(after stating the facts as above). The chief question is the charge of the judge on the issue of subornation. Whether it was a matter on which he was called upon to charge at all we have not to decide. The bankrupts corroborated each other in fact as to thfe subornation, and, if the judge had said nothing on the subject, the point might have arisen whether the doctrine requires more than that corroboration is necessary if the case is to go to the jury at all. In fact, he went further and charged them that no corroboration was necessary as to the subornation. If this is wrong in law, we cannot treat it as harmless. While it is extremely improbable that the jury should in fact have believed one bankrupt and not the other, the charge was equivalent to saying that they might wholly ignore one. This we think would have been an improper instruction, even though the judge need have said nothing on the subject. Thus it seems to us that the point was introduced into the case and must be decided.
It is, of course, well settled that the perjury must be proved by two witnesses or by one with corroboration, U. S. v. Wood, 14 Pet. 430, 10 L. Ed. 527; and this is also true in cases of subornation, Hammer v. U. S., 271 U. S. 620, 46 S. Ct. 603, 70 L. Ed. 1118. It is a doctrine alien to the common law, imported from ecclesiastical, and eventually from civil, law procedure, like the same rule in treason. U. S. v. Robinson (D. C.) 259 F. 685. Until its abolition in 1640, Star Chamber had in practice had exclusive jurisdiction over perjury, and its procedure was taken from the ecclesiastical courts. Wigmore, § 2040. There were practical reasons why the doctrine should be taken over, along with the jurisdiction, which were later rationalized into the notion that one oath would do no more than balance the other. Perhaps, too, the fact that perjury was an offense cognizable by spiritual courts may have contributed. Nevertheless, Russell (volume 1, p. 478) says that it never was the practice to extend it beyond the assignments of perjury, and this was recently decided in Atkinson v. State, 133 Ark. 341, 202 S. W. 709. We can see no reason not to accept the gloss, made in ignorance of its historical origin, as a proper limitation in application. The doctrine itself has indeed a rational basis when applied to mere recantations, though it must be owned that, if extended to the oath of another than the perjured wdtness, it is hard to justify in a court of common law.
The point now at bar was specifically reserved in Hammer v. U. S., but, so far as we can find, has been uniformly ruled in favor of the prosecution when it has arisen. Com. v. Douglass, 5 Metc. (Mass.) 241; State v. Waddle, 100 Iowa, 57, 69 N. W. 279; Boren v. U. S., 144 F. 801 (C. C. A. 9); U. S. v. Thompson, 31 F. 331 (C. C.). In People v. Evans, 40 N. Y. 1, to be sure, some of the language does look the other way; but the *715assignments o£ perjury were themselves supported by a single uncorroborated witness only, and it was unnecessary to add that the subornation was subject to the same rule.
In respect of the second error, it is to be observed that the judge considered the application for a new trial, and denied it in his discretion. We might stop for that reason, Mattox v. U. S., 146 U. S. 140, 13 S. Ct. 50, 36 L. Ed. 917; Harrison v. U. S., 7 F.(2d) 259 (C. C. A. 2); but it is plain anyway that his discretion was well exercised. The defendant learned of the publication in season, and had his opportunity to stop the trial and start afresh, or to find out whether in fact it had hurt him. He chose not to inquire, but to take his chances on the result; now, though he has lost, he must abide his choice. Spreckels v. Brown, 212 U. S. 208, 215, 29 S. Ct. 256, 53 L. Ed. 476; Marrin v. U. S., 167 F. 951 (C. C. A. 3). Nor is there in fact the slightest reason to suppose that he was mistaken in thinking that he had not suffered.
With the contradictions of the bankrupts we have nothing to do. They came before the jury in any case with the major contradiction of abjuring their own oaths; that once accepted, the jury were not likely to stop at minor divergences, especially as it is apparent that they were ignorant men ill-versed in the language, and as it is probable that on cross-examination they were confused between their direct testimony given at the trial and that at the bankruptcy examination. At any rate the jury alone was to say at which time the truth appeared; it is enough that they could see the bankrupts and weigh their words for what they were then worth.
The errors in the admission of evidence require no notice. The matters brought out were either relevant enough to the corpus delicti to be admissible, or innocuous to the defendant.
Judgment’ affirmed.