The record in this case discloses several irregularities, which, if they do not vitiate the proceedings, may very properly tend to strengthen any legitimate claims for a new trial.
On the 14th of October, during the last October term of the Superior Court in this county, the defendant was put to plead to an information filed by the Attorney for the State, charging him with the crimes of theft and burglary. The jury were empaneled and sworn to try the issue upon the plea of not guilty, “but upon the reading of the information to the jury, it appeared for the first time to the prisoner and his counsel, that .there was a new and distinct offence charged in the information against the prisoner, to wit, a charge for burglary in breaking into a barn. Whereupon the prisoner moved the coui’t for a postponement of the trial, on the ground that such new and distinct charge, of another and distinct crime, had taken him by surprise, and that he was unprepared in his defence of the same at that time.”
The court granted the motion, but without fixing upon any time for the trial, or making any order with reference thereto; and afterwards, on the same day, adjourned the court till the first Tuesday of November, 1874. On that day the court recommenced its'session; the adjourned term being held by a different judge from the one who had held the regular term. The court was informed by counsel that the accused had pleaded not guilty to the information, and that a jury was empaneled to try the issue on the 14th of October previously, by the judge then holding the court. The counsel for the accused then moved to empanel a new jury, and the Attorney for the State opposed the motion. The court directed the trial to proceed before the jury previously empaneled, and the trial was so had.
As this case was postponed indefinitely by order of the court, after the jury was empaneled, and as the court was *263afterwards adjourned for nearly three weeks, it can hardly be presumed that the judge or the jury supposed that this case was still held under consideration. The mingling of a jury with the bystanders, and with the community generally, after a case has been commenced, is an evil always to be avoided when practicable, as such a course is almost inevitably prejudicial to a fair and impartial trial.
Several objections are taken to the proceedings on the trial, and to the instructions given in the charge. We notice a portion of them.
The accused was not entitled to the instruction asked for, that the corn-house could not be the subject of burglary. A question having been made on that point, involving both the character of the building and its fastening, the court should not have passed upon the question as matter of law, but should have explained the law as applicable to the subject, and then submitted the question to the jury to be decided by them as a matter of fact. State v. McGowan, 20 Conn., 277, 8.
The law as to the evidence of accomplices, generally, must be considered settled, if abundant discussions and frequent decisions, substantially of uniform character, can produce such a result. As long ago as 1787, in the case of The King v. Atwood, 1 Leach Cr. Cases, 464, Judge Buller, speaking by the authority of the twelve judges of England, decided that a conviction, produced alone by the testimony of an accomplice, was legal. Judge Grose, in Jordaine v. Lashbrooke, 7 T. R., 609, said this was not new law, nor founded on a new principle. Lord Ellenborough held the same in Rex v. Jones, 2 Camp., 131. It was held by eleven judges, in the case of Rex v. Sheehan, Jebb’s Crim. Cas., 54, that the testimony of an accomplice, though altogether uncorroborated, was evidence to go to a jury, and that a conviction on such testimony was legal. This court has recognized the same doctrine in State v. Wolcott, 21 Conn., 272, and in State v. Stebbins, 29 Conn., 468.
While the law as to evidence of accomplices is thus broadly laid down, it has become a rule of practice, it cannot correctly be called a rule of law, for the coui’t to caution the jury as to *264the weight of this species of evidence. The jury is usually advised that, unless the accomplice is corroborated, it is safer to acquit; and where there is no corroboration, some English judges have directed an acquittal. It has even been insisted that the corroboration must be directed towards the person of the accused, rather than the circumstances of the transaction. No question involving this, or any similar point, arises in this case.
In the case at bar there were three witnesses examined, who swore that they were accomplices in the crime charged upon the accused, and on the part of- the prosecution it was claimed that they corroborated each other. On the part of the accused it was insisted, and the court was asked to instruct the jury, that where two or more accomplices are produced as witnesses, they are not to be regarded as corroborating each other; that the same rule is to be applied, and the same corroboration required, as if there were but one.
