This is the judgment of the court.

This appeal arises from the conviction of the appellant, then a care worker, at Exeter Crown Court before His Honour Judge Cottle and a jury, of a number of allegations of sexual assault on women with a mental disorder. He is now 58, a man with no previous convictions, and a wealth of evidence supported his positive good character. The four complainants were residents in the same residential care home at which the appellant was a part time worker. All four women were profoundly disabled and wheelchair bound, three with cerebral palsy, and one tetraplegic with an acquired brain injury.

The appellant’s contention, advanced by Miss Munro QC, in short, is that the evidence against him was non-existent or so weak that the charges on which he was convicted should not have been allowed to go to the jury. He says none of the complainants gave coherent and reliable evidence on which a reasonable jury properly directed could have convicted with any safety. The Crown on the other hand, argues that the special measures properly adopted to enable the complainants to give evidence, functioned as they were meant to, and gave the women access to the criminal justice system and produced convictions which were safe.

All four women lived at a residential home, in North Devon, which was a care home as defined by section 42 (5) of the Sexual Offences Act 2003 ( “the 2003 Act” ) and was run by a charitable foundation.

TB had cerebral palsy resulting from brain damage acquired in infancy, probably as a result of abuse in the form of “shaken baby syndrome”. She had lived at the home since 1994. Like the other women she could not stand or bear weight, had to be transferred to and from her chair by means of a hoist and was entirely dependant on others for her daily care needs. She could not communicate by any means whatever other than by shouting, spitting and swearing when distressed. She could utter only occasional random words. JR had profound cerebral palsy of perinatal origin and similar disabilities. She was capable of using a form of electronic communicator to respond to questions but found its use tiring and irksome. JB suffered from the same condition as JR and had much the same level of disability. She could only communicate by movements of her eyes. In order to signify “yes” she would raise them upwards towards the ceiling. In order to say “no” she would move them, and to an extent move her head, from side to side. She too finds the process of communicating demanding. These women suffered from a mental disorder within section 79 of the Sexual Offences Act 2003, namely mental impairment within the meaning of section 1 of the Mental Health Act 1983. None of them is thought to have had any previous sexual experience.

The fourth complainant, SC, was in a different position. She suffers from spastic tetraplegia following a subarachnoid haemorrhage in 1996. She was 51, married and a mother and grandmother. In cognitive terms by comparison with the others she was relatively intact at the time she gave her ABE interview on 14 October 2008. By the time of the trial, however, she had suffered a stroke and was entirely incapable of speech or any other form of communication, so that her ABE interview was admitted under s.116 (2) of the Criminal Justice Act 2003.

The structure of the indictment and the different verdicts returned on it by the jury are highly significant.

Count 1 stood on its own, and alleged sexual activity with a person with a mental disorder impeding choice contrary to section 30(1) of the Sexual Offences Act 2003 (“the 2003 Act”). The allegation was that the defendant had touched the breast of TB in a sexual manner. The evidence on this count came entirely from a volunteer worker at the home who said she had witnessed it. The defendant denied the offence. It is accepted that there was evidence on which the jury was entitled to convict him. It is not and has not been suggested that it should have been tried separately from the other counts. When summing up the case to the jury the judge said this (at 27F)

The next group was counts 2 to 4, involving JR. All these counts (as well as the following group of counts involving JB) allege sexual activity with a person with a mental disorder by a care worker contrary to section 38(1) of the 2003 Act. Count 2 alleged an intentional and non-penetrative touching; the prosecution were alleging touching with the hand in the region of her anal hair. The jury convicted on this count. Counts 3 and 4 each alleged penetrating the complainant’s vagina or anus with his penis in circumstances where the touching was sexual and the jury acquitted on both counts.

The next group of counts involved JB. Counts 5 and 6 were specimens and alleged intentional touching of the complainant by penetrating her vagina or anus with his fingers in circumstances where the touching was sexual. The jury convicted in each case, but only by means of the alternative verdict of touching her vagina or anus without penetration. Counts 7 and 8 were specimen counts alleging that he had penetrated the complainant’s vagina with his penis; the jury acquitted on these counts.

