Broomfield, J.
1. These are appeals by eight persons who were tried by the Additional Sessions Judge of Thana and a jury for offences of dacoity under Section 395, and in some cases under Section .398 of the Indian Penal Code. The jury found a unanimous verdict of guilty against all the accused and the Judge, although it is clear from his order that he did not agree with the verdict, did not think it necessary to refer the case to this Court under Section 307 of the Criminal Procedure Code, but accepted the verdict and imposed various sentences of imprisonment.
2. The offences are alleged to have been committed on the night of September 4, 1941. The prosecution story is that one Rambhilaksingh, who was afterwards made an approver, and accused Nos. 5 to 8 were at a place called Wangaon where they live and were joined by accused Nos. 1 to 4 who came from Bombay. They proceeded first of all to a village called Tanashi, which they reached about 8-30 p.m., and broke into the house of one Waman and carried off the sum of Rs. 12, after unsuccessfully attempting to extort more property. They then proceeded to another village Khetkhadi which they reached about midnight. There they broke into the house of one Kashinath and were trying to break open a box belonging to him, but as the people of the village collected they all went off without any booty. It is part of the prosecution case that six of the dacoits, i.e. accused Nos. 1 to 6, after returning to Wangaon, proceeded from there to the village of Sakhara, which is three miles from Wangaon on the side away from Tanashi, and spent the rest of the night in the house of one Karim. Waman informed the police patel on September 5 and the matter was reported to the police.
3. On September 9, Ramabhilaksing was arrested in connection with another dacoity, and in consequence of statements made by him accused Nos. 6, 7 and 8 were arrested at once and accused No. 5 on the 15th. On that day the police heard about the visit of six persons to Karim's house at Sakhara and the Sub-Inspector went and made inquiries there. The police seem to have searched Karim's house and also the houses of accused Nos. 5 and 6. An axe was found in the house of No. 6. Karim was taken to a Magistrate to have his statement recorded on the 21st and apparently in consequence of information given by him the police took him with them to Bombay and in a house pointed out by him they found accused Nos. 1, 2 and 4. Accused Nos. 1 and 4 were arrested at once, No. 2 a little later. Accused No. 3 was arrested on October 3. On October 11 identification parades were held at which some of the accused were identified by witnesses from Tanashi, Khetkhadi and Sakhara. Details of these identifications will be given in a moment There was another identification parade on October 14 when accused No. 3 is said to have been identified by Karim. On October 23 the approver was sent to a Magistrate and his confession was recorded on the 27th. In due course the police committed the eight accused and the approver for trial, but the approver was tendered a pardon and gave evidence for the prosecution.
4. The backbone of the prosecution story is the evidence of Ramabhilaksing, and the first point to be noted is that he is not a satisfactory witness even as approvers go. The learned Judge has not given us any indication as to his demeanour or the manner in which his testimony was given, but there are obvious defects in it. He has contradicted himself on several important matters. It appears that he did not inform the police that accused No. 2 was one of the dacoits at all, although in his evidence he is quite definite that he was. He told the police that there were only eight dacoits altogether including himself, whereas in his evidence he says that there were nine. He told the police that immediately after the commission of the dacoities accused Nos. 1 to 4 went by train from Wangaon Station to Bombay where they belong. He has not said anything about that in his evidence in Court and professes to remember nothing about it, and it is fairly obvious that the reason is that his statement that these accused went to Bombay that night is quite inconsistent with the prosecution story. A number of witnesses have been called to show that these four accused and Nos. 5 and 6 were seen that night in the village of Sakhara. Not only does the approver contradict himself but he is contradicted by other witnesses for the prosecution, for instance as regards the weapons that were used by the dacoits. He has mentioned a pistol and an axe as well as sticks, but he says nothing about any gun or any spear, though several witnesses have mentioned a gun and a spear as having been carried by the culprits. Obviously therefore Ramabhilaksingh is a sort of witness whose evidence requires very substantial corroboration before it would be safe to rely upon it.
