England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT FAMILY DIVISION
Mr Justice Hayden
ZC16C00911

		Royal Courts of Justice Strand, London, WC2A 2LL
		14 November 2019

Lord Justice Peter Jackson:

Summary

1. Three years ago, on a Sunday morning in November 2016, S, a 10-year-old girl, was found dead in her bedroom. Also in the home were her parents and her five siblings, including two older brothers. S died of strangulation and had suffered recent injuries to her genital area. The police investigation was deficient. No criminal charges have been brought. An inquest was opened and adjourned.
2. The local authority issued care proceedings to protect the siblings. It alleged that S had been sexually assaulted and killed but that it was not possible to say who among the parents and older brothers might be responsible. The family deny this. Its position has been that S's injuries and death must have been the result of an accident in which she fell from her bunk bed, entangled in some netting. It has also raised the possibility that S had been attacked by an intruder.
3. There have been two trials. In December 2017, Francis J found that the local authority had not proved its case and he dismissed the proceedings. The local authority appealed and the decision was overturned by this court in July 2018. The retrial came before Hayden J in early 2019. In a judgment given in June 2019, he found that S's mother had caused the genital injuries in the course of an attempt at female genital mutilation (FGM) that took place outside the home and that she had then strangled S in her bedroom after their return. The father, but not the brothers, had colluded to hide what had happened. The parents, supported by the brothers, now appeal.
4. Having heard the arguments, I am in no doubt that the appeal must succeed on the grounds that the decision was both wrong and procedurally unjust. The FGM finding was conjectural and unsupported by any real evidence. It had not been alleged, or even investigated, and therefore took the parties by surprise. It must be set aside, and as it is the foundation for the identification of the mother as having killed her daughter, that finding cannot stand either.
5. In these dire circumstances, the local authority invites us to substitute our own findings. I would decline that invitation. It would only be a proper course if there was just one realistic outcome and that is not the case here. As to whether a second retrial must be ordered, the family, supported by the Guardian, pleads that it need not and that the proceedings should now come to an end. With a heavy heart, I cannot agree. The circumstances are too serious, and a further hearing may yet provide an outcome that is of value one way or another to the surviving children. I would therefore remit the local authority's application to the family court for a retrial preceded by an early case management hearing in accordance with arrangements that have been made with the President of the Family Division.

The background

6. The proceedings concern five children – E (a boy who was 15 when S died), W (a boy then aged 13), X (a girl then aged 6), Y (a boy then aged 3) and Z (a boy then aged 6 weeks).
7. The parents were born in an overseas country where FGM is prevalent ('OC') and came to the UK as children. They married in 2000. There had been no previous concerns about the children, who appeared to be thriving. S's attendance and behaviour at school was, like that of her siblings, excellent. The entire proceedings revolve around the tragic events of a single weekend in November 2016.
8. At about 4 pm on the Saturday, while the father was at work, the mother and the children (except E, who was going to a party) drove to the home of the father's female cousin N, to whom the mother is close. N's five children, aged between 15 and 4, were there. Another 15 year-old girl was also present. The children played. S and the eldest cousin, a girl, went to the shops to pick up some food. Later in the evening, everyone had a full meal and at some point N's husband returned home. S is described as having been in high sprits throughout.
9. The mother and children left N's house at about 10 pm and went to the home of the father's aunt O, but she was not there. They returned to the family home, where they found E, who had come back from his party. The father returned from work shortly afterwards so that by 10.30 pm all eight family members were in. S, X and Y then went to bed in their second floor bedroom. The mother retired to her first floor bedroom with Z. The father and the older boys stayed downstairs watching television. E went to bed around 11.30 pm and the father and W followed at 12.30 to 1 am. At the time, shortly after the birth of Z, the father was sleeping in the other second floor bedroom with E and W. The mother was awake for much of the night with Z and she was frequently on the telephone to aunt O and others. Her final conversation with O lasted from 1.33 am to 2.23 am.
10. On the morning of the Sunday, S was found dead in her bedroom. She had occupied an upper bunk and X and Y the lower bunk. She was reported to have been found by W (then 13) with a length of decorative netting that had adorned her upper bunk bed, wrapped around her neck. Her father reported that he had removed her from the netting, but she was already dead. Her mother dialled 999 at 10.13 am and the police and ambulance service attended swiftly. During the course of the day, the police considered whether there might have been foul play, but came to the view that it had been a tragic accident in which S had been entangled in, and strangled by, the netting. No attempt was made in subsequent days to carry out a reconstruction and much potentially valuable information was lost.
11. The only account of an event in the night was one given by X (6), who was spoken to by police officers in the mother's presence. She said that she had heard a thud or thump and found that S had fallen out of bed in the night but that she hadn't told anyone because she was worried she would get told off for waking her parents up. The mother asked why she did not tell someone, and X replied that she had shouted, but no one came. X did not repeat this account under later questioning.
12. On the following Thursday, a post-mortem was carried out by the Home Office forensic pathologist Dr Cary and the paediatric pathologist Dr Marnerides. They found a mark around S's neck that was in their view indicative of ligature strangulation. They also found injury to the external and internal genitalia and the anorectal region. Their opinion was that S's death had not been an accident but a sexually-motivated homicide.
13. The day after the post-mortem, the father was arrested and questioned but released without charge. The local authority took proceedings leading to interim care orders. E and W moved to other family members, X and Y were placed in foster care, and Z remained with his mother.
14. On 6 December HM Coroner opened and suspended an inquest into S's death. The inquest has not yet been resumed.
15. On 13 December W was arrested at school and bailed after questioning.
16. On 16 December the court approved the return of X and Y to their mother under interim supervision orders. The mother and the three younger children thereafter lived with a maternal aunt, across the road from the family home.
17. On 9 March 2017 the court approved E and W returning to the family home under the care of their paternal grandfather.
18. On 7 June 2017 a decision was made that there would be no further police action in relation to the father or W.
19. In the family proceedings, expert medical evidence was obtained from, amongst others, Dr Cary and Dr Marnerides, Dr Leadbeatter (forensic pathologist) and Dr Lipetz (gynaecologist). There was agreement between them that death had probably occurred between about 11 pm and 2 am and was caused by pressure on the neck from a ligature, Dr Cary considering that it was more likely to have been caused by pulling than hanging. The most likely explanation was a deliberate act by a third party. Suicide or accident were possible but less likely. As to the genitalia, there was acute traumatic injury, likely to have occurred within 12 to 14 hours of death and to have been caused by digital penetration, though other forms of penetration could not be discounted. In the absence of a reconstruction showing how the genital injuries might have been caused by S falling onto an object (referred to as a straddling injury), Dr Lipetz considered this explanation unlikely. There was no unanimity about the presence of injury to the anorectal region and the local authority did not pursue a finding in that respect.
20. In October 2017 an experts meeting between the four experts took place. There was a substantial measure of agreement. With some degrees of nuance, their common opinion was that (1) an accident was less likely than a deliberate act by S or a third party, (2) suicide, though possible, was unlikely and (3) the death should be investigated as a homicide.
21. In November 2017 Francis J conducted a fact-finding hearing lasting 15 days. In a judgment handed down on 22 December 2017, he dismissed the proceedings. He described serious deficiencies in the police investigation and in police disclosure. He found that the local authority had not proved on the balance of probabilities that S's injuries were inflicted by anyone as opposed to being accidental or, though unlikely, self-inflicted.
22. In January 2018 the family was reunited in the family home, where they have remained.
23. The local authority appealed, and on 25 July 2018 its appeal was allowed: see A (Children) [2018] EWCA Civ 1718. King LJ, in a judgment with which the Senior President of Tribunals and Sales LJ agreed, concluded that the judge had not correctly approached the burden of proof. He had not looked at the whole picture. He had not effectively analysed the expert evidence about the genital injuries or taken them into account when considering the manner of death. Traumatic though it would be for the family, the risks to the children from the so far unexplained death of their sister made a retrial necessary.

The retrial

24. The circumstances surrounding this appeal are so unusual that I need to give a fuller than normal description of the course of the proceedings, the judgment and the submissions we received.
25. The retrial before Hayden J lasted for 18 days, beginning on 21 January 2019 and ending on 22 March. The paperwork filled 18 files. There were transcripts of the first trial and the medical, police and lay witnesses again gave evidence, with additional evidence from Dr Kolar (forensic pathologist) and Dr Malcolmson (paediatric and perinatal pathologist). On this occasion, E gave evidence. There had been a second meeting of the experts, which came to much the same conclusions as the first.
26. In preparation for the retrial, the local authority had pleaded its threshold allegations with precision. These included that:

(1) The neck injuries and deaths were caused by ligature strangulation.

(2) The genital injuries were caused by blunt force trauma which included an element of penetration either digitally or by an object.

(3) The genital injuries were caused at the same time as or within a few hours of death, and in any event following S's return home at about 10.00 pm on Saturday.

(4) All the injuries were inflicted by the mother and/or the father and/or E and/or W.

(5) One or more of those four individuals must have been aware of the manner of S's death.

(6) One or more of the four had conspired to cover up what happened in order to protect the perpetrator and had dissuaded the younger children from talking.

27. The issue of FGM had played no part in the first trial and was not considered at the second experts meeting or in the expert evidence at the rehearing. Dr Lipetz had advised that S had not suffered FGM. A search of the police disclosure material, carried out since the judgment, revealed a single record in January 2017, where Dr Cary was recorded as advising that on initial examination there was no suggestion of an attempt at FGM and the area of bleeding was not consistent with such an attempt.
28. During the retrial the issue of FGM featured only briefly in the evidence.
29. On 3 February 2019, the foster carer for X and Y gave evidence of a conversation she had overheard shortly after the children had been placed with her. X (3) was playing with toys when he stopped, looked puzzled, and said, "I don't understand, if a girl don't have a 'barelly'… how can they be circumcised?" To this, his sister Y (6) said "You don't know what you're saying", and he stopped talking.
30. On 8 February the mother was cross-examined by Mr Goodwin QC for the local authority:

"MR GOODWIN: Have you ever considered that S, in the past, should have been circumcised?