The final group of counts was counts 9 to 13 involving SC. Counts 9 and 10 each alleged sexual assault contrary to section 3 of the SOA in that he had, on two occasions chosen as specimens, intentionally touched the complainant, the touching being sexual and the complainant not consenting to it. The jury convicted on both counts. Counts 11, 12 and 13 charged rape contrary to section 1(1) of the Act on each occasion alleging a vaginal rape of the complainant. The jury returned not guilty verdicts on each of these counts.

The jury’s reasoning therefore emerges from their verdicts. It used its acceptance of the direct evidence of the able bodied volunteer in count 1 to support the evidence of the other three other complainants, but only to the extent that they alleged manual and non-penetrative touching. The jury either rejected or were unsure whether allegations of touching with anything other than the hand, or penetrative touching of any kind, was proved. The verdicts plainly reflect a very careful analysis of the evidence, in the context of the directions the jury was given about the burden and standard of proof.

No grounds of appeal are advanced against the conviction on count one.

The grounds of appeal against the remaining convictions fall into three main parts. First, criticisms are made of the failure properly to assess and examine the three complainants psychologically before the ABE interviews were carried out. In the cases of JR and JB that would or should, says the appellant, have included exploring inter alia their respective knowledge of sexual matters and anatomy and in the case of SC investigation of her manifest tendency to confabulate as a result of her particular injury (clear examples of which emerged in her evidence). On the basis of proper assessment, conducted before the ABE interviews took place, a proper use could have been made of intermediaries and reliable evidence may have been forthcoming. A better rapport could have been established with the witness, the witnesses’ knowledge of the difference between truth and lies should have been explored, and the need for the witnesses to appreciate both the need for truthfulness and the seriousness of the matters they were being asked about should have been emphasised.

The second ground is that the evidence having emerged as it did the case should have been withdrawn from the jury either at the end of the prosecution case or of the judge’s own motion, at the end of the evidence as a whole as being either non – existent or so weak and tenuous that no jury could properly convict.

The third ground is that the summing up failed to direct the jury properly as to the care to which they should look at the complainants’ evidence and that it was an unbalanced summary of the evidence, omitting many points favourable to the appellant.

Less than half a generation ago the criminal courts would not have contemplated attempting to receive evidence from persons in the position of these complainants. The Youth Justice and Criminal Evidence Act 1999 (“the 1999 Act”) introduced a radical new regime by which special measures were made available to enable vulnerable witnesses (including witnesses with major communication difficulties) to give evidence, or to improve the quality of the evidence. This is the first occasion on which the evidence of complainants suffering from such profound levels of disability has been brought to the court’s attention. We should therefore set out the relevant sections of the 1999 Act.

For the purposes of this Chapter a witness in criminal proceedings (other than the accused) is eligible for assistance by virtue of this section –

In determining for the purposes of this chapter whether any special measure or measures would or would not be likely to improve, or to maximise so far as is practicable, the quality of evidence given by the witness, the court must consider all the circumstances of the case, including in particular –

A special measures direction may provide for a video recording or an interview of the witness to be admitted as evidence in chief of the witness.

A special measures direction may, however, not provide for a video recording or a part of such recording to be admitted under this section if the court is of the opinion having regard to all the circumstances of the case, that in the interests of justice the recording or that part of it should not be so admitted.

In considering for the purposes of subsection (2) whether any part of a recording should not be admitted under this section, the court must consider whether any prejudice to the accused which might result from that part being so admitted is outweighed by the desirability of showing the whole or substantially the whole, of the recorded interview.

Where a special measures direction provides for a recording to be admitted under this section, the court may nethertheless subsequently direct that it is not to be so admitted if –

A special measures direction may provide for any examination of the witness (however and wherever conducted) to be conducted through an interpreter or other person approved by the court for the purposes of this section (“an intermediary”).

The function of an intermediary is to communicate –

Any examination of the witness in pursuance of sub section (1) must take place in the presence of such persons as rules of court or the direction may provide, but in circumstances in which –

At every stage in criminal proceedings all persons are (whatever their age) competent to give evidence.

