Bates, R. v

Background

On 3 May 2005 in the Central Criminal Court before Fulford J. and a jury the appellant, Richard Bates, was convicted of murder and was sentenced to life imprisonment with an order that he serve a minimum term of 20 years. His co-accused, James Garside, was also convicted of murder and received the same sentence. The victim was Marilyn Garside, the estranged wife of James Garside. It was the prosecution’s case that Garside had hired Bates to murder her.

Marilyn Garside was stabbed and killed on 2^nd October 2001 when she answered the front door of her elderly mother’s house in Rose Lane, Romford. The prosecution alleged that Garside was the only person who knew that Marilyn would be visiting her mother, Mrs. Barbara Rawle, that day and that she would answer the door rather than her mother, who walked with difficulty.

At about 5:45 p.m. that afternoon someone rang the doorbell of Mrs. Rawle’s house and Mrs Garside went to answer it. As she opened the door she was stabbed several times and fell to the ground. Mrs. Rawle, who had gone to her assistance, saw a man walking slowly away from the house towards the garden gate. She was afraid that he might return, so she shut and locked the front door. She did not see the man reach the gate but when she looked out again he had left and had closed the gate behind him. Mrs. Rawle summoned an ambulance, but her daughter died soon after the attack. As a result of the attack there were bloodstains in various places on and around the door. Blood stains were also found on top of the wooden gatepost and on the gate latch.

The appellant and Garside were tried together at the Central Criminal Court between April and June 2003. It was the appellant’s case that Marilyn Garside had been murdered on the instructions of her husband who had hired an unknown person to kill her. In support of its case that the appellant was the murderer the prosecution relied in part on partial profile DNA evidence which it submitted tended to show that the appellant had been present at the scene of the crime.

For much of the trial the prosecution kept open its position in relation to Garside, but after the evidence was complete it took the view that it could not invite the jury to convict him otherwise than on the basis that he had engaged the appellant to carry out the murder. As a result, despite objections from the appellant’s counsel, the judge in his summing up directed the jury that they should consider the case against Garside only if they had already reached the decision to convict Bates.

In due course, on 19^th June 2003, both men were convicted of the murder of Marilyn Garside. The appellant then sought leave to appeal against his conviction on two grounds: one relating to the admissibility of the DNA evidence, the other relating to the judge’s direction that the jury should only consider the case against Garside if they had already reached a decision to convict the appellant (what later became known as the “path to conviction”). The application was referred by the single judge to the Full Court which on 2^nd July 2004 gave leave to appeal on the second ground and quashed the conviction: [2004] EWCA (Crim) 1751. It did so, however, only because it was satisfied that counsel had misunderstood the effect of the judge’s ruling and that as a result the appellant had not had the benefit of the final speech to which he was entitled. In those circumstances the court did not find it necessary to consider the first ground of appeal and, having quashed the conviction, ordered that there should be a re-trial. However, it is relevant to mention that in the course of giving judgment Hooper L.J. said that, given the way in which the prosecution had presented its case, Garside would have been denied a fair trial if the judge had left the case against him on an alternative basis and that the judge had given the only direction that he could have given under the circumstances.

Following the quashing of the appellant’s conviction Garside also lodged an appeal against conviction. His appeal was allowed and a re-trial was ordered in his case as well. As a result a second trial of both men took place at the Central Criminal Court before Fulford J. between February and May 2005.

At the second trial the prosecution made it clear from the outset that its case was that Garside had hired the appellant to kill his wife and that it would not invite the jury to convict him unless they were sure that appellant had committed the murder. In support of its case that the prosecution sought to rely on the same partial profile DNA evidence as had been adduced at the first trial. Having held a voir dire, the judge ruled that the DNA evidence was admissible and in due course it was placed before the jury. As at the first trial the judge directed the jury that they could not convict Garside unless they were sure that the appellant was the person who had killed Marilyn Garside. In so doing he no doubt had in mind what this court had said when giving judgment on the first appeal.

The appellant sought leave to appeal against his conviction on the same two grounds as before. The single judge gave leave on the first ground, which had not been considered by the court on the previous occasion, but referred the application on the second ground to the Full Court.

In R v Doheny & Adams [1997] 1 Cr. App. R. 369 Phillips L.J. gave a succinct description at pages 371-372 of the nature of DNA itself and the process by which it is analysed in order to determine the statistical likelihood that the sample of material from which it was taken came from a particular person. However, it was common ground on the present appeal that the techniques now in use, although in substance the same as those described in that case, involve an extension and refinement of those that were in use in 1990 when that case was being prepared for trial. The jury in the present case were provided with a simplified summary of the nature of DNA and the techniques now adopted in the analysis of samples under investigation and we think it may be useful to describe those features of the process which have a particular bearing on the present appeal by reference to that summary. The process currently employed is known as ‘SGM Plus’, but no doubt as time goes on the techniques we describe will be further extended and refined and this summary will become out of date in its turn.

