(after stating the facts as above). The evidence in this case is amply sufficient to sustain a verdict for the government were the crime charged other than treason, and I shall confine myself, therefore, simply to the consideration of whether the rule has been satisfied which is peculiar to that crime; that is, whether any overt act of treason is supported by the testimony of two witnesses.
The words “overt act” go back to the original statute of the 25th Ed. III. Under chapter 3 of 21 Rich. II no overt act was necessary for two years, until it was again reintroduced by chapter 10 of the 1st of Hen. IV, the preamble to which indicates the reason for its restoration : “There was no man which did know how he ought to behave himself to do speak or say for doubt of such pains” (of treason). It is clear that men feared prosecutions pieced together inferentially from chance words or deeds which need not be charged and against which no preparation could be made. Even in the times of Píen. VIII, when the law of treason- was extended to its extremest limit, the provision was retained of some overt act, or at least some written or spoken words which evinced the traitorous design. That design was indeed the substance of the act of compassing, and, in spite of the long history of distortion which compassing endured, the logic of the relation between the overt act and the crime was kept sound. In short, the treason must be manifested by some open deed, whose happening necessarily involved the commission of the crime. This was probably what the original words meant.
*690In v ases of adhering to the enemy the same consistency was- not observed. Strictly no overt act should have been held sufficient which merely manifested a traitorous intent, because the treason lay in hostile acts, for the words “adhering” must be taken as defined by the phrase, “giving aid and comfort.” It was hardly the purpose of the 25th of Ed. Ill to allow the treason of levying war or of adhering to the enemy to be proved merely as attempt; nor is there a sound reason for supposing that the overt act was meant only to disclose a traitorous intent in treasons which involved more than intent. Probably the analogy with compassing confused the subject, but, whatever the cause, it was decided before the end of the seventeenth century (Lord Preston’s Case, 12 How. St. Tr. 646) that aid and comfort need not reach the enemy, and that, though the accused were frustrated in his attempt, it was enough if the overt act declared his intent. This has since remained the law. Rex v. Hensey, 1 Burr. 643, 19 How. St. Tr. 1342; Rex v. Stone, 25 How. St. Tr. 1171; United States v. Greathouse, 4 Sawy. 457, Fed. Cas. No. 15,254.
Nevertheless a question may indeed be raised whether the prosecution may lay as an overt act a step taken in execution of the traitorous design, innocent in itself, and getting its treasonable character only from somé covert and undeclared intent. It is true that in prosecutions for conspiracy under our federal statute it is well settled that any step in performance of the conspiracy is enough, though it is innocent except for its relation to the agreement. I doubt very much whether that rule has any application to the case of treason, where the requirement affected the character of the pleading and proof, rather than accorded a. season of repentance before the crime should be complete. Lqrd Reading in his charge in Casement’s Case uses language which accords with my understanding:
“Overt acts are such acts as manifest a criminal intention and tend towards the accomplishment of the criminal object They are acts by which the purpose is manifested and the means by which it is intended to be fulfilled.”
Therefore I have the gravest doubt of the sufficiency of the first and second overt acts of the first count and of those of the second count, which consist of acts that do not openly manifest any treason. Their traitorous character depends upon a covert design, and as such it is difficult for me to see how they can conform to the requirement. However, the point is not necessary to a decision of the case, because, though properly laid, they are proved in the same way as the third overt act of the first count, which is good as a pleading under any rule.
The critical question, therefore, is whether it is enough to prove one element of the overt act by one witness, -and another by another. It was not until chapter 12 of the 1st, and chapter 11 of the 5th and 6th, of Ed. VI that there had ever been introduced into the law of treason any quantitative rule of proof. Later, chapter 10 of the 1st and 2d of P. & M. raised a question which I need not consider because the law was settled definitely in favor of the quantitative rule by chapter 1 of the 13th of Car. II. In the trial of the Regicides (Kelyng, 9) it was decided that the statute permitted one witness to one overt act and another to another; and twenty years later, in Lord Stafford’s *691Case (T. Raymond, 407, 408), all the judges accepted the same rule. Therefore when in 1695 the same was declared by statute (7th Wm. Ill) it was only a recognition of what had long been the accepted practice.
