Craig v R [1933] HCA 41; (1933) 49 CLR 429 (29 August 1933)

HIGH COURT OF AUSTRALIA

Craig Applicant; and The King Respondent.

H C of A

On appeal from the Court of Criminal Appeal of New South Wales.

29 August 1933

Rich, Starke, Dixon, Evatt and McTiernan JJ.

Curtis K.C. (with him Woodward), for the applicant.

Manning K.C., A-G. for N.S.W. (with him McKean K.C.), for the respondent.

Curtis K.C., in reply.

The following written judgments were delivered:—

Aug. 29

Rich and Dixon JJ.

This is an application for special leave to appeal from an order of the Supreme Court, sitting as the Court of Criminal Appeal, confirming a conviction of the appellant for the murder, on 14th December 1932, of a young girl named Bessie O'Connor. The evidence called at the trial clearly established that the murdered girl had been driven in a Blue Essex Sedan motor car from Redfern to the National Park by a man who had that evening stolen the car, and that she had there been murdered by repeated blows on the head, and had been stripped naked. The issue at the trial was whether the prisoner was the man. The proof of his identity depended upon the evidence of witnesses who had seen the driver at two places at which the car stopped on its way to the National Park and of witnesses who had seen him at a place where he stopped on the return journey after the murder. The only evidence implicating him in addition to the evidence of identification consisted in a statement which one of the witnesses attributed to the driver of the car. When, on the return journey, the driver of the car stopped at Tom Ugly's Point he said, according to this witness, that he was on his way to Liverpool, that his name was Stone and that he lived there in Station Street. There was evidence that the prisoner knew a family named Stone and that he believed they lived in Station Street. Perhaps it should be added that after his arrest the prisoner was escorted to the scene of the crime when, according to the testimony of a detective, he became very agitated and said "Don't take me there."

A new trial was sought upon the ground that fresh evidence had been discovered tending to show that some other person and not the prisoner was the murderer. This evidence was laid before the Full Court. It appeared that a man named Crothers, who was in custody on 12th January 1933 upon a criminal charge then made a statement to the police that at about 7.30 on the night of the murder, when driving a car in Granville, he picked up a man whom he took to Darlinghurst where he left him promising to meet him at the same place at midnight and drive him back to Granville: that he waited for the man at the arranged place from midnight to 2 a.m. when he came carrying a parcel: that he drove the man to Granville where he dropped him: that he found the man had left the parcel in the car and that, when he unrolled it, he found it contained a pair of trousers saturated with blood and a motor tyre lever, and that he had put these in a place of concealment at the motor garage where he worked and that under the advice of fellow prisoners he resolved to report it to the police as throwing light on the murder. He described the man and said that, although he did not know his name, he had met him in Parramatta Gaol in August or September 1931. Crothers was taken by the police to the motor garage where he failed to find the tyre lever but produced a pair of trousers, which, however, were not blood stained. This incident took place five days after the arrest of Craig which was made on 7th January 1933. On 14th February, shortly after Craig was committed for trial, Crothers, who had in the meantime been released from gaol, again communicated with the police. He said that Craig had not committed the murder and he could tell more than he had at first done. He then made a more elaborate statement, describing his movements in detail, according to which, on the return journey to Granville his passenger had said that he had been out to Sutherland with a girl and had left her there. Crothers said that when he had got to the garage he examined the things left in the car and found a pair of trousers, a blood stained lady's handkerchief wrapped round the tyre lever which was also blood stained; the handkerchief had the letters B.O. embroidered in one corner. He said that he had taken the tyre lever and handkerchief and hidden them under the house of an acquaintance at Merrylands. He further said that two days before making the statement he had met the man at Merrylands: that he did not know his name but gave a description of him and of the previous occasions when he had met him. On 15th February Crothers called at the office of the Inspector of Police who said he would go with him to the address at Merrylands to get the articles which had been placed under the house. Crothers then said "No, they are not there; I have shifted them," but he refused to say to what place he had moved them. He promised to bring them to the Detective Office the following day, a promise which he did not fulfil. On the 8th June the prisoner was found guilty of murder. On the 15th June Craig, who was in the Long Bay Prison, requested an interview with his solicitor. Confined in the gaol with Craig was a prisoner who had been closely associated with Crothers. He says that he sent to Craig the name and address of Crothers as a man who could give him valuable information. Craig communicated it to his solicitor and as a result Crothers was sought out. On the evening of 19th June the solicitor's clerk and Crothers made a search at the house under which Crothers says he had placed the handkerchief and tyre lever, but neither of these things was discovered. Crothers swears that after the solicitor's clerk had left he told the occupier of the house the object of the visit. The latter thereupon called in his daughter who made a statement, to which she afterwards deposed, to the effect that three or four months before, when she was clearing out some rubbish, from under the back of the house, she noticed among the rubbish what she believed to be a lady's handkerchief, that it looked very stained and dirty and that she did not pay much attention to it, or remember what became of it, but she thought it likely that it went on to the rubbish heap. In consequence of this information, the solicitor's clerk and Crothers a few days afterwards made a further search, and a handkerchief was discovered amongst some rubbish. It was produced in Court bearing some stains which reacted to a blood test. In one corner the initials B.O. are worked. This is a brief summary of the fresh evidence relied upon.

