A, B and C (Children: Perpetrator Finding), Re

LORD JUSTICE BAKER :

1. This is an appeal against a judge’s finding in care proceedings that an injury to a small child was inflicted by his mother.

2. At the end of the hearing, the Court informed the parties that the appeal would be dismissed. This judgment sets out my reasons for joining in that decision.

3. The proceedings concern three young boys, whose names I shall anonymise as A, aged 5, B, rising 4, and C, now 18 months old but about 15 weeks old at the date he sustained the injury.

4. The boys’ parents started a relationship in 2018. In 2019, following A’s birth, they moved into a flat. In June 2020, A was made subject to a child protection plan after an incident in which the father slapped his face.

5. The following month, the parents separated. In October 2020, the mother alleged that the father had been abusive to her and was granted a non-molestation injunction against him. The father did not contest the injunction on the basis that no findings were made. In the same month, the father’s mother – the paternal grandmother (“PGM”) – reported the father to the police after he had caused damage by punching her car windscreen. The father admitted that he had damaged the car. No action was taken by the police.

6. The mother applied for a child arrangements order. In April 2021, those proceedings concluded with an order for the father to spend unsupervised time with A, progressing to overnight stays. The non-molestation order was discharged.

7. In May 2021, the parents resumed their relationship. The following month, A’s child protection plan came to an end and he was made subject to a lower level child in need plan. In October 2021, the local authority formally closed the file.

8. In March 2022, the mother gave birth to B. The parents’ relationship remained unstable. In early 2023, they separated again, with the children remaining in the mother’s care. In November of that year, however, they resumed living together and the mother became pregnant for a third time.

9. In July 2024, A told a teacher at nursery that his father had dragged him and caused scratches to his back and knee, and that there was lots of shouting at home. In August 2024, the mother gave birth to C. A few days later, all three children were made subject to child in need plans.

10. In October 2024, an incident took place in which the father was abusive to a traffic warden. Following this incident, the parents separated again and have not resumed their relationship. The children remained with the mother. The father moved back to live with the PGM. The parents agreed that he would have overnight contact at her house with all three children.

11. Contact took place between 8 and 9 November. When the children returned, there was a bruise on C’s chest. The mother complained about this to the children’s social worker. The father’s account was that this had been caused by B accidentally when playing with a toy dinosaur while the PGM was in the room. C was seen by the social worker and a doctor who accepted the father’s explanation.

12. On 13 November, the mother attended at A’s school and asked the teacher not to send A home with his father at the end of the week. She alleged that the father was shouting, controlling and violent.

13. In the early hours of 16 November, the father was involved in a fight with his brother at the PGM’s home. The police attended and the father was arrested but later released without charge.

14. At lunchtime on 16 November, the father and PGM collected the children from the mother for overnight contact. The father, who was visibly injured, told the mother that he had fallen downstairs.

15. On 17 November, the father and PGM returned the children to the mother at about 7pm. They told her that C had been unsettled throughout the visit. They attributed this to stomach discomfort. It was the evidence of the father and the PGM that they noticed nothing unusual about his left arm.

16. It was the mother’s evidence that C slept throughout the night of 17 to 18 November for 12 hours without waking for a feed or nappy change. On 18 November, the mother spent some time with a friend, LP. The evidence about what the mother did or did not notice or do that day was the subject of detailed analysis by the judge which is considered below.

17. At about 7pm, the mother called her mother (the maternal grandmother – “MGM”) and said she wondered whether C’s arm was ok. The MGM drove round to the mother’s house and saw that the arm was “not right”. She drove the mother and the children to hospital, arriving just after 8pm. The examining doctors were initially not concerned. C seemed happy and was moving his arm. It was decided, however, to admit him overnight for observation.

18. On the following day, X-rays were taken which revealed that C had sustained an incomplete fracture of the proximate left humerus. C was discharged from hospital into his mother’s care. The father (but not the mother) was arrested and interviewed.

19. On 9 December 2024, the local authority started care proceedings in respect of all three children. On 17 December, interim care orders (“ICOs”) were granted on the basis of care plans that the children would be cared for by the mother under the supervision of the MGM. That arrangement did not continue for long because of the MGM’s health problems. In the short term, the children remained in the mother’s care under the ICOs.

20. At a case management hearing in January 2025, a circuit judge sitting in the family court made an exclusion order against the father preventing him entering or returning to the address where the mother and children were living. At a further case management hearing, the judge approved an amended care plan for the mother and children to move to a residential unit in a different part of the country for a parenting assessment.