5. The only corroborative evidence practically is the identification of the accused by some of the witnesses. Waman, whose house was broken into at Tanashi, in his evidence in Court identified Nos. 1, 3 and 5 and he said that No. 3 had a pistol. But at the identification parade he was only able to identify No. 1. Parashram is Waman's son. He says that Nos. 2 and 5 were among the dacoits and No. 2 was the one with a pistol. But at the parade he was only able to identify accused No. 5. Chimi, Waman's wife, deposes that Nos. 1 and 5 came into the part of the house where she was and No. 6 came into another room. She does not explain how she was able to see him if he was in another room. At the parade she identified No. 6 and no others and it seems that in the committal inquiry No. 6 was the only one she said anything about. Javlya is a servant of Waman. According to him there were ten to fifteen dacoits outside the hedge and Nos. 1, 5, 6 and 7 were among those who came inside the house. At the parade, however, he identified only No. 6 and four wrong persons altogether. Kashinath, whose house was broken into at Khetkhadi, identified accused No. 3 as a man who fired a pistol at him. He appears to have picked out No. 3 at the parade but his evidence shows that he is by no means certain that he was the man.
6. Then there are the Sakhara witnesses. Ladkya, the police patel, says he saw six persons in Karim's house and Karim told him they were guests from Bombay. From his evidence it appears that they were accused Nos. 1 to 3, 5 and 6 and the approver. He does not mention No. 4. But his evidence before the Committing Magistrate was, in the Tanashi case, that the men he saw were Nos. 1 and 4, 5 to 8 and the approver, and in the Khetkhadi case that they were Nos. 1 and 2, 5 to 8 and the approver. So that he has given three different versions as to the identity of the men. Harji, a neighbour of Karim, deposes that he saw Nos. 5 and 6, whom he knew before, along with four or five other persons, whom he did not know, at Karim's. Laxman who also comes from Sakhara says that he was washing his face early in the morning when he saw Nos. 1, 3, 7 and 8 as well as Nos. 5 and 6, whom he knew before, passing his house on the way to Wangaon Station. At the parade he identified Nos. 5, 6, 7 and 8. As he knew' Nos. 5 and 6 before, that means nothing as regards them. His identification of accused Nos. 7 and 8 seems to be worthless, because it is not the case of the prosecution or anybody else that these two accused went to Sakhara that night.
7. There is one other witness from Khetkhadi who might perhaps be mentioned. That is Narayan, the brother-in-law of Kashinath. He claims to recognise as having been among the dacoits accused No. 3 and the approver. But at the parade he identified only the approver and curiously enough it seems that this witness Narayan is the only one of all the witnesses who was able to identify the approver.
8. It is stated by the approver and by several other witnesses that the dacoits, all or practically all, had their faces muffled. True it was only the lower part of their faces which was covered over, but even so it must have been an extremely difficult matter to observe their faces sufficiently to be able to recognise them afterwards. In view of this circumstance and in view of the way in which the witnesses have contradicted themselves as to the identity of the accused, it is quite clean that this identification evidence is worth very little and indeed almost nothing at all. It is obviously insufficient evidence to be regarded as corroborating the evidence of an approver like Ramabhilaksingh.
9. The prosecution has relied on the evidence of Karim, but he is by no means a respectable witness, and we feel considerable doubt as to what his position is in this case. He claims to have known accused Nos. 1 to 6 before, and for that reason he was not asked to identify them, although in his evidence he says that he did identify them at the parade. He was apparently asked to identify accused No. 3 at a parade held on October 14, but what the panch witness Achut says about this is that Karim first of all identified a doctor, somebody who was merely in the line assisting the police, then afterwards he pointed to accused No. 3 and said that he had 'grave suspicions about him.' We cannot consider that Karim is at all a reliable witness, and in any case the presence of these accused at Sakhara would only have an indirect bearing on the case and could not by itself prove that they were concerned in these particular dacoities.