A person is not competent to give evidence in criminal proceedings if it appears to the court that he is not a person who is able to –

The parliamentary intention which emerges from the 1999 Act is that those who are competent to give evidence should be assisted to do so. It is well understood that competence is not the same as reliability (see MacPherson [2006] 1 CR App R 30); Barker {2010} EWCA Crim 4) Provided the court is satisfied that the witness is able to understand the question put to him (or her) and give answers to them which can be understood, the competency test is satisfied. The Act further contemplates the reception of evidence in circumstances where a witness who satisfies the statutory test as to competence, may nevertheless lack sufficient communication skills to give evidence without the use of an intermediary. The use of intermediaries forms an integral part of the structure of the special measures regime. In the present case, as will be seen, two of the complainants gave evidence with the assistance of an intermediary (one of whom was not registered) by means of a process which was seen by the jury. Although arrangements were for a registered intermediary to be available to assist at trial, if required, Miss Munro decided that she should not cross-examine the complainants. The judge described that decision as understandable, but the opportunity for cross examination (with all the attendant difficulties) was available, at least in relation to JR and JB. Instead of formal cross examination of the witnesses, Miss Munro focused her attention, and that of the jury, on the areas of evidence which advanced the defendant’s case that the allegations against him were false.

Further, although section 27 (4) contemplates an inability to cross-examine as a reason for refusing to admit evidence in chief obtained by special measures, the Act leaves that decision to the discretion or judgment of the court. It therefore countenances the possibility that such evidence can be admitted even where no direct cross-examination is possible.

Accordingly, the challenge to the reliability of the special measures evidence was mounted by the defence by means of questioning the officer who interviewed the complainants, those who assisted with obtaining the ABE evidence of JR and JB, as well as the two experts who were not involved in that exercise, but were called by the Crown and gave evidence about the difficulties of relying on such evidence, and appropriateness or otherwise of the steps taken to obtain it.

On count 1, Sachiko Nisizawa gave live evidence. She was a Japanese volunteer at the residential home who herself gave evidence through an interpreter. She described how in 2007 she went with residents from the home in a big van driven by the appellant. Usually there were three residents on each trip and they went carriage riding at a stable every Monday, or every other Monday. All the residents were in wheelchairs and afterwards they usually went to lunch at a local Inn. She and the appellant helped them eat their lunch. She said she saw the appellant’s hand go under the aprons or bibs that the women wore and it was moving beneath the apron. Initially she thought he was merely holding their bodies to help them eat, but on one occasion in January or February 2008 she saw his left hand move on TB’s left breast between her body and the apron. She was hesitant about reporting the matter straight away, because her command of English then was very poor, but she eventually described what she had seen to a colleague who called in the manager and an investigation was launched. She accepted that when she first reported it to her friend she had said that “Jim had tickled TB”. The appellant’s hand was moving when he touched her left breast, she said, and she demonstrated this to the jury on her own person, showing the fingers moving over the breast.

As to Counts 2, 3 and 4 JR gave evidence through the medium of her electronic communication device. That took the form of a tablet computer mounted on her wheelchair onto which had been programmed a number of pages relating to different topics. Each page took the form of a grid on which were drawings, picture or symbols and a square for Yes or No. Though JR could not read, save to a limited extent, there was text added to each square. An illuminated cursor ran along the rows. When the cursor reached the row in which her desired answer lay she activated a switch in the left hand wing of her headrest by touching or bumping the left side of her head against it. The cursor would then run down the columns and when it reached the correct column she would activate the switch again. Where the two intersected an electronic “voice” would speak the written text. One of the pages contained pictures of all care staff at the home. Another included pictorial depictions of male and female body parts and of actions such as oral sex, sitting on someone’s lap or intercourse.