A No. My mum did that to me and I know how pain and terrible it is, so I would never wish for my child to go through what I go through.

Q Your mum did that to you?

A Yes.

Q Has circumcision ever been discussed with any of the children?

A No. The boys do it but we don't do the girls.

MR JUSTICE HAYDEN: Why is it right for the boys and not for the girls?

A Because I think ever - most of the country, they do the boys. In our country and some other parts of Africa, they do the girls as well.

Q Right?

A But they --

Q You do not think it is right for the girls?

A No. It's horrible, especially… There's two ways to do it. The one I done it is horrible.

Q I understand all about it and I agree it is an abomination, but boys have no say in their circumcisions either, so why is that right?

A It is a Sunnah from prophet [indecipherable] Saleh and boys must have, but there is no way in the Quran girls should have it. It's just the culture.

Q It is because of the Quran.

A Yes because of the culture just go through – and women been through that (inaudible) but the man must have. Every Muslim man.

MR GOODWIN: But there is an incredibly high rate of FGM (female genital mutilation) in [OC]?

MR JUSTICE HAYDEN: I think that might be a distraction in this case.

MR GOODWIN: Right, my Lord. I will not pursue that. (To the witness) Nothing that we should attach to [Y]'s comment to the foster carer about "how can a girl be circumcised if she doesn't have a penis"?

A I think Joe [the social worker] tell me and talked to me about it, [Y], so [Z], when he was circumcised, he had a little (inaudible) when he was cutting Y and Y, his friend (?) was not long ago, was still there. So he talked to the foster man –

MR JUSTICE HAYDEN: How old was he ? What age?

A [Y]?

Q What age are the boys circumcised?

A Back home they do 7, 8, but in here we do it as soon as we can. For example, after [Z] --

Q As babies?

A I done [Z] under the month.

Q In a month?

A Under the month, yes. I done after one year and-half. So, it depends how you just do it.

Q I understand."

This exchange occupied little more than one page of the thousand pages of the trial transcript. The reference is to the mother having taken Z to be circumcised by a GP at the age of about four weeks, so in early November 2016.

31. The evidence concluded on 18 February. The last witness was E, who was called to explain certain material on his phone.
32. On 26 February, counsel spoke to their extensive written closing submissions, which did not mention FGM. However, the judge himself introduced the matter during Mr Goodwin's oral submissions. He queried the assumption that the genital injuries were sexually motivated and drew attention to Y's conversation with the foster carer. Mr Goodwin reminded the judge that he had indicated that he did not find the line of questioning helpful, and the judge replied that if the local authority wanted to pursue the matter, the witness could be recalled, and that he (the judge) now thought it was important. The submissions continued:

"MR GOODWIN: Yes. Well, because, as things stand, my Lord, we are where we are, in terms of the evidence.

MR JUSTICE HAYDEN: Yes. I mean, that is the difficulty when there is a running assumption that a genital injury is sexually motivated, and the judge only really gets a chance to evaluate all the options when he's got to the end of the evidence and, particularly in this case, [E]'s evidence.

MR GOODWIN: Your lordship has far more extensive experience of FGM-type cases than I do, I certainly accept that, but I am struggling at the moment to understand what the link would be in terms of the pathology and the type of injury and the FGM thesis. There is no evidence that there was any cutting going along, going on. There is no other evidence of FGM, either picked up from the text messages sent by the family or from any other relatives. So I try to be cautious not to speculate about the significance that that episode might have.

MR JUSTICE HAYDEN: Tell me about the day, what you say about the day before [S] died, the mother and the children travelled, travelling some journeys that day."

33. Mr Goodwin then reviewed the evidence about the family's movements on the Saturday, but neither he nor anyone else attached any significance to the judge's last isolated remark and to what can perhaps be seen in hindsight as a juxtaposition of FGM and the events of that day. Mr Goodwin made clear that the local authority did not seek to draw other family members into a conspiracy or to allege that there was a plan to injure S before the return to the family home. He also cautioned the judge against attaching too much weight to the conversation with the foster parent. In response, the judge described that evidence as "one of the building blocks, this is terra firma." Further exchanges on this point included the following, in somewhat telescoped form:

"MR JUSTICE HAYDEN: And that says what, exactly, because he's talking about girls being circumcised and it is confusing him?

MR GOODWIN: That is correct. I mean, he is asking, if a girl does not have a willy-- how can they be circumcised? So query whether he has heard something about that… query whether he is asking himself –

MR JUSTICE HAYDEN: I draw the inference, to use your phrase, that he is troubled about female genital mutilation… it is on his mind for some reason.

MR GOODWIN: Well, I suppose the difficulty, my Lord, is whether or not it is on his mind for some reason because the question mark has been generated by something he has heard from a parent or whether it is something internal. …

MR JUSTICE HAYDEN: Very common, FGM, in [OC], you say?

MR GOODWIN: Well, a UNICEF report does say there is a prevalence in this age group of 95 per cent.

MR JUSTICE HAYDEN: Yes, 95/98 per cent.

MR GOODWIN: I do not want to introduce any controversial expert evidence on that front.

MR JUSTICE HAYDEN: Well let us -- We will see how we deal with the overall picture in due course."

34. As Mr Goodwin's submissions came to an end, the judge once again returned to the FGM issue:

"MR JUSTICE HAYDEN: I do not want you to go without addressing the point that I, as I am fully prepared to accept, did not encourage any further pursuit of the FGM issue in cross-examination and returned to it only after the whole of the evidence. Now, if you want to make that good and you say it is in anyway significant, I struggle to see how it is likely to go beyond a screaming disavowal, which you have already got, but if you think it needs to be pursued further than I really need to know that.

MR GOODWIN: I do not invite the court to reopen the evidence, if that is what your Lordship is alluding to."

35. Ms Ball QC then made submissions on behalf of the mother. After a certain time she came to FGM, saying that she did not know what the judge's position was, but that her client had given clear evidence about her opposition to it. The judge himself described it as "an eloquent and forceful disavowal of FGM." There was then an exchange in which Ms Ball suggested that Y's comment in the foster home might very well have been prompted by the fact that Z had recently been circumcised. The judge described this as "a very bold submission". Ms Ball continued to address the circumstances of S's death and the possibility that it was accidental. After some time, this exchange took place (again somewhat telescoped):

"MS BALL: … are the local authority really suggesting that it is a real possibility that this woman, this loving, responsible, according to all accounts, woman went upstairs, with or without the baby, and strangled or hanged that daughter, and raped her or sexually interfered with her?

MR JUSTICE HAYDEN: … Not necessarily rape or sexual.

MS BALL: Well, my Lord, what is it then? Are we going down --

MR JUSTICE HAYDEN: I do not know.

MS BALL: -- the line of FGM, because -- if we are, then I think more evidence is going to be needed.

MR JUSTICE HAYDEN: -- You will have to tell me what evidence you think is going to be needed rather than just a generalised assertion.

MS BALL: Well, medical evidence …

MR JUSTICE HAYDEN: Nobody has suggested that she has been cut.

…

MS BALL: Well, my Lord, it has not been put to my client --

MR JUSTICE HAYDEN: And I consider -- the entire wide canvas -- it has, she has been asked about FGM …

MS BALL: She has, but it had never been suggested that she has in some way interfered with her daughter or allowed somebody else to interfere with her daughter in that way, leading up to, for example, FGM

…

MR JUSTICE HAYDEN: It has not been developed, I agree -- as a theory.

…

MS BALL: … but none of us have even addressed it… We have not asked the doctors about it. I find it surprising that my Lord –

MR JUSTICE HAYDEN: Well it is surprising in so long it has not been addressed, but it emerges from the evidence as it has unfolded in what is always particularly in this case, a dynamic process.

MS BALL: Well, my Lord, the local authority did not appear to be going down that line. I mean I'm not saying that in any way prevents you from going down that line, but they were not, they have not developed it. They have not cross-examined anybody in this case to suggest that they were actually party to some preliminary, for example, activity which may have led to that, and that is wrong. That is quite wrong that they have not had the opportunity to say "No of course we didn't. No we don't agree with it."

MR JUSTICE HAYDEN: Well they have said that, have they not?

…

MS BALL: But it was never suggested that they, that they had done this to their daughter.

MR JUSTICE HAYDEN: Well, what more can they say other than disavow it …

MS BALL: Well, my Lord, if that is truly a line you are going down then I must…

MR JUSTICE HAYDEN: Well, I can assure you, Ms Ball, it is.

MS BALL: Well my Lord then I think I do need to call my client for you to hear what she says.

MR JUSTICE HAYDEN: Well, I think you should… I am not going to stop you at all if you -- I signalled that it was part of the range of possibilities that I had to consider, and I consider that the genital injuries may not be sexually motivated.

MS BALL: No, we say they are not

MR JUSTICE HAYDEN: … So if it is not sexual, what is it

MS BALL: It is an accident, my Lord

MR JUSTICE HAYDEN: Well, that is your theory."

36. Mr Bagchi for the father, added his voice to those arguing that the issue of FGM was not supported by the evidence. The judge asked:

"MR JUSTICE HAYDEN: What are the range of options?

MR BAGCHI: Accident.

MR JUSTICE HAYDEN: There is accident. The evidence in relation to that is tissue thin. What are the other options? I do not believe in… having sat through the entirety of the evidence and read all the papers, to keep you, sit here like an open fridge until you get the judgment, I like to tell you where my money is going so that you can exchange with it. That is a process that I think is essentially fair."