By sec. 6 of the Criminal Appeal Act 1912 N.S.W. the Court of Criminal Appeal is required to allow an appeal against conviction if it is of opinion that the verdict of the jury should be set aside on the ground that it is unreasonable or cannot be supported having regard to the evidence ... or that on any ground whatsoever there was a miscarriage of justice. Sec. 8 provides that the Court may order a new trial in such manner as it thinks fit if the Court considers that a miscarriage of justice has occurred and that having regard to all the circumstances such miscarriage of justice can be more adequately remedied by an order for a new trial than by any other which the Court is empowered to make. Sec. 12 authorizes the Court, if it thinks it necessary or expedient in the interests of justice, to have witnesses examined before it. There is no power conferred expressly to grant a new trial on the ground of the discovery of fresh evidence. The authority to do so is contained in the power to set aside a conviction when a miscarriage of justice has occurred and to order a new trial when the miscarriage can best be so remedied. It is evident that the exercise of a power to direct a new trial because fresh evidence is forthcoming must be attended both with danger and with difficulty. It is the function of the jury to determine questions of fact in a criminal trial. When they have found a verdict they have performed that duty. If after a verdict of guilty the mere fact that a prisoner produced further relevant evidence required the Court to vacate the conviction and submit the question of the prisoner's guilt to another jury, then in a jurisdiction where perjury is rife great abuses would ensue. A Court of Criminal Appeal has thrown upon it some responsibility of examining the probative value of the fresh evidence. It cannot be said that a miscarriage has occurred unless the fresh evidence has cogency and plausibility as well as relevancy. The fresh evidence must, we think, be of such a character that, if considered in combination with the evidence already given upon the trial the result ought in the minds of reasonable men to be affected. Such evidence should be calculated at least to remove the certainty of the prisoner's guilt which the former evidence produced. But in judging of the weight of the fresh testimony the probative force and the nature of the evidence already adduced at the trial must be a matter of great importance. In the present case counsel for the prisoner very properly as a preliminary to discussing the effect of the new evidence directed our attention to the criticisms to which the evidence of the prisoner's identification was open. Before the Supreme Court it was not denied that the verdict could be supported if the matter depended only upon the evidence given at the trial. Before this Court, however, the learned counsel submitted that whatever view was taken of the fresh evidence, the evidence of identification was so unsatisfactory that a verdict based upon it should not be permitted to stand. It is convenient to discuss this submission before stating our opinion in reference to the new evidence.