Using this device JR made video taped interviews on 4^th and 12^th June. The first of these lasted for 81 minutes. The intermediary was Amanda Postbechild a communications support officer in the employment of the charitable foundation, and not a registered intermediary. The form of the questioning, of necessity, was that closed questions only were asked. When shown the page with photographs of staff on she picked out the appellant and when asked if that was someone she wanted to talk about she responded Yes. When shown the body parts page and asked which part of the body or picture they wanted to talk about she selected “penis”. Asked if they were talking about Jim Watts’s penis she responded yes. She also selected “cheek” as another body part she wanted to talk about and the activity “touch” and the activity “punch”, agreeing that she was saying that Jim Watts had punched her. Asked where, she identified a square showing a bottom being approached by an erect penis. She confirmed that she was saying that he had punched her anus with his penis and this had happened somewhere not at the home in the van more than once. Asked if his penis had gone inside her bottom she indicated she did not understand the question but when it was repeated answered yes. Asked how she felt she selected the icon for crying.

In the way we have summarised it the evidence sounds as though it was given quite smoothly. In fact when seen on the DVD it is a tortuous and lengthy process, almost painful to watch, with long pauses as she watches the cursor move to the correct place. Sometimes it makes more than one pass down the rows before she activates her switch. When he came to sum the case up the judge, wisely as we believe, did not attempt to give the kind of summary we have set out above, but left the jury with the transcripts of this and all other interviews, from which to reach their conclusions. This unusual step was taken with the consent of both counsel, and seems to us to have been a sensible approach in the circumstances facing the court.

After about forty minutes of this process, which involved as we have said great concentration and effort on the part of JR, her answers degenerated and she began to answer “No” to most questions. She began to mention, almost at random, the names of other staff including a female employee. Not before time, it would seem to us, that initial interview was terminated on the grounds that what she was saying was unequivocally unreliable by that stage. She was plainly tired, as an expert psychologist was later to say in evidence.

Eight days later she was interviewed again in the same way. By means of the same process of closed questions and the selection of icons the narrative she gave ran as follows. She picked out Jim Watts. He had punched her and he licked her. He punched her on the bottom and the cheek. He punched and smacked her on the bottom. She was sitting on his lap but that was not where she was when he punched her. While she sat on his knee he touched her cheek with his hand and punched her cheek. Something happened to her hair when she was sitting on someone’s lap, somewhere inside the home. He touched her hair, the hair on her bottom and used his tongue to kiss her. She wanted him to do these things to her and was happy that he kissed her and touched her bottom. She was scared if she said no to such questions she shouted and she cried. Again, obviously as delivered this was not a coherent narrative in the way this judgment presents it. But this second interview contains the passage which the prosecution rely on as the evidence the jury must have accepted, about the touching of the hair on her bottom. The judge left it to the jury to make of it what they thought right, having seen the DVD and retaining their full transcripts.

To give an idea of how this evidence emerged we should set out verbatim the section of it on which the prosecution relied as indicating guilt on count 2. Up to this point she has been talking about being on someone’s lap and being “punched” on her bottom and cheek and she has said that something else has happened to her hair. She places this episode as somewhere in the home. MB is the officer interviewing, AP is the (non-registered) intermediary and “J” is the robotic voice of the computer.

DC Bishop, who conducted the interviews with each of these complainants, explained that she had taken advice beforehand from a Dr Kevin Smith, who had written on the subject of ABE interviews and who worked at the Police Crime Faculty at Bramshill and sat on the ACPO board. The decision was taken not to consult experts before the interviews but to do so after they had been concluded. Also no traditional “rapport building” was engaged in nor any examination of awareness of truth and lies or other psychometric testing.

Cheryl Berriman gave evidence of an expert nature albeit with some reluctance. Her originally intended role in the trial was to prepare JR for cross-examination in court and act as a mediator in court if asked. To that end she had assessed her needs and modified the icons on her communicator. She had added icons for “I don’t know” and “I don’t understand” and “not true,” and had also separated the Yes/No icons which had previously been adjacent on the original grid. Her assessment was that in other respects the selections of icons she had was reasonable but she stressed the need to avoid long questions and leading questions, by which she did not mean questions capable of being answered Yes or No.