10. It appears therefore that in the case of accused Nos. 2 and 4 there is nothing but the evidence of the approver. In the case of accused Nos. 7 and 8 the evidence of the approver is corroborated merely by Laxman's identification which in the circumstances stated is worthless. In the case of the others there is, besides the approver's evidence, that of one or two witnesses who claim to recognise those accused as having been among the dacoits. But that evidence is so dubious that it cannot possibly be regarded as substantial corroboration.
11. As the trial was by a jury the appeal lies only on a matter of law. But in our view the conviction of the accused in this case is in a real sense illegal. It has been laid down over and over again by all the High Courts that it is a rule of prudence and practice which practically amounts to a. rule of law that the evidence of an accomplice ought not to be acted upon unless it is corroborated as against the particular accused in material respects. It is not necessary to refer to many cases. It will be sufficient to mention one recent decision, Emperor v. Allisab, : (1932)34BOMLR1453 where this Court acting upon this rule set aside a conviction by a Sessions Judge, sitting with assessors, which was based on the uncorroborated evidence of accomplices. The fact that this trial was by a jury need make no difference if, as we hold, the. verdict was contrary to this binding rule of practice. There is no reason why a jury should be permitted to ignore a rule which is obligatory on all Courts.
12. It is not really necessary that we should find that the learned Judge has been guilty of a mis-direction or non-direction. We should in any case have been able to interfere even if the jury had brought in their verdict after 'being properly directed. But in the present case we are not satisfied that the learned Judge's charge on this point was correct and satisfactory. He pointed out Section 133 of the Indian Evidence Act, which says that a conviction is not illegal merely because it proceeds upon the uncorroborated testimony of an accomplice, and also pointed out Section 114 of the Indian Evidence Act, one of the illustrations to which says that the Court may presume that an accomplice is unworthy of credit unless he is corroborated in material particulars. He then went on to say :
As a matter of caution, we do not generally place implicit faith in the evidence of an accomplice and we always require some corroboration of his evidence.
He told the jury that they should bear in mind that
Sometimes it might be dangerous to convict only on the evidence of the approver.
He has dealt with this matter incidentally in dealing with the evidence against particular accused thus. He says further on in the charge :
It is for you to see whether you should rely wholly on the testimony of the approver and convict accused No. 2 because he was found in the room of accused No. 1.
13. And again dealing with accused No. 4 :
It is for you tat see whether you can safely act on the testimony of the approver alone and convict accused No. 4.
14. The rule of practice, however, is much stronger and more definite than that, and in our opinion, the learned Judge should have impressed upon the jury much more clearly and emphatically that it is a rule which has practically the force of law and that they ought not to convict on the evidence of an accomplice without substantial and reliable corroboration of it. We think that if the learned Judge had charged the jury in this way, it is probable that they would not have convicted these accused. We also think that if in spite of such a warning the jury had found the accused guilty, the learned Judge should have referred the case to this Court under Section 307. He seems to have thought that in view of Section 133 of the Indian Evidence Act the jury were entitled to find as they did and the verdict was not perverse and accordingly he was not called upon to make a reference. A reference to the High Court, however, under Section 307 ought to be made if the Judge is clearly of opinion that it is necessary for the ends of justice to submit the case. If a jury convicts in defiance of a well settled rule of practice and the Judge feels, as he evidently did feel in this case, that the conviction was wrong, it follows that it is necessary for the ends of justice that the case should be submitted. The learned Judge's hands were not tied in the manner he seemed to think they were.
15. However, as the breach of the rule in question is rather a matter of law than of fact, it is open to us to interfere in appeal even without a reference under Section 307. We have gone through the evidence in the case and we have no hesitation in holding that the verdict of the jury ought not to be maintained as against any of the accused. We must therefore allow the appeals, set aside the convictions and sentences and direct that all the accused be acquitted and discharged.