Wanchoo, C.J.—This is an appeal by Laloo against his conviction under sec. 302 I.P.C. by the Additional Sessions Judge of Banswara. 2. The prosecution case relates to the murder of one Chandu. This man was in the service of Saligram of Khandu. It appears that on the night between the 24th and 25th of December, 1950, Chandu deceased, and Narain, brother of Saligram, were keeping watch over a field of theirs. At about 3 or 4 A.M., Narain came away leaving Chandu still sleeping in the field. The same afternoon at about 2.30 P.M., the body of Chandu was found lying dead in that very field. It appears that he had been injured with a sword, and a number of blows had been given on his neck and other parts of his body. Information was sent to the police, and investigation began soon after. 3. The appellant Laloo was arrested on the evening of the 26th December, 1950. Before that a sword had been found lying where the dead body of Chandu was in the field, while the sheath of the sword was also found at some distance from the place. It is further said that the shirt, which the appellant Laloo was wearing at the time of his arrest, was blood-stained. The alleged motive for the murder was said to be illicit connection between the deceased and the wife of the appellant. It further appears that the appellant remained in police custody from the time of his arrest till the 2nd of January, 1951. He was put in the judicial lock-up on the 3rd of January, 1951, and a report was made for recording his confession. This confession was recorded on the 11th January, 1951, by Shri Jamnalal Laddha P.W. 12 at Kushalgarh. Thereafter, the case was sent to court. The appellant retracted this confession in the Court of Session. In the committing Magistrates Court he was not asked anything about this confession, and therefore said nothing about it, but he pleaded that he had not committed the murder. 4. There is no doubt that Chandu was murdered sometime before 2 P.M. on the 25th December, 1950. The only question, therefore, is whether it was the appellant who was responsible for the murder. In that connection the main evidence consists of the retracted confession of the appellant. 4. There is no doubt that Chandu was murdered sometime before 2 P.M. on the 25th December, 1950. The only question, therefore, is whether it was the appellant who was responsible for the murder. In that connection the main evidence consists of the retracted confession of the appellant. It has been contended on behalf of the State that the uncorroborated confession of an accused, even if it is retracted, is sufficient for his conviction, if the court believes it. As a proposition of law, this may be accepted as correct, but there is weight of authority for the view that the rule of prudence is that where a confession has been retracted by an accused at the earliest opportunity, the court should require some corroboration of it in order to convict the accused on that confession alone. It is sufficient in this connection to refer to two cases of this Court. Mst. Najali vs. Sarkar (A.I.R. 1950 Rajasthan, 5) and Balia and others vs. Sarkar (A.I R. 1950 Rajasthan, 16). The reason why it has been thought prudent by the courts to require some corroboration of a retracted confession is that the court has to satisfy itself that the confession is true. If there is nothing whatsoever outside the confession, it would be generally difficult for the court to come to the conclusion that the confession which is said to be false by the maker himself is really true. We, therefore, propose in this case to see whether the confession, which has been made by the appellant and which he retracted at the earliest opportunity, can be implicitly believed by us. If it is not possible to believe implicitly in the truth of this retracted confession, the appellant would, in our opinion, be entitled to an acquittal. 5. The trial court has come to the conclusion that there is corroboration of this confession, and therefore it should be believed. It is being urged on behalf of the State before us also that there is corroboration available on the record, and therefore we should rely on this retracted confession. We have already pointed out that the confession was made eight days after the appellant had been put in the judicial lock-up. It is being urged on behalf of the State before us also that there is corroboration available on the record, and therefore we should rely on this retracted confession. We have already pointed out that the confession was made eight days after the appellant had been put in the judicial lock-up. It may also be mentioned that the Magistrate took the necessary steps to warn the accused, and the confession therefore is admissible in evidence and there is no defect in the manner in which it came to be recorded. In this confession the appellant said that the deceased was outraging the modesty of his wife and that he used to talk to her when she went out to fetch water or when-ever she went to the fields. The appellant admonished the deceased, but he did not care for the admonition. The appellant also suspected that the deceased intended to abduct the appellants wife. Consequently, on the night in question, the appellant went to the field and found Narain and the deceased lying covered with a Gudri.. The appellant was armed with a sword. Narain later met him and enquired from him if the carts had started moving and he replied in the affirmative. Then he went on where Chandu was sleeping and killed him in the night with the sword. After killing Chandu he threw the sword away and came back to his field. 6. This was the confession made by the appellant on the nth of January, 1951. No attempt was made to prove that there was illicit intimacy between the deceased and the appellants wife, or that the deceased used to tease the appellants wife or talk with her. Thus there is no proof of any motive for the appellant to commit this crime except what we find in his own confession. On the other hand, Saligram has stated that they had not heard of any quarrel between the deceased and the accused before the death of Chandu. No evidence was produced to corroborate this part of the confession. 