37. After some further submissions, there was a brief adjournment for the parties to take stock. Upon returning, Ms Ball indicated that he would wish to recall her client and call some other lay witnesses. She then turned to the question of expert advice and referred to Re B and G (Children) (No 2) [2015] EWFC 3, where Sir James Munby P had received substantial evidence from specialists in the types and methods of male and female circumcision. Ms Ball requested permission to approach one of those witnesses, an expert in paediatric FGM cases. The judge was unwilling to entertain argument about this, saying that FGM had not taken place in this case. Instead, he identified Dr Lipetz as the expert and gave permission for a single question to be put to her. Ms Ball sought to preserve her position, but the judge made clear that he had determined the matter and that wider expert evidence was "not remotely necessary". The hearing was then adjourned for three weeks.
38. Further evidence was then filed by seven lay witnesses.

(1) The mother made a five-page statement describing her own experience of FGM and opposition to it in relation to her own children. One of her brothers has returned to OC. None of his four daughters has been subjected to FGM.

(2) Aunt O made a statement describing her own experience of and opposition to FGM. Neither of her daughters, aged 26 and 13, has been subjected to FGM.

(3) Cousin N, who had not previously made a statement in the family proceedings, filed a statement describing her experience of and opposition to FGM. None of her three daughters, aged 18, 13 and 7, has been subjected to FGM. The statement ended with two paragraphs describing the events of the Saturday and S's normal, happy demeanour that day.

(4) N's 18-year old daughter (the cousin who had accompanied S on the visit to the shops) filed a confirmatory statement.

(5) The mother's older sister filed a statement describing her own experience of and opposition to FGM. Her daughter, aged 10, has not been subjected to FGM.

(6) The mother's younger brother, a social worker who the judge described as having a very dominant role in the household, gave his opinion that the UK OC community is very much against FGM and that he himself had taken part in anti-FGM workshops. None of his four daughters has been subjected to FGM.

(7) The father filed a short statement confirming the contents of all these statements.

39. The effect of the further evidence was that all adult female members of the family had been subjected to FGM in OC, but that none of their twenty female children in the UK had experienced the practice. Generally, the mother and other witnesses described a change in culture and attitude since they were children themselves.
40. Dr Lipetz was asked two questions in writing:

"(1) To what extent were the genital injuries consistent with (a) an attempt at FGM or (b) an act preparatory to FGM or (c) a coercive genital examination preparatory to FGM?

(2) Do you have experience of examining children where there has been suspected FGM?"

Her replies were:

"(1) I am not able to answer as I have no experience of preparation of FGM or the act of cutting other than having heard women's accounts of their experiences and two cutters' accounts of their techniques. The concept of a preparatory examination prior to cutting is not something I have heard of from women who have been cut.

(2) No: I have reviewed examinations (digital images, still and video) of prepubertal children at peer review to confirm they have normal genitalia and have not had any that have had FGM. The youngest child I have seen with suspected FGM was 12, post pubertal and had normal anatomy.

Over the last 23 years I have examined many women who have been cut, all post pubertal. This has been in general gynaecology, contraception and obstetric settings (delivering babies). The women who have been cut have been from the Yemen, Sudan, Somalia, Eritrea, Liberia, Sierra Leone and Tanzania."

41. The trial resumed on 20 March 2019. The only witnesses called to give evidence were the mother and her younger brother. The mother's evidence runs to 35 pages of transcript. She described in detail what she herself had undergone as a child, much of it in response to questions from the judge. At no point did anyone ask her about events at N's home. The only reference to S was a single question to the mother from Mr Goodwin:

" Q Did [S]'s death have anything to do with FGM?

A Not at all, hundred percent no."

42. The mother's brother also gave evidence, but none of the other witnesses was required to do so, the judge having confirmed that a witness statement filed and not challenged stands as evidence.
43. Final submissions were made on 21 March. Mr Goodwin made clear at the outset that the local authority's case was unchanged by the further evidence. These exchanges followed (again telescoped):

"MR GOODWIN: … The evidence she gave and the evidence your lordship has on paper, that none of the children of this extended family have been cut. There is no contrary evidence in relation to that.

MR JUSTICE HAYDEN: There is no evidence other than assertion, is there?

MR GOODWIN: My Lord, that is right. We do not consider that there is sufficient an evidential basis to conclude that the context for S's injuries and death is related to FGM. … In our submission there is insufficient evidence to make solid conclusions about the likely motivation.

…

MR JUSTICE HAYDEN: Are you actually arguing against this, or are you saying it is not the way the local authority puts its case? It can see that there are features of the evidence which leads in that direction, but it does [not] feel confident ultimately that is going to advance that case.

MR GOODWIN: It is the latter.

MR JUSTICE HAYDEN: … And ultimately of course it is a matter for the judge.

…

MR GOODWIN: Of course, my Lord. … But we are mindful… there is no expert evidence in support of the FGM thesis.

MR JUSTICE HAYDEN: It is not a question of what evidence there is not. It is a question of what evidence there is and whether it is not enough.

…

MR GOODWIN: But we note of course that there is no evidence that FGM was practised on S, and insofar as the injuries that we see are indicative of examination, of preparatory steps prior to FGM that is a possibility. It has been argued it is no more than a possibility and therefore ultimately does not assist the local authority in defining the motivation behind what happened to her.

MR JUSTICE HAYDEN: Very well.

MR GOODWIN: So, my Lord, that is our position."

44. That concluded the local authority's submissions. The only other relevant exchanges came in the course of Ms Ball's submissions. She addressed the family evidence about FGM generally. She returned to the inherent improbability of the mother assaulting S in this way after returning home after a happy day out. To this the judge said that it should be borne in mind that the genital injuries could have been caused up to 72 hours before death, but Ms Ball noted that there was no evidence that S was unhappy or in pain when she came home that evening.
45. Judgment was reserved.

The judgment

46. In his judgment of 7 June, the judge set out the issues and summarised the history. He began his analysis with reference to the medical evidence, which he summarised in this way:

(1) The neck injury:

"… the pathologists considered that the preponderant evidence, tempered by appropriate caution, was that the injury to the neck was in consequence of third-party ligature strangulation."

(2) The genital injuries:

"… 'blunt trauma and penetration or attempted penetration by an object, finger or a penis'… represented the reasoned consensus."

The judge observed that the presence of trauma did not necessarily equate with sexual assault.

(3) Timing: the genital injuries occurred no more than 12-24 hours before death [in parenthesis, the figure of 72 hours mentioned by the judge in closing submissions had at one stage been mentioned by Dr Lipetz]. Death probably occurred in the early hours of the morning.

47. The judge rightly emphasised that the correct approach was to evaluate the evidence cumulatively, weighing expert opinion alongside other features of the evidence. He then turned to the lay evidence and I trace the remainder of the judgment in the sequence in which it appears.
48. The judge recorded the parties' cases, as already recorded. On behalf of W, the possibility of assault by an intruder was also raised, but the judge discarded this as unrealistic. In doing so, he said this:

"73. … Fact finding hearings have an investigative, dynamic complexion to them. The consequences of adverse findings against parents or carers may be profound and life changing. It is for these reasons judges frequently permit great latitude in the presentation of these difficult cases. That does not extend to the unarguable. Even in this investigative, non-adversarial, sui generis jurisdiction there must be parameters."

49. He recorded that the mother was awake for most of the night. The house was small and any noise would quickly be heard.
50. The judge then addressed the hypothesis that there had been a fatal accident in which S had sustained the genital injuries by physical contact with some part of the bunk bed as she accidentally strangled herself and fell off the top bunk. He reminded himself that there was no burden of proof on the family. He considered the evidence of Dr Lipetz that 'straddle' injuries were rare and not characterised by penetrative injury, and continued:

"83. … I have tried to envisage how a penetrative injury could have occurred by the mechanism Ms Ball suggests. I have been unable to do so. It strikes me, largely as a matter of common sense, that this is highly improbable. I do not find it difficult to see why Dr Lipetz discounts penetration in straddle injuries so unequivocally and it seems equally as unlikely in the mechanism contemplated here.

84. … I note also that there were no relevant injuries other than in the genital region.

51. The judge, again rightly, reminded himself that the view of the experts was not bound to prevail and that the court must factor in the possibility that after all the evidence is considered it may be insufficient to establish a probable cause of injury or death. He then drew his conclusions as to the causation:

"86. Here however, the individual medical opinions have been scrupulously tested, in the way that I have sought to demonstrate above, by reference to a wide panoply of other professional opinion and expertise. The experts involved have, without exception, displayed an eagerness to engage open-mindedly in enquiry by highly experienced counsel, over a wide range of possible hypotheses. Their reasoning has been, in my assessment, entirely free from dogma, nor has it been characterised by any defence of amour propre. I have, ultimately, for all the reasons set out in my summary above, come to the clear conclusion that S died in consequence of strangulation inflicted by another person. In addition, I am, with very little difficulty, satisfied that the medical evidence establishes that the genital injuries were sustained in consequence of blunt trauma and penetration, caused by an object, finger or a penis.

87. Having determined, by separate analysis, that both the strangulation and the genital injuries were inflicted by trauma caused by another person, I permit myself, at this stage and I emphasise only at this stage, to conclude that my findings in respect of each serve to reinforce the other."

52. The judge then turned to consider whether it was possible to identify a perpetrator of the assaults. He directed himself with reference to Re B (Uncertain Perpetrator) [2019] EWCA Civ 575, which confirms that the court will seek to identify a perpetrator on the balance of probabilities where possible, and will only consider making a 'pool' finding where it is not.
53. He then returned to consider the significance of the genital injuries:

"Whilst I am entirely satisfied that there has been both internal and external blunt trauma genital injury, I can find no evidence, at all, that this was a sexual assault. Accordingly, I also find no evidence that the subsequent strangulation was sexually motivated."