Miss Munro QC says that viewed as a whole JR’s evidence is simply no evidence providing any basis for support of any conviction. It was not open to the jury to select or cherry-pick parts of it as reliable. It starts with an assumption of abuse, JR was given no familiarisation with the icons that were being used and there are numerous criticisms about over complex questions; for example it is assumed that this woman who had no sexual experience knew what a penis was. As a result of this process she made demonstrably false allegations against other staff which led to a third interview in November 2008 where she had alleged similar abuse by Ben Sutton which could be shown by a means of an alibi to be plainly false. But she explored at length with Ms Cutler in a cross-examination over some fifty pages of transcript the weaknesses of this process. The judge dealt with these in his summing up at page 24 in a passage concluding with a reminder to the jury to be aware of the possibility of compliance and suggestibility in the case of both these witnesses, the issue of confabulation in relation to SC and invited them to consider “what is to be made of apparently bizarre responses, responses that appear random illogical or unconnected with the question that has been asked”.

Ms Jennifer Cutler was called by the prosecution. She was a Consultant Forensic Psychologist with experience of assessing the capacity of those with learning disabilities mental impairment and autistic spectrum condition. She was asked to examine all three complainants who gave evidence. Of JR she said that communication was quite exhausting and physically demanding for her. Her facial affect was not a reliable indicator of her emotions. When she herself interviewed her JR was not keen to use the communicator she had used in the ABE interviews, so she resorted to the use of eye movements and a “card based” system. In the case of both JR and JB it would have been in her opinion “utterly impossible” for them to share information between them. Neither was capable of cognitively assimilating information or having an understanding of sexual matters. With her, as with JB, in reaching her own assessment she used what she called a “talking mats” system which she explained. This involved the use of cards with icons of the type used in the ABE interview, which she said were commonly used for such purposes.

We note the final position in which Ms Cutler’s evidence ended. After the various criticisms of the methodology had been gone through with her she said that a question such as “Is there somebody you want to talk about?” was open and appropriate and that asking the witness to choose that person’s photograph from among fifteen members of staff was sensibly limiting the field of possibilities. Her selections from the body parts page she described as making use of icons which were in constant use all over the world to meet the needs of people with poor communication skills. The question “Are we talking about Jim Watts’ penis?”, to which she answers Yes, Ms Cutler described as “sequential” and an acceptable way to help her give her evidence. She considered the process a good practice process, a clear indication as she put it “of somebody conveying something in their own language in their own way, that was them communicating a specific experience in an open way without being led, and being able to make choices that reflected their own language” and “that type of disclosure is considered to be quite clear as an indication of sexual abuse, and in someone who is non verbal, who does not have a language to describe sexual matters, that would seem to me to be a highly plausible way for her to communicate what happened”.

Miss Munro’s case is that JR has made certain allegations which are demonstrably false. She had a third ABE interview on November 18 after she had apparently made allegations of abuse by a member of staff Ben Sutton, using the same techniques, which came to nothing in that her responses were either negative or they were shown to be false. Against her evidence the appellant was able to put the lack of opportunity, there being no evidence of the two of them being alone together for any significant time, the manifest physical difficulty of her sitting on his lap, and the fact that when other witnesses said against him that he always took JR back to the van first after carriage riding and stayed there until the others arrived he had a reason for so doing because of her known allergy to horses, again a matter he proved by documents.

The prosecution pointed to additional evidence from the manager Mr Lawrence-Parr as to an in depth conversation he had with JR after the ABE was taken in which she said she “loved” the appellant and was angry with the police. Miss Munro says that her communicator had no icon for “like” and there is no transcript of this interview. There was also a conversation with JR in the presence of her mother in October 2008 when she indicated that she said she was not angry with Jim but was angry with the police and wanted to “swear” at them, there being an icon denoting that.

JB’s evidence was given in response to closed questions answered by her eye movements as described above. There was an intermediary present at her interview, and the service manager of the home James Lawrence-Parr interpreted her eye movements, he being familiar with her and with them. She said she did not remember Jim Watts but when his description was put to her she recognised it. She was shown four photographs and, once her glasses were put on, she said (correctly) that the first three were not Jim Watts and the fourth was. She agreed that he had done something she was not happy with by touching her arm but not her foot. She did not like him touching her leg or other parts; the other parts did not include her tummy but did include her breast. The touching was done with his hand just once. He touched her breast twice and had touched her bottom with his hand inside her clothing round the front inside her pants. She did not know what a vagina was. She said his hand had gone inside her pants at the front where she has a wee and her bottom. He used his fingers inside these parts more than once. He had used another part of his body not his foot but his penis. His penis had gone inside her pants vagina and bottom. He had not kissed her mouth but had kissed her cheek and breast. She felt uncomfortable. She did not want him to touch her breasts. She did want him to put his hand down her pants. She did not want him to touch her sexually round the vagina or put his hand inside her pants. This happened when they were carriage riding inside the car and outside it.