It is unnecessary to set out in detail what the witnesses who identified the prisoner stated. It is enough to say that none of them had any prior knowledge of the prisoner, that each of them saw the driver of the car for a short time only on the night of the murder, and that of six persons who undoubtedly saw him four swore at the trial that he was the prisoner and two were not able to do so. Of the four who, at the trial, identified the prisoner as the driver of the car, one was unable to pick him out from a row of men among whom he was placed either as being the man or as being like him. Another was unable to pick him out as being the man but appears to have said he was like him, and a third picked him out only after saying that he was the type of man so far as he could recollect and the man was more tanned than this man was and wanted a shave. The fourth did pick him out without qualification. The evidence of each of them was criticized on other grounds which were discussed during the hearing before us but which we do not find it necessary to set out in detail. Evidence was also given by a witness, called for the defence, who swore that not far from the murdered girl's dwelling, at or about the time it might be supposed the car with the girl in it would leave Sydney, a sedan car stopped in the street, a girl resembling the murdered girl left it and went in the direction of the murdered girl's dwelling and after an interval returned, that he had a conversation with the man who drove the car and that he was not the prisoner. About this time the murdered girl did in fact return to her dwelling after having been out during the earlier part of the evening. This witness knew Bessie O'Connor and he swore that when he heard of the murder he concluded that the girl was Bessie O'Connor. His credibility was attacked in cross-examination, but no grounds for the attack appeared. This briefly states the character of the evidence which coupled with that relating to the allusion of the driver of the car to the name of Stone and Station Street, Liverpool, and, perhaps, the evidence of the detective as to the demeanour of the prisoner at the scene of the murder, constitutes the evidence implicating the prisoner. In our opinion it was within the province of the jury to estimate its weight and force and under a proper direction, which they in fact received from the trial Judge, they were entitled, if it carried conviction to their minds to found a verdict of guilty upon it.

We turn to the examination of the course taken by the Supreme Court in relation to the application for a new trial in respect of the fresh evidence. In doing so we agree that we should not be unmindful of the nature of the evidence implicating the accused, but we think that the learned Judges of the Supreme Court, who were all familiar with that evidence, cannot have overlooked its character. In estimating the value of the new evidence, the first and the dominant consideration is the complete lack of personal credibility in the witness Crothers. It is unnecessary to elaborate reasons for the conclusion that no jury should place any reliance upon him as a witness. The next consideration is that as to many important circumstances of his story he has given three entirely contradictory versions. Except for the physical production of a handkerchief bearing stains presumptively of blood and embroidered with the letters B.O. these facts coupled with the inherent absurdity of the story would disentitle it to any serious consideration. The only manner in which his story can be presented as rationally deserving the consideration of a jury is as that of a man who is attempting to tell enough of facts within his knowledge to exculpate Craig whilst protecting with clumsy and inconsistent falsehood some other person whose guilt he knows or suspects. The story of the handkerchief, even with the support—a slender support—of the evidence of the girl that something she took to be a handkerchief was seen some months before reeks with suspicion. We do not think it necessary to enlarge upon the details of the story of the bundle, its varying components,—trousers, tyre lever, handkerchief, the contradictory accounts of the disposal of these articles, the intervals of time between the accounts given and first search, and between that search and the finding, or the many minor improbabilities of the narrative, nor need we refer to the account of Crothers' career which shows him accustomed to the fabrication of sensations and given to seek notoriety. It is enough to say that the learned Judges of the Supreme Court investigated the matter, had the witnesses cross-examined before them, and after considering what could be urged in corroboration of some of the circumstances surrounding the narrative arrived at the conclusion that the fresh evidence ought not to be submitted for the consideration of another jury.