54. He then considered evidence voluntarily given by E. He accepted his account of the sequence of events on the days in question and his evidence that he had slept in his bed throughout the night until the following morning. He noted how cramped the space was in the bedroom occupied by the father and the boys and said that it was impossible to see how anyone could have left it in the night without disturbing the others.
55. The judge also accepted as reliable the account given by all family members that S had showed no sign of having suffered genital injuries before she returned home and that she had kissed her father goodnight before she went to bed. She had showed no emotional affect or sign of physical pain, as would have been expected had she been injured previously. The judge described this as "a very striking feature of the evidence".
56. At that point, however, the judgment began to depart from the parties' pleaded cases. He accepted detailed evidence given by both parents about the childcare arrangements at home and the mother's evidence that Z's first few weeks had been hard for her. He noted that the visit to N's house was the first time that she had left the house since the birth of the baby. He then said this:

"107. … According to M she and the younger children stayed at N's until quite late. I was told that a lot of food had been prepared and that the day was full of laughter and joking. In contrast to the detailed evidence I have recounted above I have found the account of that afternoon to be vague, generalised and ultimately unconvincing.

108. In the light of my findings, both in relation to the family's accounts and the medical evidence, it is likely that the genital injuries occurred during the late afternoon. Mr Twomey [for W] has explored the possibility that S may have been sexually assaulted during the 40 minutes that she left the house with her cousin to buy Shawarma, a type of bread. Nobody has suggested that S was distressed on her return or made any complaint at any stage. I discount it as entirely improbable that S was sexually assaulted on her visit to the shop."

57. The judge recorded that after leaving N's home, the mother and children had driven to the home of father's aunt O, who in the event was not in, but that in an earlier telephone call with the father she had not told him of her intention to do this. The judge considered the mother's explanation (that she thought the father would be displeased with her for being out so late) as unconvincing when the mother was the more dominant of the two. He considered that there was "some more sinister reason why she lied about her whereabouts to her husband." He also described as entirely unconvincing the mother's explanation that her long night-time phone call to aunt O was normal behaviour. He then considered the mother's account of her wakefulness with Z during the night and said this:

"112. What is significant about this is that the mother was awake for most of the night in this modern, modestly sized house when her daughter was, as I have found, strangled to death. It is equally clear that M had been with S the entire day, excluding the short visit to the shop."

58. He then recorded the evidence of the foster carer, who he found an impressive witness, about statements made by Y in foster care, and in particular the statement recorded above. He noted that the foster carer had formed the strong view that X saw it as her role not to let Y open up about S's death.
59. The judge next turned to the mother's evidence about her experience of and attitude towards FGM. He set out her detailed account of her own experience, describing it as harrowing and graphic. He recorded her unequivocal condemnation of the practice. Having done so, he observed:

"126. The force of all this impresses the Local Authority as sincere. I have no doubt that aspects of it are true. This said, truth and verisimilitude are, self-evidently, conceptually different."

60. The judge referred to the evidence of the mother's younger brother, and to the fact that the mother had arranged for the male children to be circumcised, something that the mother had justified on religious grounds. Of this, he said:

"Her rationalization of the practice, in this context, as an article of faith, troubled me and served, in my assessment, to weaken her articulate disavowal of FGM."

61. The judge then set out ways in which he found family accounts of the events of the morning to be disjointed and difficult to reconcile. The father's credibility was compromised by changes in his account. The fact that the mother did not go upstairs when S's body was found was noted, as was her explanation that she was too weak and distressed to do so.
62. The judge quoted the local authority's closing submission in relation to perpetration, which was in these terms:

"Identifying a perpetrator is particularly challenging here because, first, there is no firm evidence about the motive for the killing and, second, there is no clarity about whether the person responsible for her sexual assault also killed her. Was she sexually assaulted by one of her brothers then killed by her parents to avoid bringing shame on the family? Was she sexually assaulted then either deliberately killed by the same person or accidentally killed during an act of ligature restraint? We do not know the answers to these questions and would be speculating if we sought to base a decision on perpetration on either hypothesis."

63. The judge responded to this submission:

"142. … I have taken care to emphasise the many positives in this family. This is an important feature of the broader canvas of the evidence and I have it in mind when I consider the hypothesis that S may, in this essentially loving household, have been a victim of a sexual assault by one family number and killed by another. It is inherently improbable. The speculation that the killing might be motivated by the 'shame' of the sexual assault has absolutely no root at all in the evidence."

64. In the remaining paragraphs of his judgment, the judge moved to his conclusion, from which it is necessary to quote extensively:

"144. … From my summary of the oral evidence from the documents the following key features required to be identified:

i) S was in M's care throughout the 24 hours prior to her death;

ii) On her own account M was asleep for, at most, 1½ hours in this period;

iii) Excluding the physical evidence of strangulation, there were no other relevant injuries identified to S, beyond those to the genital area as described;

iv) There is no extrinsic evidence suggesting that a sexually motivated assault took place in the night in S's bedroom where two other siblings were sleeping. Nobody reports any noise or disturbance, notwithstanding their close proximity and, of course, the fact that M was awake for most of the night. What is postulated therefore is an apparently silent and brutal sexual assault (having regard to the extent of the genital injuries) followed by strangulation. On this factual matrix the explanation is simply not credible;

v) There is no physical evidence of a struggle within the bedroom, generally and none, particularly, relating to S's body or clothing;

vi) Notwithstanding that this was the first time M had been out of the house, at least socially, since the birth of the baby, who was a poor sleeper, she was out until very late in the evening;

vii) M admits that she lied to F about her visit to O. For the reasons stated above I reject her rationalisation of this lie.

145. It follows from this that the genital injuries were unlikely to have been inflicted after S returned home but, as the forensic evidence establishes, they would have been sustained during that day. There is, accordingly, no realistic opportunity for S to have sustained the genital injuries at a time when she was not in M's care. …"

65. The judge accepted the evidence of E and the father about the day and the evening but did not accept the father's account of the events of the morning:

"147. … On any view that there was a period of 40 minutes unaccounted for between F getting up and the telephone call to the emergency services."

He also considered that there was an apparent lack of curiosity on the part of the family about how S might have died.

66. The judge then stated:

"150. Having come to the clear view that the genital injuries were inflicted during the day and having entirely discounted a stranger assault occurring during the short visit to the shops, I must logically conclude that they were sustained whilst S was in M's care and/or in the company of the other women who attended N's home that day. Whilst I must not discount, entirely, the possibility of a sexually motivated assault in these circumstances, I consider it to be highly unlikely."

67. He said that he considered the foster carer's evidence relating to Y's concern about female circumcision to be significant. He referred to her "analysis" as being "well reasoned, objective and insightful."
68. The judge then came to his final conclusions:

"153. I find M's account of her voluble conversation with O between 1.33 am to 2.33 am [sic] as a 'normal' event to be unlikely. She had not been able to speak with O as she had hoped earlier in the evening and had lied about her unsuccessful visit … to her husband. I have been told that an older generation is more likely to continue the practice of FGM and I note that O falls into this category of respected older woman.

154. Having accepted the evidence of F and E to the degree that I have, I consider both were likely to have been asleep when S was killed. I do not consider that, on a proper construction, there is any evidence to suggest that there is a real likelihood or a real possibility that W could have been responsible for his sister's injuries and subsequent death. Though both the Local Authority and the Guardian contend that he should remain in the pool of perpetrators, I cannot identify any evidential basis upon which they invite me to do so. Thus, the only person in the household likely to have been awake when S was killed was her mother. Drawing all the above strands of evidence together I consider that it was she who, on the balance of probabilities, caused the genital injuries earlier in the day and strangled S during the night.

155. Mr Goodwin describes M's evidence variously as 'disingenuous', 'not to be trusted as truthful' and to be treated with 'great caution'. Nonetheless, I am told the Local Authority evaluate her denouncement of the practice of FGM as convincing. I am not prepared to make this exception to my assessment of her damaged credibility. Whilst her criticism of FGM was articulate, I was not persuaded of its authenticity. Following the discovery of S's body M told me she was too distressed and weak to go upstairs to see her daughter for herself. Certainly, when the police and paramedics were present she made no attempt to do so. She told me that she accepted what F had seen. I consider that she had no need to go upstairs to find out what had happened - she already knew.

156. I reiterate, every single adult female member of this family has been subjected to the abhorrent practice of FGM. I note that E described his family as being towards the 'traditional' end of the cultural spectrum. Despite M's denial, I consider that the genital injuries are more likely to be in consequence of some failed attempt at FGM. By this I mean, the genital injuries were sustained in a manner which was not sexually motivated. It was either an intimate investigation by way of preparation or, more likely, an actual attempt at FGM, which S was successfully able to resist. I consider the latter to be more likely because this assault is, for all the reasons I have set out, intrinsically linked to the strangulation that followed. I am not prepared to speculate about the actual circumstances. S made no complaint of what were quite significant genital injuries. Nor, I am told, did she exhibit any distress at home later that evening. Had it been a sexual assault. I am convinced she would have done."

157. I am not prepared to draw any inferences as to what M was speaking to O about during her lengthy telephone call. On the forensic evidence, S's death is likely to have occurred at some point after the conclusion of that conversation. The two may or may not be linked. Neither am I prepared to draw conclusions as to why M caused her daughter's death. What is evident is that M had not been sleeping properly since her baby was born, a period of some 6 weeks. She may not have been thinking clearly. It may be that she was angry with S or that she felt her family had been dishonoured by S in some way. What is clear, is that from the discovery of S's body this family has closed itself off to any investigation and has been determined that the full facts should not come to light. The collusion of silence is between the parents, the children are, to differing degrees, caught up in it."