7. It was possible to corroborate the confession by proving that the sword, which was recovered from near the dead body of Chandu, belonged to the accused, but no proof of that is forthcoming. No evidence was produced to corroborate this part of the confession. 7. It was possible to corroborate the confession by proving that the sword, which was recovered from near the dead body of Chandu, belonged to the accused, but no proof of that is forthcoming. One witness Hakaria was produced to show that the sword, which was recovered from near the dead body of Chandu, belonged to the father of the accused, but this witness did not say so in the Sessions Court, though he had made a statement in the committing Magistrates Court to that effect. Under these circumstances, we cannot rely on the statement of Hakaria, and there is nothing to prove that the sword, which was recovered from near the dead body of Chandu, belonged to the appellant or to any member of his family. It may be mentioned that Hakarias statement was that the police had threatened him and forced him to say that the sword belonged to the father of the accused. 8. Another way in which it was possible to corroborate this confession was to show that the shirt, which was recovered from the house of the accused soon after his arrest and was said to be smeared with blood, had stains of human blood thereon, but no attempt was made to have the shirt chemically examined. The appellant admits that the shirt is his, but that is of no use to the prosecution for unless it is proved that the shirt had stains of human blood, there will be nothing strange in finding the shirt in the appellants own house. 9. Learned counsel for the State has laid particular stress on the fact that the appellant had gone to the field that night and had been seen by Narain with a sword in his possession to show that the confession has been corroborated. Accepting the statement of Narain as correct, namely, that the appellant had gone to his field where he was sleeping with Chandu in that night and asked for fire, we cannot understand how this can be called corroboration of the confession in a material particular connecting the accused with the crime. Accepting the statement of Narain as correct, namely, that the appellant had gone to his field where he was sleeping with Chandu in that night and asked for fire, we cannot understand how this can be called corroboration of the confession in a material particular connecting the accused with the crime. The fact that he was armed with a sword is nothing, for we understand that that part of the country is dangerous and it would be quite natural for a person walking out in the night to be armed with some weapon or other. Narains statement is that the appellant asked him for some fire. There is nothing strange in a villager in one field going to another person in another field in the night and asking for a match stick or match box in order to be able to have a smoke. So the fact that the appellant was seen near the field in which Chandu was later found murdered is, in our opinion, not that kind of corroboration of the confession in material particulars which would lead us to the conclusion that the confession must be true. 10. It is said that the deceased was killed somewhere in the night. The medical evidence shows that death took place about 18 hours before the post mortem examination, i. e., somewhere in the afternoon of the 25th of December, 1950. We have evidence to show that Chandu was dead at about 2 P.M. on the 25th December, but there is no evidence to show how much before 2 P.M. he had died. If reliance can be placed on the medical evidence, the chances are that the man must have been killed during the day time, and not in the night as mentioned in the confession. It has also been urged that the appellant had stated in the confession that he threw away the sword near the deceased, and that a sword was found there. But as the appellants connection with the sword has not been established, we do not place much reliance on the recovery of the sword on a spot near the deceased. 11. It has also been urged that the injuries that have been found on the deceased also corroborate the confession. But as the appellants connection with the sword has not been established, we do not place much reliance on the recovery of the sword on a spot near the deceased. 11. It has also been urged that the injuries that have been found on the deceased also corroborate the confession. We do not, however, consider this as corroboration in material particulars, because it is hardly likely that an accused who has been in custody of the police for about seven days would speak of any other kind of injuries when giving his confession. 12. On an analysis therefore of the evidence that has been produced in this case, we are of opinion that there is no corroboration of this retracted confession on any material particular. Learned counsel for the State relied on Manohar Singh vs. Emperor (A.I.R. 1946 Alld., 15) in this connection and said that in that case the accused was convicted on his retracted confession when there was very slight corroboration of the confession available. We find, however, that in that case the motive for the crime was corroborated by independent evidence. Another case to which reference is made is Raggha vs. Emperor (A.I.R. 1925 Allahabad, 627 (F. B.)). In that case there was evidence to show motive for the murder, and there was also evidence corroborating the recovery of certain clothes which had been traced at the instance of the accused, In the case before us there is no corroboration even as to the motive for the crime. There is no other corroboration connecting the accused in any way with the crime, or with the instruments used for committing the murder. Under these circumstances, in our opinion, it would be unsafe to convict the accused on his retracted confession alone. 13. We, therefore, allow the appeal, set aside the order of the court below and acquit the appellant. He will be released at once if not required in any other connection.