Our jurisdiction is to give special leave to appeal in criminal matters with respect to which the Court, in the exercise of its discretion, thinks an appeal fitting. A criminal case must present some special features before our intervention is justified. The outstanding fact in the course adopted by the Supreme Court is their complete condemnation of the fresh evidence. We think there is no ground upon which we, who have not seen the witnesses, could, having regard to the character of the story, estimate its probability at any higher value than their Honors did. At that estimate there is, in our opinion, no departure from principle involved in saying that, even when weighed with the criticisms offered upon the evidence implicating the prisoner, the evidence now produced is not of such a character as to require a new trial in order to avoid a miscarriage of justice.

The case, in our opinion, presents none of the features which would make it fit for the grant of special leave to appeal. We think the application should be refused.

Starke J.

Motion on behalf of Eric Roland Craig for special leave to appeal to this Court from the judgment of the Court of Criminal Appeal of New South Wales dismissing his appeal against his conviction on indictment for murder.

In my opinion this Court ought not to interfere with the course of criminal justice unless it is shown that exceptional and special circumstances exist, and that substantial and grave injustice has been done.

The ground founding the present motion is the weakness of the evidence connecting the prisoner with the murder, coupled with fresh evidence laid before the Court of Criminal Appeal. The evidence was given partly on affidavit, and partly on cross-examination before the Court itself. It was anxiously considered by the Court, which reached the conclusion that the evidence was untrustworthy. The greatest weight must be attached to the finding of a tribunal which has seen and heard the witnesses: it has had an opportunity of observing the demeanour of these witnesses and judging their veracity and accuracy such as no other tribunal can have. It would be dangerous, indeed it would be subversive of the whole administration of criminal justice, if this Court intervened and gave leave to appeal from a decision reached in such circumstances.

Evatt and McTiernan JJ.

The present applicant, Eric Roland Craig, is under sentence of death, having been convicted by a jury of the murder of a girl named Bessie O'Connor. Two juries to whom he had been given in charge upon the same indictment failed to reach an agreement. In his report, the learned Supreme Court Judge who presided at the third trial stated that he anticipated a similar result—disagreement.

From his conviction the applicant appealed to the Full Court of the Supreme Court, sitting as the Court of Criminal Appeal under the statute of 1912. The Court was specially constituted to consist of the three Judges who had presided at the three trials. This course was taken, we are given to understand, at the request of the applicant's counsel, because he intended to concentrate upon obtaining the grant of a new trial owing to the discovery of fresh evidence, the general effect of which was to implicate in the murder a person other than the applicant. We are not aware of any precedent for such a course, and sec. 11 of the Criminal Appeal Act seems to suggest that the Judge of the Court of trial should not sit as a member of the Court considering appeals from convictions before him. In the present case the attention of the learned Judges was directed to the authenticity and value of the new evidence, and diverted from what is the outstanding feature of this case—the absence of any evidence implicating the applicant in the crime, other than the evidence, such as it was, directed to the question of identity.

The applicant now seeks to invoke the jurisdiction of this Court to hear appeals from the judgments of the Supreme Court. That jurisdiction is undoubted. The Commonwealth Parliament has imposed one condition, and one condition only, upon its exercise in criminal cases, viz., the grant of special leave (Judiciary Act, sec. 35 (1) (b)). That condition itself involves the remission to this Court of the preliminary question whether any given case presents features of sufficient gravity to warrant a review of the decision of the Supreme Court. Whilst the decision of such preliminary question necessarily carries with it an exercise of discretion, the discretion cannot and should not be limited in advance by the adoption of any general rule tending either to place an embargo upon, or to facilitate, appeals in criminal cases. Each case should be left to depend upon its own peculiar circumstances.

If the preliminary question is answered in favour of granting special leave, this Court is entitled, indeed it is bound, to pronounce the judgment or order which the Supreme Court, acting as the Court of Criminal Appeal, should itself have pronounced. In other words, the High Court becomes, pro tanto, the Court of Criminal Appeal. When, therefore, it is asserted that this Court "is not a Court of Criminal Appeal," the assertion is only true in the sense that there is no appeal as of right to this Court in criminal matters. If special leave is granted, the functions of the High Court are assimilated precisely to those of the Supreme Court, i.e., the Court of Criminal Appeal.