69. The underlined passage was added after the parties had asked for clarification of the finding that there had been "some failed attempt at FGM".
70. On 14 June 2019, the judgment was formally handed down at a hearing convened for the making of orders. Ms Ball sought clarification of the finding at [150] that the genital injuries were sustained "whilst S was in the care of her mother and/or in the company of the other women who attended N's home that day". This, she asserted, amounted to adverse findings against third parties who had not been heard. The judge stated that the finding was against the mother and that he had made no findings against "the other women".
71. The order made on 14 June records these findings of fact:

(1) The injuries to S's genitalia and neck were deliberately inflicted;

(2) The genital injuries were caused by some act preparatory to or due to a failed attempt at female genital mutilation;

(3) The neck injuries caused S's death.

(4) The perpetrator of S's injuries was her mother.

(5) The mother and father colluded to hide the cause and the perpetrator of S's injuries.

(6) No findings were made against either E or W and in particular no findings that they had inflicted or colluded in the infliction of S's injuries, or colluded to hide the cause or the perpetrator.

Most of the argument on this appeal has centred on the second of these findings and its impact upon the first and fourth findings.

72. The judge made interim care orders in respect of X, Y and Z on the basis that they could remain in the family home if their mother moved out, something she was prepared to do under protest and subject to her intention to appeal. He also made a female genital mutilation protection order in respect of X. He granted a short stay of the interim care orders and made a reporting restriction order; these were extended by this court. On 3 July 2019, I gave permission to appeal.

The grounds of appeal

73. The grounds of appeal can be distilled in this way:

(1) The judge was wrong to find that the genital injuries had been caused in connection with FGM, and not at the same time as the neck injuries.

(2) The finding about the genital injuries being caused at N's home was unfair to the mother, who had not been put on notice of the possibility of such a finding. It was also unfair to N and to any others who were present, as they had not been given an opportunity to be heard.

(3) The judge was wrong to find that S had been strangled by another person, or that the mother was responsible.

(4) Inadequate consideration was given to the possibility that all the injuries were caused by accident.

(5) The judge was wrong to find that the parents had colluded to prevent the circumstances of S's death coming to light.

(6) He failed to take into account significant evidence undermining each of his conclusions.

(7) Alternatively, he should have found that the evidence was insufficient to discharge the burden of proof.

The submissions of the family

74. The parents and the brothers make common cause. They argue that, despite his self-direction, the judge wholly failed to look at the totality of the evidence. Instead he developed a theory of his own and strained to fit the facts of the case into it.
75. As to the evidential basis for and procedural fairness of the FGM finding:

(1) It was an extraordinary finding that had not been sought by the local authority, even after the judge showed an interest in FGM. It could only have arisen from an erroneous assessment of the evidence and/or an unfair process.

(2) There was no supportive medical evidence:

S had not been subjected to FGM.

There was no medical or other evidence that her injuries might be the result of a preparatory act or a failed attempt.

The injuries bore no distinctive stamp of an attempt at FGM (see Dr Cary's view).

There were no accompanying injuries of the kind that might have been caused by a struggle.

(3) There was a lack of appropriate expert evidence:

The judge only allowed a further question to be asked of Dr Lipetz, who was not qualified to advise on whether these injuries might be associated with attempted FGM or whether there is a practice of making preparatory examinations.

The judge refused to permit an expert on FGM to be instructed to advise on whether such a practice exists.

There is no literature describing it.

The other experts were not asked about the judge's theory.

(4) The judge did not show his hand. If he was thinking that an assault may have occurred at N's home, he should have been explicit about that so that the parties could address it. It had never occurred to the parties that this finding would be made.

(5) As it was, the evidence was incomplete:

The police had not investigated events at N's home.

Because FGM was not pursued by any party and arose at a late stage, questions were not asked of earlier witnesses.

The focus of the additional statements was almost entirely on FGM generally, and not on the events at N's home.

It was never even put to the mother that she had caused or allowed S to be injured at N's home.

Nor was it put to N.

There were a number of older children at N's home that day, who would have been well-placed to know whether S had been assaulted or not but, because it was only in the judgment that it was suggested that an event had taken place there, they were not asked.

(6) In particular, the judge did not consider that his finding meant that W (then 13), whose evidence about his movements was accepted, was present throughout the time that he found S to have suffered the genital injuries. W had made a statement containing a paragraph that described the normal events of the day. He would have been a critical witness in relation to any FGM event that might have occurred at N's home, but he was never asked.

(7) The judge ignored reliable accounts of S's demeanour throughout the day.

(8) Reasons relied upon by the judge to discount a sexual assault during the day apply equally to an FGM-type event. No pain or distress was recorded from what would have been an intensely painful experience. (Dr Lipetz advised that an intimate examination of a child of this age is so painful that it is conducted under anaesthetic.) Yet S ate a normal meal and was described as being happy and lively, both at N's home and on return to the family home.

(9) The judge had no proper basis for saying that the mother's views of male circumcision, as being based on religious obligation, weakened her disavowal of FGM.

(10) The judge cherry-picked the evidence about the family history:

While he gave significant weight to the fact that all adult females had been subjected to FGM, he ignored the unchallenged evidence that

This had happened to them at between 3 and 7 years of age.

None of the twenty female children of the adult family members had been subject to FGM.

All witnesses expressed abhorrence of the practice, and the mother's brother had campaigned against it.

None of the female witnesses has any criminal record or propensity to act unlawfully: FGM has been a criminal offence since 2003.

No professional had ever expressed relevant concerns about the children's safety and the evidence was replete with positive statements about the parents and children.

There was no basis for the judge to consider that the interaction between the mother and O was sinister or that a phone call in the night (which the mother had described before her phone was examined) was abnormal when he had evidence from both women showing that it was not.

(11) When further evidence was given on 20 March by the parents and the mother's brother, it was not substantially challenged and it was never suggested to either of them that anything had happened at N's home.

(12) Excessive weight was given to Y's comments in foster care, which did not even amount to an account of a factual event. No reference was made to the fact that Z had just been circumcised, which was plainly a possible explanation. It was not "a bold submission".

(13) In reaching the conclusion that the parents wished to stop Y 'opening up' the judge ignored the evidence that they had argued for X and Y to receive immediate therapy and that nothing similar was said in therapy when it was eventually arranged.

(14) Excessive weight was given to the mother concealing the attempted visit to O from the father, and the judge did not give himself a Lucas direction about it.

(15) The judge gave no proper reason for finding at [108] and [150] that the genital injuries had been inflicted 'in the late afternoon' or 'during the day', indeed there was no proper basis for this conclusion.

(16) The judge said that the genital injuries and the cause of death were intrinsically linked, but he did not justify this. His assertion that the mother may not have been thinking clearly or that she was angry with S or felt she had dishonoured the family in some way was not evidence-based. In fact the evidence showed that S was a much-loved child and the circumstances in the home overnight and of the discovery of the body make the judge's conclusion highly improbable. Nor had the link between FGM and the death been explored during the evidence.

76. As to the identification of the mother as the perpetrator of the genital injuries:

(1) The finding is unsatisfactory as it is not clear who has inflicted the injuries: is it the mother, or others, or the mother and others?

(2) Insofar as the finding is against the mother, the judge's rejection of her account of the events of the afternoon as vague and unconvincing was unfair as no one was aware that it was a relevant area of investigation.

(3) The court had unchallenged statements from N and from her 15-year-old daughter, who would have been likely to have been present at any FGM event.

(4) The findings could have repercussions for N or any other adults present. They were not on notice and the process was obviously unfair to them within the terms of Re W (below).

(5) There is an inconsistency between the finding in the order at B(iv) and the judgment at [150].

(6) Despite the judge's attempt at clarification at [156], his finding is not amenable to clarification because it is unsustainably uncertain.

77. As to the identification of the mother as the perpetrator of the neck injuries:

(1) The finding was an extraordinary and extreme finding made in a cursory manner, seemingly based on the mother having been awake and having taken part in an FGM event.

(2) It was made without the judge giving due consideration to:

The nature of the relationship between the mother and S.

The lack of any credible motive.

The inherent implausibility of her doing such a thing.

The improbability of her doing so undetected.

The position and the state of rigor mortis in which S was found.

The absence of any sign of struggle on the body or in the room.

The absence of the mother's DNA on the netting.

The account given by X of having heard a thump in the night.

(3) It was bolstered by an interpretation of the mother not having gone upstairs, which did not take account of her profound shock and distress, evident in the 999 call.

(4) It was based on impermissible exclusionary reasoning: not father or boys (though not explained in the case of W).

(5) The suggestion that the mother was feeling tired, angry or dishonoured was never put to her on any of the occasions when she gave evidence.

(6) The finding is so egregiously flawed as to cast doubt on the judge's rejection of the possibility of accidental causation.

78. As to the rejection of an accidental causation:

(1) Having discounted a sexual motivation, the judge gave scant attention to the unifying explanation of an accident. It was not incumbent on the parents to prove it.

(2) However, there were a number of indicators that it was a real possibility. It was physically plausible and may explain the sound heard by X in the night. It had not been ruled out by the expert evidence, indeed six doctors had been prepared to accept it as a possibility. However, the judge did not engage with it.

(3) The judge ignored factors that pointed towards an accidental explanation for S's death: the lack of any cries or other sounds despite the presence of two children in the same bunk and three people next door; the lack of any marks of restraint when S was described as a lively child.

79. As to the findings of collusion by the father:

(1) The judge was wrong to find that there had been a 40 minute gap between the discovery of the body and the emergency call. The timing was based on when E turned on his phone at 9.17 and the mother's 999 call at 10.13. E, whose evidence the judge generally accepted, gave evidence that he may have woken, turned on his phone and gone back to sleep. He strongly denied that there had been a 40 minute delay before the police were called.

(2) The judge paid insufficient account to the extent of the family's distress when finding their accounts of the morning's events inconsistent. Nor did he explain how any inconsistencies indicated collusion. He paid no account to innocent alternative possibilities for discrepancies, as discussed in Lancashire County Council v C, M and F [2014] EWFC 3 at [9].