21. Expert medical reports were commissioned from a paediatric radiologist and a paediatrician. They reached the following conclusions. The fracture had been sustained between 9 and 19 November 2024. There was no medical or organic cause for the fracture which had been sustained as a result of a significant force applied to the bone. It would not have occurred through normal handling, rough play or inexperienced parenting, and was therefore sustained non-accidentally. Any person causing the fracture would have realised the force used was excessive and inappropriate. At the time the fracture was sustained, C would have been in pain and shown signs of distress lasting for some minutes. Thereafter, movement of the arm would have caused him discomfort and he would have been distressed when his arm was handled. This expert evidence was not challenged by the parents. It was accepted that the only persons who could have inflicted the injuries was one or other of the parents. Each parent denied that he or she was responsible.

22. The findings sought by the local authority were set out in a threshold document. In summary, they were as follows: (1) The injuries to C were caused by one of his parents. On the available evidence it was not possible to say which parent caused the injuries. (2) C suffered significant physical and emotional harm as a result of the injury. (3) All three children were placed at risk of suffering significant physical and/or emotional harm by being subjected to and/or witnessing the incident and/or similar incidents. (4) The father caused A significant physical and emotional harm through physical abuse when he slapped his face in June 2020 and dragged him across the floor in July 2024. (5) The mother failed to protect the children by resuming her relationship with the father in May 2021 and continuing this until October 2024 despite having witnessed and/or knowing about his physical abuse of A. (6) The perpetrator of C’s injury failed to seek timely medical attention for him when this occurred despite being aware of an injury having been caused. (7) The father perpetrated domestic abuse on the mother, including in the presence of the children. The parents’ relationship was characterised by frequent arguments witnessed by the children. In or around the middle of June 2020, the father “got into the mother’s face,” (as it was phrased in the threshold document), shouted and was verbally abusive towards her whilst A was in the home. He was intimidating, aggressive, coercive and controlling towards her, and pressured her into having sex. (8) The father became involved in a physical fight with his brother in the early hours of 16 November 2024. (9) The mother failed to protect the children from being placed at risk of and/or caused significant emotional and/or physical harm by becoming embroiled in and/or exposed to domestic abuse perpetrated by the father.

23. The fact-finding hearing took place over three days in August 2025. Judgment was reserved and handed down on 22 August.

24. The judge began by summarising the background as set out above. She recorded that the parties had agreed a note of the relevant law which she had read in its entirety. She set out some of the salient legal principles, including (at paragraph 28) the following summary about the identification of the perpetrator of injuries to a child: “It is always desirable, where possible, for the perpetrator of non-accidental injury, to be identified both in the public interest and in the interests of the child. The court should first consider whether there is a list of people who had the opportunity to cause the injury, then consider if it can identify the actual perpetrator on the balance of probabilities. Only if it cannot identify the perpetrator to the civil standard of proof, should it go on to consider in respect of those people on the list, is there a likelihood or real possibility that person A or person B was the perpetrator of the inflicted injury?”

25. She then summarised the evidence given by the witnesses who had been called at the hearing – the mother, her friend LP, the MGM, the father and finally the PGM. In each case, having recited the witness’s evidence, she set out her view of their reliability. (1) The mother “I initially thought that she was an honest and reliable witness who said she could not remember because she did not want to accidentally mislead the court. I came to believe that it was a default or a convenient refuge from facing difficult questions, particularly when inconsistencies or untruths were being pointed out. I found the mother to be a poor witness, with numerous inconsistencies and contradictions in her evidence.” (2) LP “My impression … was that she was a good witness, who gave a clear and consistent account.” (3) The MGM “My impression of the maternal grandmother was that she lacked confidence in her own evidence, and she too frequently defaulted to, I do not know or I cannot remember, particularly when she thought she may have said something inconsistent or unhelpful to the mother.” (4) The father “My impression of the father was that he was frequently emotional during his evidence, and appeared genuinely remorseful about his behaviour, both towards A in 2020 and towards the mother in the course of their relationship. He accepted much of what was put to him in cross-examination. There was a notable increase in the things conceded by the father in his oral evidence, compared to the time that his written evidence had been prepared. I did not see that so much as inconsistency, but evolution of insight and acceptance. He did not get riled, although he clearly found the experience of giving evidence difficult. I consider that the father gave reflective evidence, and I am satisfied that he was doing his best to be honest.” (5) The PGM “My impression of the paternal grandmother was that she was clearly minimising the circumstances of the fight, although in her police statement about it, she had given graphic and frank details. She was clearly reluctant to admit that the father had scared her in 2020, although again at that time she had given a candid account to the police. It was her who reported her son to the police and, as I have already said, her exasperation or despair over C’s crying seemed very real.”