Again we should give a flavour of how this evidence came out by setting out verbatim the passage of her interview relied on by the prosecution in support of counts five and six. Again MB is the interviewing officer. JLP is James Lawrence- Parr and HD is Helen Donnelan who is a registered intermediary. This passage comes early in JB’s interview, and up to this point she has indicated that the appellant had done something that she did not like by touching.

After a break in the interview (which lasted 64 minutes in all) she said that the appellant did not have to change her pad inside her pants. He put his hand in for sexual reasons. There was another female resident there, Shirley. He got her to touch Shirley not her breast but between her legs. He did touch the hair underneath her pants not with his fingers but with his penis more than once.

Ms Cutler agreed that she was a highly suggestible witness and likely to be compliant in her answers to questions asked by a person in authority such as the policewoman or, particularly, Mr Lawrence-Parr. As to the latter Ms Munro points to 28 occasions in her interview when Mr Lawrence-Parr asks her to confirm or clarify an answer and she answers yes. We have looked at those answers on the DVD, and in most if not all of those occasions it appears to the court that the original answer probably was a “yes”. There is no instance where it is clear to us that Mr Lawrence-Parr’s question converts a No to a Yes. But it was the case that JB has no facility to say anything other than yes or no and she cannot indicate a lack of comprehension or cannot answer “I don’t know”. Again in her case there was no rapport building or proper preparation of her for the interview, says the appellant. There is no evidence that she was ever alone with him. As with JR, the appellant’s complaint is that the jury was not entitled to have been sure about the two allegations on which it convicted when so much of the rest of her evidence was plainly unreliable.

Ms Cutler interviewed JB in April 2009 as part of a lengthy and careful process. She established that she did know the difference between men and women and with the use of her card system and, with all her expertise and after careful ground-laying elicited a Yes answer to the question “Did Jim put his penis between your legs?”.

The evidence of SC was given without the use of an intermediary. The day before her interview on 14 October 2008 the interviewing police officer had spoken to her and asked her if there was anything that she wanted to say as the police had received information that something inappropriate had gone on between herself and a member of staff of a sexual nature. She immediately replied that there was, saying “Jim has been disgusting and touched my fanny”. She had no difficulties with comprehension and gave a coherent account, albeit in response to closed questions. The major problem with her evidence was a manifest tendency towards confabulation.

Though she could not initially remember the name she agreed that she knew Jim Watts and said he was horrible, always touching her private parts. He touched her fanny with his hand and nothing else. He was just playing with her. He also touched her breasts inside her clothing and inside her bra; when he touched her fanny it was inside her pants. He was smiling, threatening her not to tell or he would “get” her.

She then embarked on a crescendo of further allegations of a more and more serious nature. He made her touch him inside his trousers and would rub his penis against her on her legs and fanny. It got hard and wet when it went inside her several times a week in her bedroom. He kissed her cheeks and mouth putting his mouth on her breasts and fanny and put his tongue in her fanny. He never touched her bottom. He took her for rides alone and did it in the van. He held a knife to her throat, a carving knife. He rubbed himself against her at Ilfracombe beach when JB was with them in the van and he had sexual intercourse with JB. He put her on a fold-out bed erected in the back of the van and he also did the same to JB.

There were clearly many parts of her evidence where she was confabulating. The clearest examples would be her misnaming of her home village, her husband’s occupation and of the fact that she told her husband about what was happening to her and that he had threatened to kill Jim Watts. She also said she had told the staff who had done nothing. When her husband gave evidence he was clear that she had never told him anything. She finished by saying she had been raped daily by the appellant in her room behind a locked door. Her door in fact had no lock, and it would have been quite impossible for such a course of conduct, had it truly happened, to have gone unremarked, the appellant having no cause to visit the area of the residents’ rooms on any but the most rare occasions.