While we are not prepared to sanction to the full extent claimed, the doctrine insisted on by the defendant’s counsel, and though we think the instruction given to the jury by the court below was correct as far as it went, still, the exigencies .of the case demanded more full and more specific instruction as to whether several accomplices could be considered as corroborating each other. We dwell on this point at some length the more readily, because we do not find that it has received much attention in courts of last resort, though it has frequently been decided at nisi prius.
So important has it been deemed that the testimony of .an accomplice must be corroborated to give it weight, that Judge Ellsworth, in giving the opinion of the court in a case already quoted, State v. Stebbins, says: “ To omit it is now held a clear omission of a judicial duty, and becomes a ground, perhaps, for granting a new trial.” 29 Conn., 473.
If it be thus important to caution the jury as to this evidence unless corroborated, it is exceedingly important that it be explicitly understood what is meant by corroboration. There was no omission in this case to charge as to the necessity of corroboration, but as the charge stands, we think the *265jury would have been warranted in supposing that by law accomplices might be regarded as corroborating each other. To that, certainly as applicable to this case, we cannot assent.
The testimony of an accomplice is suspicious because it comes from a polluted source. It is defective in credit. No doubt, under some circumstances, such defect might be diminished by additional testimony of like character, also defective in credit. But such testimony must be independent, not connected together, to be of any avail. Any understanding, connivance or conspiracy among the witnesses, destroys altogether any cumulative or corroborative force which increasing the number of witnesses might otherwise produce. If the accomplices in this case had been confined- in separate jails, and had had no opportunity of being together to prepare a uniform story, neither having heard the testimony of the others till he had given his own, their testimony, if coincident, would certainly have been reciprocally corroborative. But it seems that these witnesses were together, and so similarity of statement counts very differently from what it otherwise might. Merely adding to the number of broken reeds gives no increase of strength. The confirmation required to give the testimony of an accomplice the necessary weight should come from an unimpeached source, not from partners in guilt.
Such are the views of the text writers who discuss this subject, with perhaps a single exception. 1 Phill. Ev. (Cow. Hill & Edw. notes,) 4th Am. ed., 113; 1 Greenleaf Ev., § 381; 2 Taylor on Ev., 798. Chief Baron Joy favors the idea that one accomplice may corroborate another. Joy on the Evidence of Accomplices, 100 et seq.
The doctrine of the decided cases is very uniform.
In Rex v. Noakes, 5 Car. & Payne, 326, Mr. Justice Little-dale in summing up said, (Mr. Baron Bolland and Mr. Justice Alderson sat with him): “It is not usual to convict on the testimony of one accomplice without confirmation, and in my opinion it makes no difference that there are more than one.”
In Magill & others’ ease, Northeast Circuit, Irish Circuit Cases, 1841, 2, 3, p. 418, the prisoners were indicted for larceny. Five witnesses were called for the crown. Mr. O’Ha*266gan, of counsel for the prisoners, claimed that these witnesses were accomplices, and so there was no case to go to the jury. .Perrin, J., said, “Four of the witnesses are accomplices, and if the case rested with them I should direct the jury to acquit the prisoners, but I do not think it clearly appears that the fifth witness was an accomplice, and the case must go to the jury.”
In Regina v. Aylmer, 1 Craw. & Dix Cr. Rep., 116, the prisoners were indicted for stealing four sheep. All the witnesses • were concerned in the commission of the felony charged. Bushe, C. J., remarked—“It will be necessary to sustain the charge by other evidence than the uncorroborated. testimony of accomplices.” Scott, Q. 0., submitted that, though the evidence of each witness was tainted by the fact of his having been an accomplice in the commission of the crime, yet the witnesses were respectively corroborative of each other. Bushe, C. J. “ If in this case there could have been no communication between the approvers, the testimony of one might be brought forward to support the testimony of another. This was done in the case of The Wild Q-oose Lodge, where the approvers had been confined in separate gaols so that there could have been no communication between them. In this case the approvers have had an opportunity of concocting a story among themselves.”
In view of the various questions appearing on this record, we think there should be a new trial, and so advise the Superior Court.
In this opinion the other judges concurred; except Phelps, J., who did not sit.