1. In this case the verdict of the jury dearly cannot stand and the conviction must be quashed. I think it is my duty to order a retrial for dacoity, if the prosecution think there is evidence of dacoity, or for any act of robbery or violence included in the original charge of dacoity which they think they can prove to the satisfaction of the jury.
2. The ground upon which I quash the conviction is that there was something more than a mere misdirection. There is a finding which, as it stands, convicts the accused of an offence wholly unknown to law. The charge was a charge against nine persons for having taken part in a dacoity themselves in addition to a tenth person, who was the approver. It may be that there were other persons who took part in this dacoity and that evidence was laid before the jury that there were other persons. But it is quite clear that the attention of the jury was not drawn to such evidence, if it existed, by the Judge in his summing up He began his charge thus-' This is a case of dacoity.' Now a dacoity involves the co-operation of five persons, but he did not explain this. He continues: 'The prosecution story is that the nine accused and the approver dacoited the shop of one Budhai.' He then considers the evidence of the approver, and the evidence in detail against each of the accused in a very 'careful and clear summing up, which I have no doubt the jury carefully considered because they showed discrimination in their verdict. But the verdict ultimately returned by the jury was that seven of the accused were not guilty, but only two. Now the accused who have been convicted might be guilty of dacoity with the approver and any other persons amounting in number to five or more, although unknown. The other persons might be the other accused, even although the jury were not satisfied that the other accused were the persons, and therefore so far as the jury were concerned they were persons unknown. And if the learned Judge had slopped for a moment to consider the possibility of the jury reducing the number of guilty persons to less than five and had told the jury that any one or more of the accused might be found guilty of dacoity if the jury were satisfied that they had taken part with five or more persons, there could have been no ground for complaint. Similarly if he had gone on to say that if the jury were not satisfied that five persons ever took part at all, none the less they could convict any of the accused of any of the other offences which go to make the offence of dacoity, no complaint could have been made about it. The result is that this Court has before it a conviction of two persons and two persons alone for dacoity upon a finding by the jury that the only other persons except the approver who were suggested to have taken part in this dacoity did not take such part. This is not a mere irregularity; it is a matter of substance. You cannot take part in a dacoity with one person; the number must be as least five. Therefore, unless the jury were satisfied that there were at least five persons in the dacoity and found so by their verdict, they could not lawfully convict two men of dacoity. This is in accordance with the principle laid down in England in The King v. Plummer  2 K.B. 339, where a conviction was quashed, although the accused had pleaded guilty.
3. It is also complained, and the Government pleader concedes it, that so far as the heads of the charge to the jury in this case are concerned, they do not show that the legal definition was ever explained to the jury in the sense of making them understand what a dacoity is. If the Judge did so, he has at any rate not made it clear in the heads of charges that ha did so. There is a decision of this Court Emperor v. Baij Nath Weekly Notes 1903 p. 232, where it was held that under Section 377 of the Code of Criminal Procedure, a Judge is called upon to show by the heads of his charge to the jury in what way the law on the subject was explained to the jury. I agree with that ruling. To hold otherwise might be to render nugatory an appeal to this Court on a question of misdirection, because the Judge might otherwise merely say that he explained the law on the subject, which means no more than what he thought was the correct statement of the law. I may further point out that the authority to which I have just referred lays down that the heads of a charge to the jury ought to show that the evidence was properly laid before the jury. Before I take leave of the charge to the jury in this case, I would merely observe that it is not as a general rule a sufficient compliance with Section 367 for a Judge to say in the heads of the charge to the jury: 'I do not propose to discuss the defence. It has hardly been touched in argument.' The pleaders are not infallible and sometimes they do not make the best of their clients' case. A jury, who are not trained in the ways of a court of law, might not appreciate the full value of that statement. I think, whatever his own opinion of it may be and whatever he may think of the method of the prisoner's counsel in dealing with his own evidence, the Judge should at any rate state to the jury accurately the substance of the evidence relied upon by each accused. It is for the jury and the jury alone to decide whether that evidence is worthy of consideration,
4. It has been urged before me that I might, if the evidence justified, alter the conviction into one which would be according to law, if found in fact. Undoubtedly an appellate court in the case of a trial by jury may look at the evidence to see whether the misdirection has caused a miscarriage of justice in this sense, namely, whether, in spite of the misdirection, the conviction and the verdict are not justified in law as they stand, and if they are then the court of appeal will refuse to interfere. I have already pointed out that I cannot do that here, because in my view the conviction cannot stand. It may also look at the evidence to see whether there is a case worthy to be sent down for a retrial. But in my view it is impossible for an appellate court to take the further step, namely, to look at the evidence and find, as a fact, the accused guilty of any offence with which they have not been specifically charged in the trial court and which was not laid before the jury. In my view the whole essence of a trial by a jury is that, if an accused has a right to a jury, he is entitled as of right to call upon the prosecution to obtain a verdict against him on the faces laid before the jury before he can be convicted at all. It is not for the appellate court to look at the evidence with a view to see whether another jury might not have arrived at a verdict of guilty upon another charge. That is to usurp the function of the jury and to substitute the appellate court's opinion for the verdict of the jury. On a reference to the High Court under Section 307 different considerations arise, but in an appeal the right of an accused to a verdict of the jury alone is in fact jealously guarded and preserved by the Code, Section 418 of the Code of Criminal Procedure provides that an appeal may be on a matter of fact as well as on a matter of law except where the trial was by jury, in which case an appeal will lie on a matter of law only. Section 423 of the Code gives wide powers to an appellate court as to the alternative courses open to it in an appeal from a conviction, and they include altering a finding and the sentence, which of course means arriving at a new finding for itself upon facts, Without anything to justify it in the section, to hold that that power is given to a court of appeal in the case of a trial by a jury would be inconsistent with the, provisions of Section 418. Section 423 must be read with Section 418, and where facts are in issue the absolute finality of the verdict of a jury on a question of fact must be given effect to. But the section itself provides this, I think, in sufficient, even if not in clear language: 'Nothing herein contained shall authorize a court to alter or reverse the verdict unless the verdict is erroneous owing to misdirection by the Judge.' I think that provision is in favour of the accused, that is to say, an appellate court may alter or reverse a verdict in the sense of preventing a conviction taking effect as the result of a misdirection. But it clearly cannot substitute its own verdict for that of the jury. It seems that the view I take in this matter has been followed for many years, at any rate in Calcutta, as illustrated in Wafadar Khan v. Queen Empress (1894) I.L.R. 21 Calc. 955, and I am told in other provinces as well. I quash the verdict and the conviction and direct the case to be sent back to the Sessions Judge to be retried with a jury for such offence as the prosecution see fit to charge within the scope of the evidence given at the first trial. The question of the personnel of the court I leave to the Sessions Judge.