The requirement of two witnesses probably had its rise from the canon law (Lord Stafford’s Case, T. Raymond, 408) ; and in fact heresy and treason were necessarily regarded as closely analogous^ at a time when the axiom was everywhere accepted, cujus regio, ejus religio. It was the rule of canon law, borrowed from the civil, that two witnesses were ordinarily necessary (Burns, Ecclesiastical Law, vol. 3, p. 304), and whether it had any genuine roots in Roman law it is not necessary to consider. Such a procedural requirement, wherever it comes from, implies a system of trial not rational in its processes at all, one in which the cause was tried by the sacramental nature of the oath itself (Wigmore, § 2032), one where the witnesses were guarantors or sponsors for the parties (Imperatoris Iustiniani Institutiones, Moyle, pp. 632, 633). This idea, common enough in archaic law, was preserved in the later civil and in the canon law. It would be a complete misunderstanding to suppose that when applied to treason it only meant that the prosecution’s witnesses to any overt act should number at least two. In the sense of the rule he is not a witness who testifies only to an isolated and neutral fact, which is relevant because it rationally corroborates the story of a direct witness. When evidence is estimated quantitatively it is the support of the oath that counts, and the witness is no neutral narrator of past truth. He, is “on the side” of him who calls him; he is of “his tail,” as it were, of his array against the opposite array. It is, of course, quite true that in the ultimate test the jury will decide rationally; but I have now to consider only this formal requirement, which had its origin in other and wholly different ideas, and under which a witness is only he who can “back up” the story of the party to whom he belongs. Treason requires two such witnesses to the overt act. It has on this account always been necessary to produce direct, and not enough to produce circumstantial, proof. Rex v. Lowick, 13 How. St. Tr. 267, 305; Burr’s Case, 25 Fed. Cas. 55, 176. And so, taken historically, there seems to me no doubt as to the correctness of the defendant’s position.
There is, however, a further argument which comes near to demonstration. I have spoken of the origin of the second alternative of the 7th of Wm. Ill, which first, so far as I know, was applied in the Regicides Case, 1660. A similar development in the law of heresy, which, as I have shown, is nearly akin, is of much importance. As trials for heresy were conducted under the canon law, it was in theory necessary that the charge should be proved by two witnesses; but very early the craving to secure convictions at any cost led to precisely the same expedient, eventually legalized .in the sixteenth century, as took place in England in respect of treason in the seventeenth. While the canon law required two “contestes,” and while it was laid down distinctly by Escobar that heresy might not be proved by one testis to one declaration of the accused and another to another, nevertheless *692exactly that practice, i. e., contestes in genere, became prevalent and-was finally authoritatively recognized in 1590.1
This development of the rule of contestes, both in heresy and treason, shows that the pressure of necessity substituted as an avowed, though an unorthodox, equivalent, separate oaths to two separate instances (“overt acts”) of the same general character. Yet each of these oaths must directly support one of the several instances ("overt acts”). Moreover, it. was always the rule in treason that the Jury must believe both witnesses. Rex v. Castelmaine, 7 How. St. Tr. 1067, 1111, 1112. It scarcely needs much historical imagination to see that, if such an expedient were chosen as a relief from the stringency of the law as it stood, it could only have been because the existing rule demanded the support by two witnesses of all the facts making up the specific charge (the “overt act”). If it meant no more than that one direct witness must be corroborated in the sense that later and more rational procedure understands that word, no need for the equivalent devised as such relief could ever have existed. It is, of course, much harder to obtain one direct witness to each of two several instances than a corroborating witness to some innocent detail of the story of a single witness. Indeed, na prosecution has much chance of success without some corroboration, and the rule, if so interpreted, means in practice nothing whatever.
The original draft of the Constitution upon which the’ debates took place contained no other requirement than that “no person should be convicted” unless “on the testimony of two witnesses,” following the Statutes of Ed. VI. The debate occurred on August 20, 1787, and so far as was reported in Madison’s Journal (Doc. Hist, of U. S. vol. 3, pages 568-571) contains no allusion to the practice which had in*693terpreted this language exactly as the 7th of Wm. Ill provided. Mr. Dickinson raised the question whether the proposed change should include both alternatives of the 7th of Wm. Ill, i. e., two witnesses to the same overt act, or one to one and another to another. It was decided to add only the first of these, though, as James Wilson observed, “treason may sometimes be practised in such a manner as to render proof extremely difficult, as in a traitorous correspondence with the enemy” — an observation prophetic of the situation here at bar. The Constitution, therefore, though quite unwittingly, restored the rule to what was probably its original form in England, and was certainly the ’original and orthodox form for prosecutions for heresy. Moreover, the convention intended to establish the stricter of the two alternatives permissible under the 7th of Wm. Ill, to whose penalties the delegates had indeed nearly all been liable. Whether they did wisely in declining to observe James Wilson’s warning is not to the point; they meant to take over less than they found, but in the form they found it, and they accomplished what they intended.
The authorities offered in support of the opposite view do not seem to me persuasive. They are two: First, a ruling of Mr. Justice Paterson in United States v. Mitchell, 2 Dall. 348, Fed. Cas. No. 15,788: and, second, the case of Regina v. McCafferty, 10 Cox, Cr. Cas. 603. Justice Paterson’s ruling in U. S. v. Mitchell is not wholly clear. He may have meant that the assemblage at Couch’s Fort, which moved toward Gen. Neville’s, was in itself an overt act of levying war, and, if so, as he says, it was amply proved. He seems to have preferred the interpretation of the.whole act as one; that is, the assemblage at Couch’s Fort and the attack upon Gen. Neville’s house, If the attack upon the house was an essential part of the act, it must be confessed that there were not two witnesses to each substantial part; for the second witness to the acts at Gen. Neville’s said no more than that it ran in his head he had seen the defendant there. There can, however, be little doubt that in fact the assemblage in warlike posture, and the advance to the attack, was a sufficient act of levying war, although hostilities had not yet actually begun, and so much was conceded on all sides in Burr’s Case. If more was intended it can scarcely be held to be an authority for so much.
Regina v. McCafferty, supra, seems to me no authority for the doctrine that one may tack together the testimony of two witnesses to separate parts of the same overt act. The statute there applicable was the 7th of Wm. Ill, and six judgments were written. Of these, one (Keogh, J.,) does not deal with the point, and two (O’Brien, J., and Piget, C. B.) expressly declined to recognize such a doctrine. George J., seems to have proceeded upon that clause of the statute which permits the proof of two overt acts by separate witnesses, relying upon the informer, Corydon, for only the conspiracy at Chester. O’Hagan, J., relied upon the informer for the conspiracy at Chester, and thought that “one or more of the overt acts” had been proved by many witnesses. He seems to me to lend no countenance to the theory that the testimony can be, tacked. The same is not true of the judg*694men.t of Fitzgerald, J., who in one part held that the rising in Dublin was sufficiently proved by several witnesses to the rising, and that the informer’s testimony to the defendant’s connection with it might be proved by him alone. Yet in another part of his opinion he seems to suppose that the fourth and fifth overt acts of coming to Ireland with others and their assemblage there might be treated as separate, and that they were independently proved. His position is therefore not wholly consistent.
Finally it may be urged that the rule in perjury is an analogy which should be followed, since it quickly became -rationalized into requiring only corroboration of a single direct witness. No doubt the rule in-perjury was in its origin also, canonical (Wigmore, § 2040); but it seems not wholly clear to me that .it first got recognition in the common-law courts only a century after the Star Chamber was abolished. It is at least open to argument that already in Star Chamber the rule had suffered some abatement by the growth of a more rationalizing habit, and that its fate oscillated later between growing rational methods and literal obedience to a rule whose origin was not understood, and for which, as so often happens, a false explanation had been devised. At most, I can see no more in the analogy than that in perjury the rule has suffered change, while in treason it has not, and Wigmore, a high authority, clearly supposes that the rule in treason has not become like that in perjury. Section 2042.
I conclude, therefore, that it is necessary to produce two direct witnesses to the whole overt act. It may be possible to piece bits together of the overt act, but, if so, each bit must have the support of two oaths; on that I say nothing. In the' case of none of the overt acts at bar was the necessary evidence produced. The gravamen of the charge depended for direct support on Victorica alone. For the rest, the case rested upon circumstantial evidence, which, while well-nigh conclusive in fact, was not direct as required. There seems to me no question-whatever that without disregarding the whole theory of the Constitution I could not allow a verdict to stand if I received it. I must therefore direct it for the defendant.
Lea, History Spanish Inquisition, vol. II, pp. 562-563, Book VI, c. 5:
“The theory that it required two witnesses to prove a fact was developed into the rule that they must be contestes — that is, two witnesses to the same individual act of heresy — before it could be accepted as proved. It is often found urged in the argument for the defense that the witnesses are singulares and not contestes; but in practice such a defence was usually disregarded, or at' most only led to the unfailing resource of torture. * * *
“Even in the seventeenth century Escobar affirms the rule absolutely; if one witness swears that he heard Pedro say in the market place that God is not a Trinity, and another that he heard him say so in a house, it does not convict him, for neither fact is legally proved. Such a definition, however, threw too many obstacles in the way of the prosecution not to be eluded, and in fact there were classes of cases, such as solicitation in the confessional, in which it was impossible to have more than one witness to each individual act. So, in prosecutions for Judaism, in which evidence frequently covered a long series of years and infinitesimal incidents in daily life, concurrent witnesses to any single one could scarce be had. Vet the claims of the Inquisition to extreme benignity required this to be understood, as Escobar expresses it, while in practice it was disregarded. It was discovered that witnesses could be contestes in genere .when they testified to different acts of heresy and thus make full proof. It is true that Hojas, after citing authorities on both sides, concludes that the rule requiring two concurrent witnesses to a fact must be observed; but one of his authorities asserts that the contrary is the rule in practise; and the Suprema affirmed this July 27, 1590, by ordering that, where formal heresy is concerned, depositions as to different ceremonies and points of faith are to be held as contestes.”