Ms Cutler described the process of confabulation in relation to those who had suffered acquired brain injury, who will therefore retain some memory of their lives prior to the injury. In essence it is the triggering in the patient of the recognition of the feeling that she would normally have a particular piece of information but does not. She therefore generates that information by the process known as confabulation. It is not, therefore, a process comparable with conscious invention. It can, said Ms Cutler, be quite extreme in its form and is usually easy to identify.

In respect of this witness (Ms Cutler having said it was “utterly impossible” for JR and JB to have contaminated each other or been contaminated as witnesses) Ms Munro raises the possibility of contamination. The appellant was suspended from his duties on 6 March 2008 and there had been a staff meeting about the allegation in August. A memorandum from Mr Lawrence-Parr to all staff on 26 September 2008 expressed concern that members of the public had become aware of detailed information relating to the investigation. Ms Munro argues that speculation must have been rife amongst staff and, though no staff member accepted that there had been discussions with or in the presence of residents, there existed the risk that this witness may have picked up some of the allegations that were being made and known about whom they were being made. We regard this as a speculative argument. SC’s husband was a regular visitor and had heard nothing from his wife about any such allegations; if she had heard staff talking about the reason for the appellant’s suspension we cannot believe she would not have told him about them and then about her own experience at his hands.

On 13 October 2008 DC Bishop spoke to SC as a result of JB naming her as someone who had been abused by the appellant. We have already set out her immediate unprompted reaction to that approach. The Crown unsurprisingly place reliance on this answer. The appellant argues that she failed to repeat that allegation, with similar alacrity at least, at the interview. She appeared then not to know who the police officer was talking about, and did not recall it until, having rejected a list of other names, the name Jim was put to her. Once that was done she immediately was asked what had happened with this man and said “Horrible” and that he was always touching her in her private parts, her fanny, touching her with his hands. She was asked whether he had touched her with anything else and she said no. He was just playing with her. He did touch her breasts inside her clothing and inside her bra.

From that point on she made her series of increasingly serious allegations demonstrating her tendency, says the appellant, to confabulate together with compliance and suggestibility. Her evidence is therefore worthless, it is argued, and the jury cannot extract one allegation from this otherwise bizarre interview and find it proved to the criminal standard.

There is no doubt that the judge applied the correct test in making his decision and he followed the established principle that, where the reliability of witnesses was crucial to the decision, generally speaking the judge should allow the matter to be tried by the jury. He recognised the unusual features and difficulties of the case, but posed the question whether the evidence was so demonstrably unreliable that it would be unsafe and wrong to leave any individual count to the jury. He acknowledged the significance of the matters elicited in cross-examination of Ms Cutler, but he concluded that the jury was fully equipped to assess the principal witnesses and their reliability.

In our judgement his decision was correct for the reasons which he gave. As the judge put it, to have removed the case from the jury would to have been to usurp their function, which was to decide the reliability of the three complainants in the light of such support as the jury believed that evidence received from other sources.

At the close of all the evidence, and in particular the evidence of the appellant, no further application was then made to him. Perhaps more important, we cannot see that any new factor had emerged at that stage to cause him to reach a different conclusion.

We discern an entirely logical thread running through the verdicts returned by the jury. Once the jury accepted the evidence of the eye witness in count 1, that meant that they were entitled to treat it as evidence which fortified the remaining complainants’ evidence of manual touching but without penetration. It was also significant because it showed in the eyes of a jury which was sure that the offence had taken place, that the appellant was prepared to interfere sexually with a disabled and highly vulnerable woman, and, for good measure, to do so in a setting where the presence of others meant that he was prepared to take the risk that his conduct might be seen. However nothing in count 1 supported any propensity in the appellant to go as far as the various penetrative sexual acts which were also described by the complainants. Moreover having examined the evidence ourselves, it was not unreasonable for the jury to have been troubled by what we shall describe as the practical difficulties of such sexual activity, including as it did, penile penetration, and the evidence which tended to confirm that the appellant lacked the necessary opportunity. However these problems did not arise in relation to the relatively fleeting forms of contact involved in the counts where convictions were returned, and the allegations made in these counts were supported, at any rate as to propensity, by evidence of the eye witness in count 1 and it remains a further striking feature that in this particular case, there could be no possible question of collusion between any of the complainants, and no possible question either of contamination in relation to either the evidence of JR or JB, and in reality, the possibility of contamination was excluded in relation to SC.

