a suspicion of participation in his crime; but it would be monstrous that they should suffer from that theory when the same facts are rationally explainable on other theories.