(3) He overlooked a very large number of significant matters pointing away from collusion or silencing of the children. These include:

The acute improbability of the mother showing such distress during the 999 call if she had known for some hours that S had been dead.

The assessment of the family by the police officers and ambulance personnel that the family was being open with them.

The mother's immediate confidence that S had not taken her own life (which, if she was guilty, deprived her of a defence).

The parents' willingness for the children to have therapy.

The mother's account of questioning the boys about what had happened, even though it upset them.

The mother's long-term co-operation with the police throughout a most distressing time for the family.

The positive impression formed of the mother by the police liaison officer and the Children's Guardian.

The local authority's general view of the mother as reflected in its willingness to use her as a protective carer for 2½ years, in contrast with its submissions about her in court.

The fact that the police had never treated the mother as a suspect.

(4) It is not possible to see how the judge could find that only the parents had colluded but the brothers had not.

80. It is further said that in contrast to his giving little attention to the evidence about the family's very positive characteristics, the judge gave undue prominence to their origins and assessed their religious and cultural identity in an unbalanced way. The wider canvas showed no relevant family pathology, no mental illness or personality disturbance, and no relevant substance abuse. These matters were treated in a manner that was discriminatory in terms of Art. 14 as applied to Arts. 6 and 8.
81. Finally, the judge did not address the possibility that the evidence fell short of proof to the necessary standard. This had been accepted by this court on the first appeal as being a possible outcome of last resort: King LJ at [61].
82. If the appeal is allowed, it is submitted this court could not substitute its own findings, and on behalf of the boys, Mr Tughan QC and Mr Twomey QC strongly oppose the imposition of adverse findings when two trial judges have declined to make such findings against them. Mr Bagchi also points out that the judge accepted that the father had been asleep when S died, and argues that this court should not go behind that finding.
83. All family members argue that in the exceptional circumstances the case should not be remitted for a second rehearing.

The submissions of the local authority

84. In their written presentation, Mr Goodwin and Mr Parker defended the judge's decision on these broad lines:

(1) Overall, the findings were sound and integrated. The finding that the injuries were inflicted was based on solid expert evidence.

(2) There was nothing medically implausible about the genital injuries being inflicted on a separate occasion to the neck injuries. The medical evidence of a likelihood of strangulation was solid.

(3) The judge was entitled to reject an accidental explanation. Dr Cary considered it but had never seen such a case in this age group. Dr Kolar was of a similar view. They and Dr Leadbeatter considered that the likelihood of strangulation was significantly increased if the genital injuries were shown to have been inflicted ones.

(4) As to the genital trauma, the judge was entitled to find that there was no sexual motivation. The rejection of sexual motivation necessarily opened the door to other motivations. The findings did not suppose any distinctive procedures associated with FGM.

(5) The refusal to allow expert evidence deserves careful attention, but no formal application was made and the value of such evidence would be doubtful. In the event, Dr Lipetz' evidence did not exclude some preparatory examination having occurred.

(6) The steps taken by the judge in adjourning for further evidence and submissions ensured that the process was fair. It is accepted that detailed questions about events at N's house were not asked. W could have given more information if he wished, and the absence of any evidence that he saw or heard something amiss does not affect the finding.

(7) The judge was entitled to reach his conclusion about perpetration, even though it did not accord with the local authority's case. The mother's wakefulness was highly significant. As to improbability, the judge carefully emphasised the positives in the family. Once he had found the injuries to have been inflicted, the relative improbability of a family member being a perpetrator dissolved. There was no need for him to speculate about a motive.

(8) The judge's credibility findings differentiated between the boys (broadly credible), the father (credible in part) and the mother (credibility absent). These conclusions were reached as a result of the trial process and should not be disturbed.

(9) The judge confirmed after handing down judgment that his findings were against the mother and not against "the other women". No other effort was made to seek clarification of any aspect of the judgment about FGM. In any case, the mother could have carried out the assault undetected. So there is no ambiguity, nor any unfairness to others. Re W is distinguishable because of this, and because the finding made against the mother was within the parameters of the case and all parties had the opportunity to address it with evidence and submissions.

(10) As to collusion, the timings from phones alongside the family's own evidence showed a 40 minute unexplained period.

(11) Finally, the judge himself recognised that a case may fall short of proof.

85. In oral submissions, Mr Goodwin was more accepting of what he described as difficulties in the judgment. He said that the local authority had considered what its proper stance should be in response to the appeal. It had decided that it should not submit to the appeal, both out of respect for the unique position of this experienced judge and to assist this court with possible counter-arguments. The local authority is not wedded to the FGM finding, but it defends to the hilt the finding that S sustained inflicted injuries, that this was not suicide, and that there was no intruder. It also gives no ground in relation to the findings of parental collusion and it strongly resists the suggestion that there should not be another rehearing if the appeal is allowed. With somewhat less force, Mr Goodwin argues for this court to substitute a four person 'pool' finding if the appeal is allowed only in relation to the perpetration finding and not in relation to the injuries as having been inflicted. He argued against a finding that only the parents should be placed in the pool.
86. Mr Goodwin suggests that the real issue on this appeal is not whether there were inflicted injuries. The judge's finding about that was favoured by the experts and it was not outlandish. The real mystery concerns the identification of the perpetrator. That has been made difficult by a closing of ranks in the family which had cut off lines of enquiry. He took us to passages in the evidence that showed the improbability of S having been assaulted and killed without the mother having heard anything, and to other passages in the parents' evidence about the events of the night that he suggested were unsatisfactory.
87. Mr Goodwin observed that the challenge facing the judge cannot be overstated. He realistically accepts that if the FGM finding against the mother is unsound, it formed such an important part of the judge's overall reasoning that the finding that she had caused S's death cannot stand.
88. In arguing that the process was procedurally fair, Mr Goodwin points to the adjournment to allow for further evidence. However, he fairly conceded that the local authority was as surprised as the other parties to hear the judge's theory about FGM when it was first aired and that the lack of notice of the specific finding was procedurally difficult. No one had put to the mother that an event associated with FGM had been carried out, either in the family home or in N's home. The judge's reference to "other women" at N's home was not understood.
89. Mr Goodwin further accepted that the experts had not entirely dismissed the possibility that there had been an accident. He also agreed that it is difficult to compartmentalise the judge's conclusions and to substitute a finding against the boys after they had been exonerated at a trial at which E had given evidence accepted by the judge.

The position of the Children's Guardian

90. The Guardian played no part in the fact-finding appeal, beyond observing that the rehearing had not clarified for her what had happened to S and who, if anyone, was responsible for her untimely death. She did not anticipate the issue of FGM being elevated to the significance ultimately placed upon it in the judgment.
91. As to her position if the appeal succeeds, the Guardian questions whether a third hearing would be likely to resolve the disputed issues in a way that would fairly and properly shape any welfare decision. She draws attention to the heavy burden on the family of proceedings that have lasted for so long.

Appeals from findings of fact

92. An appeal court will rarely even contemplate reversing a trial judge's findings of primary fact: Re B (A Child) [2013] UKSC 33 per Lord Neuberger at [52]. Unless a finding is insupportable on any objective analysis it will be immune from review: ibid per Lord Kerr at [108]. The judge has had the opportunity to make a comprehensive assessment of all the information – written, verbal, non-verbal and visual – when reaching a conclusion. This court should therefore only interfere with findings of fact in limited circumstances, for example where there has been a material error of law, or the making of a critical finding of fact which has no basis in the evidence, or a demonstrable misunderstanding of relevant evidence, or a demonstrable failure to consider relevant evidence: Henderson v Foxworth Investments Ltd [2014] UKSC 41; [2014] 1 WLR 2600, per Lord Reed at [67]. Without such error, an appeal can only succeed if the appeal court is satisfied that the decision cannot reasonably be explained or justified and is one that no reasonable judge could have reached: ibid at [62, 67].

Principles of fact-finding

93. Although he was plainly aware of the nature of his task, the judge did not give himself a conventional self-direction in relation to fact-finding and to matters such as the possible significance of lies.
94. The starting point, as King LJ said in this case, citing Baroness Hale in Re B (Minors) [2008] UKHL 35; [2009] 1 AC 11 at [70], remains that the facts must be proved on the simple balance of probability. Neither the seriousness of the allegations nor the seriousness of the consequences makes a difference. The inherent probabilities are simply something to take into account in deciding where the truth lies.
95. Findings of fact must be based on evidence, including inferences that can properly be drawn from the evidence, and not on suspicion or speculation: A (A Child) (No 2) [2011] EWCA Civ 12; [2011] 1 FLR 1817.
96. The court is not bound by the cases put forward by the parties, but may adopt an alternative solution of its own: Re S (A Child) [2015] UKSC 20 at [20]. Judges are entitled, where the evidence justifies it, to make findings of fact that have not been sought by the parties, but they should be cautious when considering doing so: Re G and B (Fact-Finding Hearing) [2009] EWCA Civ 10; [2009] 1 FLR 1145, where Wall LJ said this:

"15. I am the first to acknowledge that a judge … is entitled to take a proactive, quasi-investigative role in care proceedings. Equally, she will make findings of fact on all the evidence available to her, including her assessment of the parents' credibility; she is not limited to the expert evidence. I am also content to decide the question in this appeal on the basis that a judge … is not required slavishly to adhere to a schedule of proposed findings placed before her by a local authority. To take an obvious example: care proceedings are frequently dynamic and issues emerge in the oral evidence which had not hitherto been known to exist. It would be absurd if such matters had to be ignored.

16. All that said, however, the following propositions seem to me to be equally valid. Where, as here, the local authority had prepared its Schedule of proposed findings with some care, and where the fact finding hearing had itself been the subject of a directions appointment at which the parents had agreed not to apply for various witnesses to attend for cross-examination, it requires very good reasons, in my judgment, for the judge to depart from the schedule of proposed findings. Furthermore, if the judge is, as it were, to go "off piste", and to make findings of fact which are not sought by the local authority or not contained in its Schedule, then he or she must be astute to ensure; (a) that any additional or different findings made are securely founded in the evidence; and (b) that the fairness of the fact finding process is not compromised."