In these circumstances we are unable to accept the submission that the verdicts were logically inconsistent. The reality is that the verdicts reflected the requirement on the jury to reach different verdicts on each count, to approach each count separately, and not to convict of any count where they entertained a reasonable doubt.

Ms Munro, however argues that, even if the verdicts could not be described as inconsistent in the classical sense, in such an unusual and complex case where her arguments may not fit readily into any particular category or attract any particular label, the unsatisfactory nature of the evidence from these witnesses and the extent of proven false allegations made by them is such that this court should say that their evidence should have been rejected by the jury in its entirety, or, putting the same point a different way, that it was not open to the jury to select any part of that evidence and treat it as reliable; the verdicts are accordingly unsafe even if some logical basis for the verdicts can be discerned.

The Crown respond that all the arguments presented to this court were made to the jury, including those parts of the expert evidence which could have been construed as critical of the methodology used to obtain this evidence and the conduct of the interviews. Mr Dunkels argues that the appellant’s submissions are an invitation to this court to supplant the jury’s decision and replace it with its own judgment. In our judgment he is right. We firmly believe that, even in a case of difficulty and complexity such as this, the primacy of the jury in our criminal justice system has to be respected, particularly where matters of reliability and the assessment of witnesses lie at the centre of the case. This applies to the assessment of any competent witness, whether disabled or able-bodied. The appellant was represented skilfully and with determination by experienced and respected counsel. All the features drawn to our attention by her were drawn to the attention of the jury. The summing up directed the jury concisely, carefully and clearly. We do not detect the unfairness or lack of balance for which it is criticised. The verdicts were the responsibility of the jury. The ordinary principles governing criminal trials require both the judge and the jury to face the realities which can sometimes arise where special measures are put in place, but these arrangements do not alter the principle that the primacy of the jury should be respected. There is, in reality, no sensible basis on which we, sitting in this court as a constitution of three judges, should set aside the verdict reached by the jury on the grounds that they are unsafe. Accordingly the appeal against conviction is dismissed.

The sentence passed totalled 12½ years’ imprisonment made up as to 12 months on count 1, 18 months consecutive on count 2, 5 years consecutive on count 5, 5 years concurrent with count 5 on count six, 5 years consecutive on count 9 and 5 years concurrent with count 9 on count 10.

The judge, unsurprisingly, said that this case posed a sentencing problem the like of which he had never previously encountered. We are not surprised. Until recently there was no possibility of the court even considering the evidence of complainants with serious communication difficulties. Although we understand the judge’s reaction, he was not, in our view, right to say that the Sentencing Guidelines were not of a particular assistance to him. We have considered them.

We have taken account of the fact that the offences of which the appellant was convicted did not involve penetration. Therefore the recommended sentencing range was 26 weeks to 2 years with a starting point of 12 months. Although the judge in sentencing sought to reflect the vulnerability of the victims, the public revulsion for what he had done and the clear breach of trust there is a sense in which by definition all section 38 offences will include such factors. No other unambiguous aggravating factor can be discerned.

If the appellant had been convicted of the serious penetrative sexual offences of which, in the end, the jury acquitted him, a sentence in the region of, or indeed in excess, of the total sentence imposed by the judge may well have been appropriate. But the essential feature which we must bear in mind is that the convictions reflected non-penetrative touchings. However there were four different vulnerable victims of this abuse of trust, and subject to the totality principle, consecutive sentences were appropriate.

Properly reflecting the verdicts of the jury, these sentences must be quashed. In the absence of a guilty plea, the total sentence will be 4 years’ imprisonment. That will be made up as follows: