Prinsep and Stephen, JJ.
1. The three appellants ham been convicted by a jury of dacoity. There can be no doubt that the dacoity was committed, but it was not known by whom, until about eight months afterwards. Through information obtained one Fatik was arrested. He was examined as a witness under conditional pardon, and it is on his evidence that the conviction of the appellants almost entirely depends. In laying this, evidence before the jury, the Sessions Judge told them: “If you think that the approver’s story is worthy of credit in itself, you have to consider whether it has been corroborated on material points.” He then described what in his opinion were “the points of corroboration,” and he told the jury that “the above are points on which the Evidence has been corroborated and that corroboration is full and complete, if you believe it; you have to consider these points and decide whether the approver has been corroborated in material points; and if you find that to be so, then you have in his story sufficient evidence to connect all three accused with the crime.”
2. This was not a proper way to place the case before the jury. The Sessions Judge should have told the jury that, although the law permits them to convict on the uncorroborated evidence of ant accomplice, it is not the practice of our Courts, which nave consistently held that it is not safe or proper to convict on such evidence without some corroboration sufficient to connect each of the accused with the offence committed. With this caution the Sessions Judge should have hid before the jury the evidence corroborating the statement of the accomplice. In regard to the nature of the corroborative evidence, it must be confirmatory of some of the leading circumstances of the story of the approver- as against the particular prisoner–Queen v. Kalla Chand Doss (1869) 11 W. R. Cr. 21. Facts which do not show the connection of the prisoner with the commission of the offence with which he is charged are no corroboration in the sense in which the word is used in such cases, although they may tend to show that certain portions of what the accomplice says is true–Queen v. Nawab Jan (1867) 8 W. R. Cr. 19, 25. per Macpherson J. So also it has been held that as a general rule Courts ought not to convict upon an accomplice’s testimony, unless it is confirmed not only as to the offence, bat as to the identity of the individual prisoner as the person or one of the persons, who participated in the offence, and juries ought to be so advised and directed–Palamsam (1863) Weir 535. Lastly, it was held by a majority of a Full Bench of this High Court that it would be an error in summing up, if a Judge, after pointing out the danger of acting upon the uncorroborated evidence of an accomplice were to tell the jury that the evidence of the accomplice was corroborated by a fact, which did not amount to any corroboration at all–Eiahee Buksh (1886) 5. W. R. Cr. 80-88 This was quoted and followed by Macpherson J. in, Queen v. Nawab(sic)n (1867) 8 W. R. Cr. 19. 25.
3. In this case the Sessions Judge properly told the jury that they had to consider whether the story of the approver had been corroborated in material particulars by the other witnesses, and that, unless they were so satisfied that there is such corroboration they cannot convict. We have now to consider what the Sessions Judge laid before the jury as corroborative evidence.
4. The approver stated that he with the three appellants and others, amongst whom was Babuali, used to assemble at the house of Moharali Karikar, where they used to have a singing party, and that one day they assembled there and started in two parties to Gopalpur to commit this daooity. The Sessions Judge has laid before the jury evidence that the singing parties used to take place. That surely was no evidence connected with this dacoity. The Sessions Judge next laid before the jury evidence to show that Babuali spent a good deal of his time at Gopalpur, from which he asks the jury to find that Babuali had an opportunity to gain information that Mangan’s house, where the dacoity was committed, was worth robbing,’ and that from this there is some corroboration of the approver’s story that he got information on which they committed the dacoity. It does not follow that, because Babuaji used to go to Gopalpur, that he acquired this information or that he gave it to the approver, and in no point of view could this be regarded as evidence corroborating the statement of the approver that the appellants were present at this dacoity.
5. The Sessions Judge next commented on the evidence of the boy Fatik, a cowherd, who was on the premises robbed, and professed to recognize one of the appellants, Jamiruddi, and after po(sic)ifiting out that this boy told the police that he could recognize; nobody, he asked the jury to consider whether, notwithstanding this, his evidence was not reliable. We think that the Sessions Judge should rather have told the jury that this evidence could not he safely relied upon. However, this is not material for the purposes of this appeal, on which we cannot deal with the case on the facts.
6. The approver stated that on arriving at Mangan’s house, where they committed the dacoity, they armed themselves with sticks, which they took from an adjoining pan-garden. On this the Sessions Judge lays before them as corroborative evidence the statement of Bhuban Das, who owns a pan-garden and who states that next morning he found some sticks missing. This is not evidence corroborating the statement of the approver, that these appellants committed the dacoity. The approver next stated that after committing the dacoity they left on the Sankurpur math a bati and two thalis, which they had taken from this house. The statement of the boy Fatik who found these articles is laid before the jury as corroborative evidence, though, as in the case of the sticks, it does not in any way affect the presence of the appellants at the dacoity. For the same reason any conversation that Bahadur may have had with the approver after the dacoity is no corroborative evidence against the appellants. That Jadu Nath Pramanick may have met the two parties on the evening just before the dacoity armed with sticks does not necessarily prove anything; and, if they were so armed, it is not easy to understand why they should have taken sticks from the pan-garden just before the dacoity. The Sessions Judge, however, did not lay any stress on this evidence. We may here point out that the statement of a third person, who was told by Jadu Nath of this occurrence, is not admissible as evidence except to corroborate or contradict the evidence of Jadu Nath. The last point enumerated by the Sessions Judge is regarded as very important corroborative evidence. It is the evidence of an Honorary Magistrate, who went with the police and was shown by the approver the different places that they went to. This may be accepted as corroborating the approver’s story of the occurrence, but it certainly does not corroborate his statement, that the appellants were with him. It can be properly regarded only as illustrating the statement made by the approver.