As is well-known, DNA is a complex molecule in the form of a double helix. DNA analysis ultimately relies on the fact that different regions (or “loci”) contain repeated blocks of material known as “alleles”. The loci are given individual designations (“D3”, “D8” etc.) and the analysis is directed to 10 loci at which the alleles are known to vary widely between individuals. Although the loci at which the alleles are found are the same in everyone, the number of blocks making up the alleles at each locus differ from person to person. An allele formed of 17 blocks would be described as “allele 17”. At each locus there are two alleles, one inherited from the father and one from the mother, so, for example, a person might have alleles 14 and 17 at locus D3. That is normally designated “D3 14, 17”. In addition to the 10 loci the analysis also includes a sex indicator, amelogenin. This is ‘X,X’ in females and ‘X,Y’ in males.

A person’s DNA profile is currently built up by reference to the alleles present at the chosen 10 loci and the sex indicator. This represents an advance on previous techniques which we understand were limited to 6 loci. In due course it may be possible to refine the technique still further by including additional loci. The identification of alleles is carried out by gel electrophoresis. This process uses an electric current to draw samples of DNA through a gel and separate the alleles. Lasers are used to detect coloured markers that have been applied to the sample earlier in the process and the resulting data are fed into a computer which produces the results in graphical form. The interpretation of the graphs calls for a high degree of skill and experience and can give rise to differences of opinion, as indeed occurred at the trial in the present case. However, it is unnecessary to describe that aspect of the process in any greater detail because it was accepted that for the purposes of the appeal the summary of the results produced by the prosecution could be accepted as correct.

If a fresh sample of DNA from a single contributor is obtained the analysis will produce a complete profile for the person from whom it was taken. Such a profile will identify 2 alleles at each of the 10 loci together with the sex indicator. (We use the term “complete profile” in the sense that it is complete in relation to the 10 loci analysed, although many other loci exist in respect of which no analysis is undertaken.) When testing material for a match with a particular suspect the first step, therefore, is to obtain a complete profile of the suspect’s DNA for the purposes of comparison. A profile of DNA obtained from stains, hair or other materials found at a relevant location can then be prepared in the same way and the two compared. Data drawn from empirical research is available to enable analysts to calculate the statistical likelihood of any person within the population having a particular allele at a particular locus. Using that data it is possible to estimate the statistical likelihood that a particular sample of DNA originated from the person whose profile is being used for comparison. This is usually referred to as the “match probability”.

The process of preparing a sample for analysis involves copying a known amount of the DNA within it a number of times by a process known as a polymerase chain reaction (“PCR”). It is unnecessary for the purpose of this case to describe that process in any detail; its significance for present purposes lies in the fact that it can generate pieces of DNA which were not present in the original sample. These are known as “artefacts”. The most common of these artefacts appear on the graph as a low peak one unit below the true peak which denotes an allele. These are known as “stutters”. Stutters are frequently observed in profiles produced by the method we have described and account has to be taken of them when interpreting the results of the analysis, but they are well documented in the scientific literature and it is accepted that they do not prevent a reliable analysis from being produced by an experienced analyst.

The procedure as we have described it assumes that a full profile can be obtained of the DNA recovered from the scene of the crime or other relevant location and that the sample contains the DNA of only one person. However, in practice samples often contain the DNA of more than one person, in which case the analysis will produce what is known as a “mixed profile”. A mixed profile can be identified by the presence of more than two alleles at any single locus. In such cases it is necessary to identify the number of contributors to the profile and to establish separate profiles for each of them. This gives rise to certain difficulties in the interpretation of the results of the analysis to which we shall return. Moreover, even in a case where there is only one contributor to the sample, it may not be possible to obtain a complete profile, that is, to identify two alleles at each of the 10 loci. A profile in which, for whatever reason, some alleles cannot be identified is referred to as a “partial profile”.

In a mixed sample originating from two or more persons it is often the case that one person (the “major contributor”) will have contributed much more of the DNA present than the others (the “minor contributors”). That results in higher peaks appearing on the graph at the locations of the major contributor’s alleles and lower peaks appearing at the locations of alleles obtained from the minor contributors. Where the major contributor and a minor contributor have the same allele at the same location the peak produced by the minor contributor’s allele will be hidden by that produced by the major contributor’s allele. This phenomenon is known as “masking” and may account for the apparent absence of an allele belonging to the minor contributor. The presence of a stutter in the profile of the major contributor may also mask an allele in the profile of the minor contributor.