97. As to fairness, the decision in B (A Child) [2018] EWCA Civ 2127, in which Newey LJ and I participated, confirms that:

"15. It is an elementary feature of a fair hearing that an adverse finding can only be made where the person in question knows of the allegation and the substance of the supporting evidence and has had a reasonable opportunity to respond.  With effective case-management, the definition of the issues will make clear what findings are being sought and the opportunity to respond will arise in the course of the evidence, both written and oral."

98. Re W (A Child) [2016] EWCA Civ 1140; [2017] 1 WLR 2415 concerned unexpected findings against professional witnesses, as to which MacFarlane LJ said this:

"95. Where, during the course of a hearing, it becomes clear to the parties and/or the judge that adverse findings of significance outside the known parameters of the case may be made against a party or a witness consideration should be given to the following:

a) Ensuring that the case in support of such adverse findings is adequately 'put' to the relevant witness(es), if necessary by recalling them to give further evidence;

b) Prior to the case being put in cross examination, providing disclosure of relevant court documents or other material to the witness and allowing sufficient time for the witness to reflect on the material;

c) Investigating the need for, and if there is a need the provision of, adequate legal advice, support in court and/or representation for the witness."

99. Finally, the judge did direct himself with reference to the well-known dicta of Dame Elizabeth Butler-Sloss P in Re T (Abuse: Standard of Proof) [2004] EWCA Civ 558; [2004] 2 FLR 838 at [33]:

"… evidence cannot be evaluated and assessed in separate compartments. A judge in these difficult cases has to have regard to the relevance of each piece of evidence to other evidence and to exercise an overview of the totality of the evidence in order to come to the conclusion whether the case put forward by the local authority has been made out to the appropriate standard of proof."

Conclusion: The FGM finding

100. The questions for every fact-finder are What, When, Where, Who, How and Why? Their significance and difficulty varies from case to case. Some answers will be obvious while other questions can be extremely hard or even unanswerable. Sometimes a question may not need answering at all. At all events the questions come in no set order and each inquiry will suggest its own starting point. It will no doubt find apparently solid ground and progress from there, but conclusions can only ever be provisional until they have been checked against each other so as to arrive at a coherent outcome. At each stage, regard is had to the inherent probabilities and improbabilities surrounding what are inevitably abnormal circumstances.
101. This was certainly a challenging investigation. That is confirmed by the fact that two senior judges have come to different conclusions on very similar evidence. What then did that evidence show?
102. There was clear evidence of what injuries S had suffered. They were traumatic injuries and it was unlikely that whatever caused them would have passed unnoticed in a small, crowded house.
103. There was dependable evidence of when the injuries occurred, with two overlapping windows of time for the neck injuries and the genital injuries. This was bound to raise the question of whether there was any good reason to believe that the two injuries were not caused at or around the same time.
104. As to where the injuries were sustained, the evidence pointed to the neck injuries and death as occurring in S's home and probably in her bedroom. That raises the similar question of whether there was any good reason to believe that the genital injuries were caused elsewhere.
105. The crucial question of how S came by her injuries and death – whether it was by assault or suicide or accident – was significantly illuminated by the experts, without their evidence being in any way definitive. Other permutations, such as accidental suicide, were not espoused by any party. The police investigation did not give all the help it might have done.
106. The next question, if these were inflicted injuries, was who might realistically have been responsible, given all that was known about the family and S's place within it? Was it one person or more than one, and was there collusion before or after the fact? That called for a consideration of motive and why such a thing might have happened.
107. Once all these questions had been considered, it was the court's task to decide what facts emerged and to consider whether they satisfied the threshold for making public law orders. In undertaking this task, the judge was operating within the framework of a sophisticated forensic process. There had been a police investigation, for all its faults. A large amount of information had been gathered in the course of two trials. The local authority had framed its case meticulously. The court had the benefit of expert opinion of the highest calibre and very experienced legal representation, all co-operating to assist the court to reach a sound conclusion.
108. It was against this background that the judge developed his own case theory. In such a vexed case, he was bound to consider all the possibilities but, as he himself said, there must be parameters. In particular, a judge who believes he alone may have discovered a path that has not been revealed to other experienced professionals is bound to reflect on why that might be so. The situation in this case fell squarely within the sound guidance found in Re G and B – cited at [75] above – which bears repeating:

"Where, as here, the local authority had prepared its Schedule of proposed findings with some care, and where the fact finding hearing had itself been the subject of a directions appointment at which the parents had agreed not to apply for various witnesses to attend for cross-examination, it requires very good reasons, in my judgment, for the judge to depart from the schedule of proposed findings. Furthermore, if the judge is, as it were, to go "off piste", and to make findings of fact which are not sought by the local authority or not contained in its Schedule, then he or she must be astute to ensure; (a) that any additional or different findings made are securely founded in the evidence; and (b) that the fairness of the fact finding process is not compromised."

In my judgment, the judge did not heed this guidance.

109. The judge was entitled to start from the premise that the medical evidence suggested that both sets of injuries were inflicted. He was also entitled to consider the possibility that the two types of injury were not sustained in the same incident and to challenge the assumption of a sexual motive. However, he was also bound to reflect on how the unprecedented finding he went on to make would fit with the evidence as a whole. In my view he should have seen that it faced insuperable difficulties, conceptually, evidentially and procedurally.
110. Conceptually, a case theory that S might have been injured in separate incidents at separate locations was a complicating hypothesis. It required there to have been two opportunities for her to exhibit distress and two events for the perpetrator(s) to cover up. Some solid evidential support was therefore needed to open the door to a hypothesis of this kind, but there was none. On the contrary, the evidence confirming the normality of S's behaviour during the day pointed the other way.
111. Evidentially, the difficulties were manifold:

(1) The FGM theory had no convincing roots in the evidence. The justifications given in the judgment were:

(a) Y's conversation with the foster carer.

(b) The mother's reason for distinguishing between male and female circumcision.

(c) All the adult women having suffered FGM.

(d) The older generation being more likely to continue FGM and O being a respected older woman.

(e) The mother staying out late on her first excursion since Z's birth.

(f) The mother misleading the father about her plan to go to O's house.

(g) The vague and unconvincing account of the events of the Saturday.

(h) The phone calls in the night.

(i) The improbability of an assault passing undetected in the crowded home.

(j) The mother's lack of credibility in relation to events on the Sunday.

With respect to the judge, the first eight of these matters are flimsy and contestable foundations for such a serious finding and the last two are non-specific and offer no support for it.

(2) The finding followed, indeed it depended upon, an incomplete assessment of the evidence. For example, the striking fact that twenty relevant children living in this country had not suffered FGM was self-evidently a matter that needed to be considered, and was presumptively more relevant than what had happened to their mothers in OC. Similarly, no attempt is made to reconcile the finding with the stance of the mother's influential brother. Other unchallenged evidence provided by the family on the issue of FGM was given no weight.

(3) Further, and to my mind critically, the judge was bound to descend to practicalities. But he did not have the necessary evidence to do that. He had a level of background information about the events of the day but, as no one was focusing on N's home as a potential crime scene, no detail. The police had not investigated at that address and, so far as I am aware, the court had no information about the most basic matters, such as the layout of the home, or any accounts from those present. The mother and N were there throughout and N's husband arrived later. There were no fewer than eleven children in the property, of whom seven were aged between 10 and 15. Any finding that S was seriously assaulted there required the court to descend to a thorough assessment of all available evidence. It could not be reached by a process of logic.

(4) Insofar as the judge did rely on logic, it is difficult to follow his reasoning. His twofold conclusion at [145] was (i) that the genital injuries were unlikely to have been inflicted after S returned home and (ii) that the forensic evidence established that they would have been sustained during that day. In support of the first conclusion, he had set out in the preceding paragraph a list what he described as key features of the evidence. None of these features could in my view be taken to support the theory that there had been separate assaults at separate locations. As to the second conclusion, the forensic evidence did not point to the genital injuries as having been caused outside the home, it merely permitted it. However, building on this reasoning, the judge concluded at [150] that because the genital injuries were inflicted during the day they must have happened in N's home. This process of reasoning does not appear to be logically sound or based in the evidence. It is also unclear what the judge meant when referring to "the other women" and the expansion of [156] does not provide clarity. The result of all this is that the reader can have no real sense of what has actually been found in relation to the genital injuries.

(5) The finding had no roots in the medical evidence. Such evidence as there was did not support any kind of attempt at FGM.

(6) Despite the idiosyncratic subject matter, the judge declined to allow specialist cultural evidence to be called, and as result we know nothing about whether there is such a concept as a preparatory examination, nor what the likelihood is of a mother who is not a cutter attempting FGM on her own child.

(7) Perhaps because the inquiry did not descend to practicalities, it led to the equivocal finding at [150] that the mother and/or other women were responsible. This, as Ms Ball says, creates an impression of a group of women gathered to carry out FGM, when the evidence was that there were no other women.

(8) Ultimately, in the words of Re G and B, there was no good reason for the court to depart from the way in which the local authority had framed its case. The course taken by the judge is the more surprising where the parties had repeatedly warned him about the lack of evidence to support the theory, or at any rate the theory as they understood it to be.

112. The judge's conclusion also faces insurmountable procedural objections, which need not be elaborated once the transcript has been read. He revealed his thinking about FGM in general but at no stage did he unveil the specific finding that he went on to make. The parties were blindsided by a finding that not only departed from the way in which the case had been put but contradicted it.
113. In sum, the substantive and procedural criticisms of the FGM finding, more fully catalogued at paragraphs 75-76 above, are convincingly made out. Indeed they are two sides of the same coin: had the finding had a sound basis in the evidence, it would not have taken the parties by surprise. Per Henderson, this was a critical finding of fact that had no basis in the evidence and there was a demonstrable failure to consider relevant evidence. The process was also fundamentally unfair, not only to the mother but to N, because the finding had simply never been litigated.
114. For all these reasons, the judge's finding in relation to the genital injuries must be set aside.

Conclusion: The other findings

115. The judge considered the two sets of injuries to be intrinsically linked [156] and, as the parties accept, the setting aside of his FGM finding undermines his other findings against the parents. Nor is it realistic for us to preserve just some of the findings, for example the finding that these were inflicted injuries, where all issues are so interlinked.
116. I have in any event certain concerns about the further findings. Firstly, the judge's statement at [157] that on the forensic evidence S's death is likely to have occurred at some point after the conclusion of the mother's conversation with O is not factually correct. Secondly, also at [157], the judge declined to draw conclusions as to why the mother caused her daughter's death. The court is not obliged to discover people's motives, but it is obliged to factor in the possibilities when considering the probabilities. Here, it was necessary for the judge to engage more fully than he did with the question of motive and to consider whether the lack of any apparent motive cast light on the other questions he was considering. Thirdly, it is possible that the prominence given to the FGM theory distracted attention from the competing theories of homicide and accident as promoted by the parties, and from any other possible explanations. Fourthly, given the very close quarters in which the family lived, it is unclear to me how the judge reasoned his conclusion that collusion, if it had occurred, was limited to the parents alone.
117. At all events, the judge's other findings of fact must also be set aside.

Conclusion: Substitution of findings

118. Mr Goodwin asks us to substitute a four person pool finding that S's injuries and death may have been inflicted by one or more of her parents and brothers. That is resisted by the family.
119. There is such a pressing need for a solution to these proceedings that a substituted finding should be made if it properly can be. But that can only happen where there is just one realistic consequence to the appeal being allowed. Here, even if we felt able to answer the questions 'what, where and when', there are too many possible permutations to 'how, who, and why'. Also, there are obvious difficulties in this court imposing findings of perpetration against the father and brothers in the light of the judge's assessment of their credibility.
120. I would therefore decline to substitute findings.

Conclusion: Rehearing

121. Mr Twomey, with the support of the other advocates for the family, invites us not to remit the matter for a second retrial. He draws attention to Oxfordshire County Council v DP, RS & BS [2005] EWHC 1593 (Fam), where the question was whether a fact finding hearing should take place at all. At [24], McFarlane J identified factors that should be borne in mind when deciding whether or not to conduct a particular fact finding exercise:

"(a) The interests of the child (which are relevant but not paramount);

(b) The time that the investigation will take;

(c) The likely cost to public funds;

(d) The evidential result;

(e) The necessity or otherwise of the investigation;

(f) The relevance of the potential result of the investigation to the future care plans for the child;

(g) The impact of any fact finding process upon the other parties;

(f) The prospects of a fair trial on the issue;

(g) The justice of the case."

122. Mr Twomey points out that even in a case of serious allegations, the family court is not obliged to direct even a first retrial. In Re J (Vulnerable Witness) [2014] EWCA Civ 875; [2015] 1 FLR 1152 this court allowed an appeal from sexual abuse findings. It did not order a retrial (i) because it did not consider that a finding of fact was open to the court on the evidence as a whole; (ii) because it was highly unlikely that the vulnerable complainant would be able to engage in a further hearing; and (iii) because no greater clarity would be achieved by a retrial. The court therefore put a stop to the proceedings: see [101-103].
123. Mr Twomey also submits that in criminal proceedings, second retrials do occur but are unusual. The convention (but it is no more than that) is that where two trials have ended in jury disagreement, the prosecution will normally not seek a third trial. The ultimate question is whether the interests of justice justify a second retrial: R v Bell [2010] 1 Cr App R 27. There must be a searching examination of why a third trial is justified if there have been no irregularities in the first two: R v Burton [2015] EWCA Crim 1307.
124. Against this background, Mr Twomey submits that:

(1) A second retrial would not serve the interests of justice and would be unjust and oppressive. It would constitute a breach of the family's entitlement under art. 6 and at common law to a fair hearing within a reasonable time. The family has already faced what he describes as two botched trials in which different cases have been levelled against them by three arms of the state. Their physical and mental health is threatened. For E (18) and W (15) in particular, a continuation of proceedings that have consumed so much of their childhoods would be unfair and an abuse of process.

(2) The delay has made a satisfactory outcome to a further trial less likely. The evidence has become stale and the ability of lay witnesses to recall events and give their accounts effectively has been hampered by constant repetition and by hearing the accounts of others. This must confound assessments of credibility.

(3) The forensic value of a further trial is doubtful. All matters are in issue. The evidence against the family members is weak and circumstantial. Even if a four person pool finding was made, its value is limited. Risk can be managed without findings, and no harm has come to the younger children in the last three years.

(4) Mr Twomey also renews a submission that a pool finding that included a child who is not a carer could not satisfy the threshold. I do not consider that this is likely to be well-founded, but we did not hear full argument and it is unnecessary to take time on it on this appeal.

125. In response Mr Goodwin accepts that another trial would be an abuse of process if there was no realistic prospect of a third judgment finding that the s. 31 criteria were met. Here, however, whatever the challenges of a further trial, there is a very solid evidential foundation for the conclusion that S's injuries were inflicted by a third party, that she did not commit suicide, that there was no intruder, and that there has been family collusion. The local authority does not accept that it is in a position to protect the younger children by other means and it notes that Y is now approaching the age at which S died. It considers that the finding it seeks would have a forensic value and would allow it to take any measures that were in the interests of the children.
126. The views of the Guardian have been recorded above. All parties are agreed that our decision cannot be influenced by speculation about what course the inquest into S's death may take.
127. In deciding this issue, this court has a broad discretion, and the discretion is particularly broad when we are considering whether to order what is probably an unprecedented second rehearing. All relevant matters must be taken into account. Here, I take account of:

(1) The seriousness of the issues. In a case of this extreme gravity, a party, here the local authority, should in my view only be shut out from a determination of its case if there are strong countervailing reasons.

(2) The interests of the children. If it is possible, it is in the interests of all the children, and the youngest three in particular, for there to be valid findings about their sister's death and for measures to be taken for their protection if that proves to be necessary.

(3) The likely evidential result. Although it is profoundly unsatisfactory that there is still no clarity about how S came by her injuries and death, there is no reason to believe that a rehearing cannot provide a legally valid conclusion that would make the matter clear, or at least clearer. This case is different to Re J, where the evidence was incapable of supporting findings. Here, two appeals have succeeded because of errors of process and not because the evidence is incapable of justifying a s.31 finding. If the local authority was to succeed, the court might face a very difficult welfare decision, but that is not a good reason for abandoning the proceedings.

(4) The fairness of a further trial. I am alive to the concerns expressed about litigation fatigue and its possible effect on the integrity of the trial and the assessment of witnesses. At the same time, it is at least possible that some of the professional evidence already heard can be carried forward so that it need not be given again. As to the lay evidence, the trial judge will no doubt be alert to the risks that have been pointed out to us. I would therefore not accept that there cannot be a fair trial.

(5) The impact upon the family. This carries significant weight where proceedings have already lasted for three years and where a further trial must considerably extend that period. The emotional cost to the family of the loss of S and of the continuous proceedings cannot be overstated. The Guardian has made the observation that they have not yet properly been able to mourn S's death. I am mindful of all this, and take it into account. In the end, however, it has not been shown that the burden of continued proceedings would be disproportionate to the seriousness of the matters in issue.

128. After full reflection, I would therefore hold that in these most unusual circumstances a second retrial is unavoidable and that it would serve the interests of justice. I would stress that nothing that I have said when addressing the very specific circumstances of this appeal is intended to limit the scope of the retrial or to influence the court's assessment of the evidence or its consideration of all the possible explanations for these tragic events.

Outcome

129. I would allow the appeal, set aside the judge's findings and remit the matter for an early case management hearing and for subsequent rehearing in accordance with arrangements that have been made with the President of the Family Division. By agreement of the parties, the Female Genital Protection Order will be set aside. The judgments and order of this court will be sent to HM Coroner.

Lord Justice Newey

130. I agree.

Lord Justice Underhill

131. I have no doubt that this appeal must be allowed, for the reasons so carefully explained by Peter Jackson LJ. That means that this Court has had to allow an appeal against two successive judgments of the Family Court on the same issue. Such a situation is, fortunately, vanishingly rare; but it is not the less a cause for deep regret. However, I think it is fair to both Judges to say that they were faced with a very difficult task. The circumstances of S's death are extremely puzzling, and the evidence about it is unusually unsatisfactory, partly as a result of an inadequate police investigation. That makes it less surprising that two conscientious judges should have gone astray, albeit in different ways, in trying to find a path to a fair decision.
132. The much more difficult question is what order we should make as a result of allowing the appeal. I am sure that it would be wrong for this Court to make a finding of its own; but the real question is whether the local authority's application should be remitted for what will be a third hearing. I feel acutely the force of the argument that it would be wrong to put this family, and indeed the witnesses who would have to be called, through the ordeal of such a hearing, particularly when there can be no certainty – bearing in mind the evidential problems to which I have already referred – that it will result in a definitive finding or one which will lead to a straightforward welfare decision. I have wrestled with this, but in the end I think the balance must come down in favour of a further hearing, for the reasons given at para. 127 of Peter Jackson's LJ judgment. His points (1) and (2) carry great weight. A child has died, and (then or shortly before) sustained significant genital injuries, in the middle of the night in a house occupied only by members of her own family. It is of the greatest importance, from the point of view of all the members of the family, that valid findings be made as to how that happened; and, like Peter Jackson LJ, I am not persuaded that findings may not still be made which will enable the local authority to reach a properly-informed decision about how to proceed.