As will appear from our opinion, the present case is sufficiently special in character to require the grant of special leave.

The murderer of Bessie O'Connor was, undoubtedly, the person who, on the night of December 14th last, drove the Essex sedan car to Sutherland, with her as a passenger. Later in the night the murderer drove the car back to a lane near Centennial Park, Sydney, distant about a quarter of a mile from the garage from which, early in the evening, it had been unlawfully taken. The only real question in issue was whether the Crown produced proof, sufficient to exclude all reasonable doubt, that the driver of the car was Craig.

The learned trial Judge said to the jury, when directing them on the question of the identification of the accused: "There are no outside circumstances which would assist you one way or the other as pointing towards or away from the accused." With this view we agree.

We consider that no inference of guilt could safely be drawn from the evidence of the witness Brown to the effect that the driver of the car said that he was acquainted with a certain family and a certain street in Liverpool, and gave his name as Stone. There is only one other piece of "outside" evidence (to use his Honor's phrase) upon which the Crown sought to rely at the trial. Craig was arrested on January 6th, and subjected to an interrogation at police headquarters for many hours throughout the night and early morning, the detectives conducting the examination in relays. On the next morning he was taken with a large number of police officers to the scene of the crime at Sutherland, being told, "We are going to where the girl Bessie O'Connor was murdered." He was also told where the pool of blood had been and the position where the body was found. The evidence relied on as tending towards proof of Craig's guilty knowledge is that, as the police car was pulling up, he was "very agitated," and exclaimed "Don't take me there!"

The police method employed was stated by a witness to be "usual." It is one thing to seek corroboration of the voluntary confession of a person under arrest by taking him to the scene of the crime so that he may be asked to point out in detail what he has already admitted. But it is a very different thing to attempt to extract an admission or confession when, as in this case, the person under arrest has not made any admission of guilt. The latter method may easily lead to abuse.

In the present case, we think that Craig's exclamation "Don't take me there!" might well have been excluded from the consideration of the jury. It is a statement which is quite consistent with the natural revulsion of an innocent person against his being compelled to visit the scene of a murder. This evidence was apparently relied upon by the prosecution, the learned trial Judge saying to the jury: "It is suggested by the Crown that it is some indication of guilt on his part in some way."

The main evidence implicating Craig was that given by certain persons to the effect that he was driving the Essex sedan car on the night of December 14th.

In criminal cases, where the only real issue is the identity of the accused with the person who was performing some apparently innocent act at a time long before the trial, all the surrounding circumstances have to be carefully considered, for we at once enter what has been described as that branch of proof "so notoriously delicate as proof of identity." A short consideration will demonstrate the truth of this description.

An honest witness who says "The prisoner is the man who drove the car," whilst appearing to affirm a simple, clear and impressive proposition, is really asserting: (1) that he observed the driver, (2) that the observation became impressed upon his mind, (3) that he still retains the original impression, (4) that such impression has not been affected, altered or replaced, by published portraits of the prisoner, and (5) that the resemblance between the original impression and the prisoner is sufficient to base a judgment, not of resemblance, but of identity. It therefore became necessary, in the present case, to pay attention to the following circumstances:—(1) Whether the witness was a stranger to the driver of the car, (2) whether the driver had any special peculiarities which, at the time, impressed themselves upon the witness, (3) the length of time which elapsed between December 14th and (a) the time when the witness first described the driver or (b) the time when the witness saw the accused person, (4) the description of the driver given by the witness before seeing the prisoner, and (5) the circumstances under which the prisoner was first seen and identified by the witness as the driver.