If only a partial profile can be obtained from the sample under test there will be some loci at which only one allele, or perhaps no alleles at all, have been found. That may be due to a variety of causes which include masking, the loss of some molecules from the sample and the tendency of molecules with a high molecular weight to degrade. In very rare cases there may be no allele at that locus. Such “voids” are potentially significant because, if the missing allele did not match either of the alleles at that locus of the person under investigation, it would establish conclusively that he (or she) had not provided that sample of DNA. Every partial profile carries within it, therefore, the possibility that the missing information excludes the person under investigation, but there is currently no means of calculating the statistical chances of that being the case.

With that introduction we can turn to the circumstances giving rise to this appeal. Samples of material were obtained from seven locations in the vicinity of the murder. Most were found on or near the front door but one was found on the wooden gate post at the end of the garden. They were all tested in accordance with the procedure described earlier and produced a series of mixed partial profiles. The results were presented in the form of a table produced by Miss Bridget March of the Forensic Science Service who had carried out the analysis. The table included for the purposes of comparison the full DNA profiles of the appellant and Marilyn Garside. There is no dispute for present purposes that the alleles identified in the table of results had been correctly called.

One of the samples taken from the front door handle (sample area 4a) contained DNA from several contributors and gave no reading for sex indicator. However, the results of analysis showed one allele at each of loci D8 and D19 which did not match the appellant’s profile, so it was clear that he cannot have been a contributor to it. It plays no further part in our discussion. The other samples contained a major profile and a minor profile. Although the experts could not exclude the presence of more than one contributor to the minor profile, it was accepted that there was nothing to indicate that there had been more than one contributor and the argument before us proceeded on the assumption that there had in fact been only one.

All the samples tested contained the male sex indicator (X,Y) showing that one of the contributors was male. Moreover, the other sample taken from the front door handle (sample area 4) and the sample taken from the gate post (sample area 2) each contained eight alleles matching those of the appellant. (For reasons which will become apparent, we think that it is helpful to concentrate on those two samples, although the principles applicable to them are equally applicable in the case of the other samples.) However, none of the samples tested showed any alleles at loci D16 or D18 and in the case of locus D2 one allele (allele 22) was found in three of the samples (including samples from areas 2 and 4) and no alleles were found in the remaining three.

When calculating the probability match for each sample Dr. Evett, the expert statistician called on behalf of the prosecution, attributed a value of 1 to each of the voids, treating it as neutral. On that basis he calculated the probability match in the case of samples 2 and 4 to be 1 in 610,000. This means that one person in every 610,000 has the particular combination of alleles found in those two samples. Put another way, if the total population of this country is assumed to be about 60 million, there are only about 100 people who share this particular combination of alleles. Moreover, if one assumes that about half of the population are male and half female, the presence of the male sex indicator means that in statistical terms this combination of characteristics is shared by about 50 males, some of whom are likely to be quite young and some quite elderly.

In the case of a full profile match the match probability is estimated as being about 1 in 1 billion (1,000,000,000) for an unrelated person. A full profile match is therefore very strong evidence that the sample in question came from the person who provided the DNA for comparison. A match probability of 1 in 610,000 is clearly much less compelling, but it may have strong probative value if, for example, there is other evidence linking the accused to the offence.

As the table of results in this case shows, at least twelve of the appellant’s alleles were consistently absent from the samples tested, but it was accepted that for all the loci apart from D2 and D18 their absence is satisfactorily explained by masking. It was agreed that their absence can therefore properly be treated as neutral for the purposes of calculating a match probability. As a result attention was directed to the voids at loci D2 and D18, particularly that at D18 where no alleles were reported.

At the trial Mr. Miskin Q.C. submitted that the judge should exclude the DNA evidence on the grounds that it is impossible to ascribe any statistical value to the potential exculpatory effect of the voids in a partial profile. Professor Balding, the expert statistician called on behalf of the defence, said that it is not correct to ascribe a value of 1 to the voids because of the potential exculpatory effect of the unreported alleles, but since it is not currently possible to calculate that effect in statistical terms it is not possible to produce a true match probability. Thus, in the case of the two samples taken at areas 2 and 4, the true statistical match probability is not 1 in 610,000 but something rather lower.

The judge rejected that submission. In a detailed and lucid ruling he explained the nature of the evidence and described the choice facing the court in the following terms:

“13.

What are the consequences of the impossibility of assigning a statistical weight to the voids? The alternatives are to exclude the evidence entirely or to admit it subject to an appropriate warning to the jury of the limitations of the evidence, and particularly highlighting the fact that although what was found was consistent with Bates’ DNA profile, the voids at D2 and D18 in particular may have contained an allele or alleles, the presence of which would have been wholly exculpatory.

14.

In arriving at the correct conclusion it is important to remember that scientific evidence frequently only provides a partial answer to a case, or to an issue in a case. However, the test of admissibility is not whether the answer is complete, but whether science can properly and fairly contribute to the matter in question. . . . . . . . . ”

He held the evidence was admissible and that it could safely be put before the jury who could evaluate it together with Prof. Balding’s reservations and his explanation of significance of the voids. He was fortified in his conclusion by the fact that Prof. Balding in his evidence on the voir dire had himself agreed that there was no objection to adopting that course if the judge was of the view that the jury could grasp the concepts involved.

The primary ground of appeal in this case rests, as did the application before the judge, on the impossibility of ascribing any statistical value to the potential exculpatory effect of the voids in a partial profile. In the hands of Mr. Miskin this ultimately found expression in two submissions: (a) that the effect of the decision of this court in Doheny and Adams is that only statistical evidence can properly be placed before the jury in relation to DNA analysis and that in the case of a partial profile the inability to take account of the potential exculpatory effect of voids invalidates any match probability; (b) that to invite the jury to assess for themselves the evidential value of a partial profile, having explained to them the potential significance of the voids, is to invite them to embark on an exercise which they are ill-equipped to undertake because it requires them to weigh up something which is inherently unquantifiable. He submitted that whether or not the test put forward by the judge in paragraph 14 of his ruling was correct in relation to other kinds of scientific evidence, it was not the correct test to apply in the case of DNA analysis.

Perhaps the first point that should be made is that the evidence derived from the testing carried out by the Forensic Science Service in the present case was presented to the jury in the form of statistical match probabilities of the kind contemplated in Doheny and Adams. Moreover, we can find nothing in that case to support the proposition that only match probability calculations which take into account the statistical value of every conceivable possibility are admissible in evidence or that evidence based on partial profiles must be rejected in every case. That does not reflect the approach taken by Prof. Balding in his evidence to the judge on the voir dire, nor do we think it is consistent with good sense. The court in Doheny and Adams was primarily concerned to provide guidance of a general nature in relation to the presentation of DNA evidence and to expose and eradicate the so-called “prosecutor’s fallacy” which elevated the significance of the evidence beyond its proper level. It was not concerned with the distinction between full profile evidence and partial profile evidence, although there are passing indications in the judgment that the court may have had partial profile as well as full profile evidence in mind - see, for example, the reference to “the frequency with which the matching DNA characteristics are likely to be found in the population at large” at page 374E. Moreover, it is necessary to bear in mind that in one sense all profiles currently obtainable are partial inasmuch as present techniques only allow testing at 10 loci. We were told that at the time when the tests considered in Doheny and Adams were performed the analysis was carried out by reference to 6 loci; now it is carried out by reference to 10 and we were told in the course of argument that advances in technology may make it possible to test by reference to many more. It remains the case, however, that the presence of one allele at one locus that does not match the profile of the comparator is sufficient to exclude that person as the contributor.

We agree that the starting point, as Mr. Hilton Q.C. for the prosecution submitted, is to ask whether the evidence is relevant in the sense of being probative of a matter in issue between the prosecution and the defence. In our view it plainly is. It is important to understand that the results of the testing procedure and the statistical analysis based on them indicate what proportion of the population has the reported alleles at the relevant loci. In the case of the samples taken from areas 2 and 4 it is one person in 610,000 in each case, or a total of roughly one hundred persons in a population of 60 million. That would remain the case even though there might be an allele in one of the voids which exculpated the appellant. If, on the other hand, a “missing” allele matched the appellant’s profile, the match probability would be reduced and the chances that the sample had been contributed by the appellant increased accordingly.

We consider that the judge’s approach to the question was entirely correct. We can see no reason why partial profile DNA evidence should not be admissible provided that the jury are made aware of its inherent limitations and are given a sufficient explanation to enable them to evaluate it. There may be cases where the match probability in relation to all the samples tested is so great that the judge would consider its probative value to be minimal and decide to exclude the evidence in the exercise of his discretion, but this gives rise to no new question of principle and can be left for decision on a case by case basis. However, the fact that there exists in the case of all partial profile evidence the possibility that a “missing” allele might exculpate the accused altogether does not provide sufficient grounds for rejecting such evidence. In many cases there is a possibility (at least in theory) that evidence exists which would assist the accused and perhaps even exculpate him altogether, but that does not provide grounds for excluding relevant evidence that is available and otherwise admissible, though it does make it important to ensure that the jury are given sufficient information to enable them to evaluate that evidence properly. Moreover, as the court observed in Doheny and Adams at page 373D, the significance of DNA evidence depends to a large extent upon the other evidence in the case. By itself such evidence, particularly if based on a partial profile, may not take the matter far, but in conjunction with other evidence it may be of considerable significance.

In the present case the judge summed the case up to the jury in an exemplary manner. He dealt with the DNA evidence fully and fairly in simple and straightforward terms and took particular care to ensure that they were made aware of its limitations and of the potential significance of the voids. We have no doubt that the jury were capable of understanding the evidence and of making proper use of it, guided as they were by the summing up. The fact that they could not assess with any statistical accuracy the chances that there might have been a “missing” allele which exculpated the appellant did not prevent them from making proper use of the evidence for what it could establish. Moreover, this is not a case in which the DNA evidence stood alone; there was other evidence linking the appellant to the scene of the murder which provided the context in which its significance fell to be assessed.

In our view the judge was right to admit this evidence. We are unable to accept the submission that the conviction is unsafe on this ground.

The second ground of appeal is that the judge was wrong to direct the jury that they should only consider the case against Garside if they were sure that the appellant was the person who killed Marilyn Garside. The single judge referred the application for leave to appeal on this ground to the Full Court, mindful, no doubt, of the fact that the court had considered and ruled on that question on the first appeal.

There was no argument as such on this matter before the trial judge. Counsel for the appellant did raise the question, but only in a way which suggests that he was seeking to ensure that his concern appeared on the record. The judge made it clear that he intended to direct the jury in those terms and nothing more was said about it.

The issue which the appellant seeks to raise on this occasion is precisely the same as that which was before the court on the first appeal and it arises in the same context, save for the fact that at the second trial the prosecution had made its case clear from the outset and counsel for the appellant (who had also appeared for him at the first trial) had the benefit of this court’s analysis of the position. That included, of course, its finding that on the previous occasion there had been a misunderstanding as to the effect of the judge’s ruling so far as the appellant was concerned. The judge at the second trial directed the jury in accordance with the observations of this court on the first appeal, to which we have referred, and in those circumstances we had some doubt whether it was open to the appellant to invite the court to reconsider the question. However, we allowed Mr. Miskin to make his submissions and are content for present purposes to deal with the question on the merits.

In our view the considerations which arose at the second trial were for all practical purposes precisely the same as those which arose at the first trial. The prosecution’s case was that Garside had engaged the appellant to kill his wife and that the appellant had done so. It was no part of its case that, if the appellant did not kill Marilyn Garside, she had been killed by some other person whom Garside had hired to carry out the murder. From Garside’s point of view that was an important matter because it limited the case against him and so restricted the range of possibilities which he and those representing him had to address. It would have been quite unfair to him for the judge in his summing up to have left the case to the jury, explicitly or by implication, on the basis that they could convict Garside even if they were not sure whether the appellant had committed the murder. In paragraph 30 of the judgment given on the first appeal Hooper L.J. said

“Given the way in which the prosecution had presented its case, Garside would have been denied a fair trial if the judge had left the case against Garside on an alternative basis. Thus, the only way for the judge to have given effect to what Mr. Miskin wanted him to do would have been for the jury to be directed that when considering Garside’s case, they could only convict Garside if they convicted Bates, but that when considering Bates’ case they could convict Garside without convicting Bates. The jury would have been utterly bewildered if they had received that direction. The only direction which the judge could have given in the circumstances was the one he gave.” (Emphasis added).

In the light of that observation it is difficult to see what other direction the judge at the second trial could have given. Mr. Miskin submitted that the effect of the judge’s ruling was to elevate the direction from one dealing with the evidential position to one of law and that as such it was incorrect because a secondary party may be convicted despite the fact that the principal is acquitted (for example, because the evidence admissible against one is not admissible against the other). As a matter of law, that may be so, but that is not the issue in this case, as the court’s judgment on the first appeal makes clear. The issue here is simply whether, in the light of the way in which the prosecution had put its case against Garside, it would have been fair to him to leave the case to the jury on an alternative basis. Plainly it would not.

For these reasons we consider that the judge was clearly right to direct the jury as he did and accordingly leave to appeal on this ground is refused.