7. Here we would draw attention to the Judgment of Peacock C.J.–Elahee Buksh (1866) 5 W. R. Cr. 80., in which he points out the necessity for guarding against the introduction of the names of persons committing a crime as one of the chief reasons for requiring corroboration of the statement of an approver in order to convict a person accused by him. Before we comment further on the charge of the Sessions Judge on this part of the case, we would notice that he adds to this he vidence the confession of one of the-appellants, Moharali, recorded on two occasions under Section 164 of the Code of Criminal Procedure.
8. The first examination was recorded on 2nd May by an Honorary Magistrate in the presence of the Joint Magistrate. The second was recorded on 20th May by another Magistrate, who had been deputed to hold an inquiry on the spot, apparently to test the approver’s statement by going over the ground with him. Why he should have recorded a further examination of Moharali is not apparent, nor can we learn why Moharali after his statement of the 2nd May should have been remanded to police custody. The reason given for extending his detention in police custody is, moreover, altogether insufficient. We have read the examination of Moharali, and we do not find that it amounts to a confession. He states that he was accidentally present at the dacoity and an unwilling witness to it, and in no way does he admit that he otherwise took part in it. At the trial he retracted this statement, which he declared to be untrue and made under pressure and the instruction of the investigating police officer. Even as it stands, it is, as the Sessions Judge pointed out to the jury, different in many respects from the approver’s story. After having said this, we think that the Sessions Judge was not justified in adding, “if you believe the approver’s story, then you may conclude that Moharali is trying to shield’, himself in his confession and has introduced some falsities into it in, order to show that he himself was unaware that it was to take place and only happened to be there by chance and was forced into it against his will.” This was not leaving this part of the case for the consideration of the jury.
9. In regard to one of the appellants, Hadan Mullah, there is nothing except the evidence of the approver. In respect to another, Jamiruddi, there is only the evidence of the boy Fatik, who said for the first time at the trial that he recognized this man amongst the dacoits. Against Moharali we have only his statements recorded under Section 164 of the Code of Criminal Procedure, which, as already stated, he retracted, stating that they were extorted from him and, were untrue. We think, however, that there has been misdirection by the Sessions Judge in his charge to the jury. He has laid before the jury evidence against the prisoners corroborating the statement of the approver’s evidence which cannot properly be so regarded, inasmuch as it relates to matters not affecting any of them. In so doing he has acted in a manner, which is declared in the cases of Elahee Buksh (1866) 5 W. R. Cr. 80,88. and The Queen v. Kalla Chand Doss (1869) 11 W. R. Cr. 21. to amount to misdirecting the jury.
10. In considering whether we should order a new trial, we think that we should follow the rule laid down by Peacock, C.J. in Elahee Buksh (1): “It would lead to defeat and not promote Justice, if a verdict were set aside and a new trial granted for a defective summing up with reference to the weight of evidence in a case in which the High Court would, upon evidence given in a trial, have affirmed a conviction if, instead of a trial by jury, the trial had been before a Judge and assessors. It appears to us that the question to be considered is not whether, upon a proper summing up of the whole evidence, a jury might possibly have given a different verdict, but whether the legitimate effect of the evidence would require a different verdict. If the Court is of opinion that the evidence would not on any proper view of the case support a conviction, it would be worse than useless to send back a case for a new trial in order that a jury may have the opportunity of convicting on such evidence upon a proper summing up.”
11. Acting on this rule, we think that Jamiruddi should be acquitted. On a proper view of the evidence of the boy Fatik it (sic) not be relied upon and accepted as corroborating the evidence (sic) the approver. Against Hadan Mullah there is no corroborative evidence, and therefore he must be acquitted. The case against Madar Ali is somewhat different, as in his examination before the Magistrate he admitted his presence at the dacoity: still he did not admit participation in it, and his statement, if received at all, must be accepted in its entirety, and there is this difficulty, he retracted it in the subsequent proceedings; and though there is no evidence to show that it was extorted, and we have the Magistrate’s statement that it was voluntarily made, the character of the statement made certainly does not impress us so as to mate us believe that it was freely made. We, moreover, strongly disapprove of the manner in which, as already stated, the Sessions Judge has asked the jury to consider the variation in his statement from that of the approver.
12. Under such circumstances we think that it is not necessary in the ends of justice that he should be retried.
13. We accordingly direct that he be also acquitted and released.