We think that these considerations are well illustrated by the evidence of the witness Harvey. Harvey was a garage proprietor, who gave evidence at the trial that the Essex sedan stopped at his garage at Brighton-le-Sands, on the evening of December 14th. According to him, the driver wanted water put into his radiator, and drove off as soon as Harvey had performed this service. Harvey said at the trial that the driver was Craig. But the most extraordinary part of his evidence is his description of the girl, who did not leave the car. He said

There was a girl sitting in the front seat of the car, on the left side. She was a girl with a full face. She had rather bright eyes. She rather struck me as being a happy sort of girl, rather wide mouth. She had a long mouth, I would say. She gave me the impression that she was rather happy. She had that look. I should say she was about 18 years of age. I have the impression that she was wearing some beads around her neck. I could not say what colour they were.

It seems reasonably clear that in giving this detailed description Harvey is unconsciously relying upon the photograph of Bessie O'Connor, which was published in the newspapers as early as December 17th, or upon some other description of her, rather than upon his real recollection of the girl who was in the car. He seems to have had no occasion or reason for specially noting the features or characteristics of the girl. Although he saw newspaper photographs of the murdered girl, he refrained from volunteering information about the girl in the car. On December 23rd, an offer of a substantial reward was published. On January 6th Craig was arrested. It was not until ten days later, on January 16th, that Harvey placed himself in contact with the police as a prospective witness, although, during the month which had elapsed since the murder, officers of police had, from time to time, visited his garage. The coronial inquiry did not take place until January 29th. Harvey was not called as a witness. According to him "the police sent him home." It was not till March 24th, more than three months after the visit of the car to his garage on December 14th, and only five days before the commencement of the first trial, that Harvey was taken to the gaol to see if he could identify Craig. The latter's photograph had frequently appeared in the press since his arrest. There was a "line-up" which included Craig; and Harvey, pointing to Craig, said "This is the type of man so far as I can recollect ... the man was more tanned then, and he wanted a shave. ..." After "communing with himself," as he called it, Harvey asked Craig to speak, and his final opinion that Craig was the driver was partly based upon the supposed resemblance of a voice as to which no peculiar quality is suggested, and which Harvey had heard only once in his life, for a very short time, and long before.

We are of opinion that it would be most unreasonable, indeed highly dangerous, to allow a conviction to proceed upon the basis of such evidence of identification.

With regard to the other witnesses, not one of them had any prior acquaintance with the driver of the car and each one of them saw him for a very short time on the night of December 14th, and all of them must have known by December 16th, at the latest, that the driver of the car was wanted, and whether any special impression on their minds had been made. The value of their evidence is greatly weakened by their lack of opportunity for observation, their failure to record any peculiar characteristic of the driver, and their all having seen published photographs both of the girl in the car and of Craig. The witness Watson was not taken to identify Craig until January 17th, but failed to identify him upon a "line-up" stating that he did not take particular notice on the night in question. The witness Weeks did not identify until January 29th, when Craig, who had been charged with the murder, was sitting at the bar-table in the Coroner's Court along with his solicitor. This was an unsatisfactory identification, and clearly Weeks was in considerable doubt about the matter, as is shown by his statement that he could not "swear" to Craig. Mrs. Watts was given, or sought, no opportunity of identifying Craig before January 20th. She was in touch with the police as early as December 16th, being taken out to identify Bessie O'Connor. She made a statement to the police at once, describing the driver in detail and as having "fairly broad Irish features." But to such description, Craig would appear not to answer.

Whilst there is no other evidence implicating Craig, the evidence of Brown, Stiff and Lawrence considerably weakens the conclusion that Craig was the driver. Brown certainly had special opportunities for observing the driver of the car. The witness Lawrence says that he saw a girl, whom he believed was Bessie O'Connor, leave a car near her home in Redfern and go towards her home, while the driver got out and waited for her upon the pavement. He spoke to the driver, and says that the driver was not Craig. He made a statement to the police upon the day after the murder, which was in accordance with his evidence, and he was taken to see whether he could identify Craig on January 7th or 8th. The witness was subjected to a somewhat harsh cross-examination at the hands of the Crown Prosecutor, which, so far as the evidence discloses, was quite unjustified.

The cases to which Mr. Curtis referred in his impressive argument show clearly enough that, in England, the Court of Criminal Appeal does not hesitate to set aside a conviction where there is a reasonable possibility that witnesses as to identification have been mistaken. The course of decision becomes the stronger when it is remembered that the English Court is not empowered to order a new trial, but must, if it intervenes at all, quash the conviction and terminate all proceedings upon the indictment. The reason for the special care which the English Court has accorded to cases dependent upon evidence as to identity is probably to be discovered in the historical fact that a proved instance of miscarriage of justice through honest mistake in identification (the case of Beck) led to the establishment of the Court of Criminal Appeal in England. So, too, in Scotland, the Slater Case, in which evidence of identity played so important a part, led to the establishment of the Scottish Court of Criminal Appeal.

In our opinion, the summing-up of the learned trial Judge upon the question of the identity of the driver with the prisoner, was insufficient to bring home to the minds of the jury the possibility or probability of mistake upon the part of the witnesses. Moreover his Honor said: "A witness can state his belief and if that is supported by a sufficient number of other persons who also state their belief, or if supported by outside circumstances, that is sufficient for the jury to determine the matter upon."

It seems to us, that, in this passage, his Honor indicated to the jury that the absence of corroboration might be balanced by "a sufficient number" of other witnesses. With all respect, this seems to us to amount to an invitation to the jury to pay regard to the mere fact that four persons gave their sworn opinion at the trial that the driver of the car was Craig and to consider the matter upon the footing that the evidence of each witness acquired additional probative value by reason of the presence of the testimony of the other three. This was calculated to induce, and quite possibly had the effect of inducing, the jury to aggregate rather than to analyse the evidence of these witnesses and to diminish the significance of the failure of other witnesses to identify Craig. The jury was not warned, as we think they should have been, as to the special care with which they should approach and weigh the evidence of identifying witnesses in cases where there is no other evidence implicating an accused person. An illustration of such a warning is contained in the charge to the jury of Lord Guthrie, who presided at the Slater trial (Notable British Trials—Trial of Oscar Slater, 3rd ed. (Roughead) (1929), pp. 238, 239). He said:—

Next we must consider the evidence of identification and its value. Not a word too much has been said on that matter by the Lord Advocate and Mr. M'Clure. It is extremely important. I express the point thus—it would not be safe to convict the prisoner merely on the evidence of personal impression of his identity with the man seen flying from the house, on the part of strangers to him, without reference to any marked personality or personal peculiarities, and without corroboration derived from other kinds of evidence. My proposition involves a distinction between the identification, by personal impression, of a strange person, and the identification, by personal impression, of a familiar person. Suppose that a father told you that his son, who was resident in his house, had been seen by him in Princes Street yesterday. That would be admirable evidence. But if a person who had only seen the son once in his life told you that he had seen him in Princes Street yesterday, that would be evidence of slender value, unless the son had a marked personality, or unless he had some peculiarity about him, such as a very peculiar walk, or unless there were corroboration, such as that the man, when spoken to, answered to the name of the particular individual. ... Then, again, people differ as to the extent of a resemblance, or even whether there is any. You may have seen a strong resemblance, but one of your friends says that he can see no resemblance at all, and, when the two people are brought together, you see that there is nothing but a very general similarity. That applies to the personal impression of a stranger in reference to a stranger.

When, many years after the trial, the charge of Lord Guthrie came to be reviewed before the Scottish High Court, under the provisions of the Criminal Appeal (Scotland) Acts 1926-1927, the validity of this portion of the charge was sustained. In our opinion a warning of a similar character was called for in the present case. It was not given, and the jury's attention was not expressly directed to the importance of the various matters to which we have referred in this judgment.

For the above reasons we think that the conviction should have been quashed by the Supreme Court and a new trial ordered. As has already been pointed out, on the appeal to the Supreme Court, counsel for the prisoner devoted his efforts to the significance of the new evidence which suggested that a person other than Craig was implicated in the murder. The chief witness called to substantiate this suggestion was one Crothers. He was cross-examined before the Full Court, and the learned Judges gave no credence to his evidence. A handkerchief alleged to have belonged to Bessie O'Connor was produced to the Court, and there is evidence upon which it might be found that the handkerchief did belong to her. The learned Judges did not, moreover, reject the possibility that the handkerchief was hers, but rather proceeded upon the basis that, even if it was hers, Crothers' account as to how he came by it, was not thereby corroborated. It is correct that the mere finding of the dead girl's handkerchief, if such it was, would not corroborate Crothers' account of the events of the evening of December 14th. None the less, if the handkerchief did belong to Bessie O'Connor, the conclusion that Crothers is an untrustworthy witness would not necessarily destroy its importance and significance in the minds of a jury. There is nothing suggested which would account for the presence of Bessie O'Connor's handkerchief at the place and time it was found, except (1) a conspiracy to pervert the course of justice, to which some person in possession of a handkerchief of the dead girl was a party, or (2) Crothers' having obtained possession of the handkerchief under circumstances tending to implicate either him or some other person, or both, in the crime of murder, or of being accessory after the fact of such crime. A jury might reject the former and accept the latter hypothesis.

There is another important aspect of this branch of the case. On January 12th last, Crothers made a statement to the police, and again on February 14th he made a signed statement to the police. There was considerable inconsistency between the statements, but, in the second statement he purported to disclose the existence of a bloodstained handkerchief, with the letters "B.O." embroidered in the corner. The officers of police suspected Crothers' bona fides and undertook the responsibility of spending no further time in investigating his story. They say that they inquired into the movements of the person who, according to Crothers, deposited the handkerchief in the car, and that, a satisfactory alibi being obtained, they dropped the matter. By that time, of course, their efforts were being directed towards proving that Craig was the driver of the car. Crothers' statement, for whatever it was worth, tended to implicate someone else. Under such circumstances, we think Crothers' statement should have been made available to those advising the defence. It is quite possible, even probable, that Crothers would not have been called for the defence, the success of which might have been seriously imperilled by tendering evidence of such a character from such a source. On the other hand, those advising the defence are obviously the best judges as to whether it is desirable to call a witness, or, without calling him, to make an endeavour to use information obtained from him. Of that opportunity, whatever its value would have turned out to be, the prisoner was deprived by the officers of police failing to disclose Crothers' statements, having decided for themselves that his evidence was not to be believed. Under analogous circumstances, which may easily be imagined, such a decision on the part of those engaged in proving that A is guilty of a crime, where the defence is that the crime was committed by B, might lead to a serious miscarriage of justice.

This case is therefore quite distinct from the ordinary case in which it is alleged that new evidence has been discovered after a conviction. Here, on the contrary, most of the evidence was available before the first trial of the prisoner in the form of statements in the hands of the officers of police. Indeed, the first statement was made by Crothers to the police before the coronial inquiry, and it seems to us that the Coroner should have been afforded an opportunity of deciding whether to call Crothers or not. As it turned out, the substance of Crothers' statements did not come to the notice of the prisoner or his advisers until after conviction and sentence of death.

Whilst, therefore, expressing no opinion that the new material, of its own weight and force, should produce a different result upon a new trial, we are clearly of opinion that the circumstance of its non-disclosure to the defence supports the conclusion to which we come independently that this is a case where the Court of Criminal Appeal should have ordered a new trial.

For the reasons we have given we think that the case warrants the granting of special leave. The matter having been fully argued on both sides, this Court should order that the conviction be quashed and a new trial had.

Application for special leave to appeal, by majority, refused.

Solicitors for the applicant, C. Jollie Smith & Co.

Solicitor for the respondent, J. E. Clark, Crown Solicitor for New South Wales.