26. Central to the judge’s analysis was her impression of the evidence given by the mother, her friend LP and the MGM about the events of 18 November 2024. Her summary of the mother’s evidence identified a number of inconsistencies, including about when she first noticed anything unusual about his left arm. She told the hospital that she noticed his arm was floppy when she changed his nappy and that lifting his arm made him uncomfortable. In her response to the schedule of findings, however, she denied that the arm was floppy or not moving in the morning, and in a position statement it was said on her behalf that it was only in the evening that she noticed a pain response when she moved the arm. LP’s evidence, however, was that within a couple of minutes of her arrival at 9.30 that morning, the mother said she was a little bit worried about C’s arm because it was floppy. LP noticed that C was moving both of his arms but later she had seen some discomfort when he was being put into his pram suit. At that point she told the mother she should get the arm checked. The MGM’s evidence was that she had received a call from the mother in the evening and that the mother had explained that when she tried to put C in his baby grow, he began to cry with a different cry. She explained she had been with LP for the day and had been wondering if C’s arm was okay. She had not told her mother that LP had advised her to get medical advice. She had not told her mother that C’s arm had been floppy since the morning. When the MGM arrived at the house, she could see it was “not right”.

27. The judge concluded by setting out her analysis and findings over thirty paragraphs. She found that there was an unhealthy relationship between the parents, characterised by numerous arguments, separations and reconciliations. She found that there were occasions when the parents’ heads were in close proximity during arguments and that the father pressed his head against the mother’s. On the other hand, she rejected the mother’s allegation that the father smashed things in the house and did not find that he was coercive and controlling towards her, nor that he had pressured her into having sex. She found that, as he admitted, the father had slapped A’s face in 2020 but not that he had dragged A across the floor in 2024. She found that the mother had exposed A to emotional harm by repeatedly questioning him about C’s injury. She found that both parents failed to protect the children from emotional harm as a result of exposure to arguments.

28. Turning to C’s fracture, the judge summarised the unchallenged medical evidence. She then said: “118. The court must, of course consider the wider evidential canvas. In this case, there is no dependency or misuse of illicit drugs or alcohol by either parent. There were arguments with shouting and verbal abuse directed by each parent towards the other. I find there is no evidence of propensity towards physical violence towards a child. There was the incident in May 2020, in respect of which the father has demonstrated significant remorse. Since then, he has completed parenting work, anger management work and talking therapies. He has been assessed as safe and competent to have unsupervised care of his children.

119. The mother confirmed that since that time she had never seen him physically abuse any of the children or had reason to think he was hurting them. Even after the separation following the incident with the traffic warden, the mother immediately agreed to the children having unsupervised and overnight contact with their father. She trusted him not to hurt the children and/or that the paternal grandmother and great grandmother were a protective factor, although there was no stipulation that they must be present.”

29. The judge then said that she was “deeply troubled by the inconsistencies and contradictions in the mother’s evidence generally, and the evidence of the mother and maternal grandmother in respect of the events of 15th to 18th of November, in particular.” She gave a number of examples – the fact that the MGM spent two hours in C’s company on 15 November without holding him; a conflict between the mother’s oral evidence and contemporaneous text messages about whether C was taking formula milk without difficulty; and the mother’s evidence that C, aged three months, had slept through the night of 17 to 18 November for 12 hours without waking for a feed or nappy change. She continued: “123. The mother’s evidence about when she noticed something was wrong with C’s arm was wholly inconsistent and seemed to be, at the very least, altered to avoid criticism for delay in seeking medical attention. At no point did the mother call the father to ask questions about C’s arm or whether anything had happened. I would expect that to be the first thing she did when she had concerns and at the very latest, when she was at the hospital. The mother was unable to explain why she had not.

124. Given the mother’s wish for the social worker to be immediately notified when C sustained a bruise to his chest the previous week, I am surprised that she did not call the social worker to raise concerns about C’s arm. Again, she was unable to explain why.”

30. In contrast, the judge found the father’s evidence “more reliable”. He now accepted the “toxic dynamics” in his relationship with the mother. The judge found his remorse about that, and about the incident when he had slapped A in 2020, was genuine. She accepted the evidence of the father, and the PGM, about C’s distress over the weekend of 16 to 17 November which they attributed to stomach discomfort. She remarked that this had been “consistent from the outset” and was “entirely plausible”. She observed that the father and PGM had not noticed anything wrong with his arm, but observed this was “possibly because they simply were not looking at it and because the crying was continual.” Although LP had noticed something about C’s left arm, she said that he was still using it. The doctors at the hospital did not realise he had sustained a fracture until the x-rays were conducted the following morning.

31. The judge noted that the PGM’s evidence about the sleeping arrangements, with all doors being left open, was unchallenged. She concluded: “I find that the father was in the company of [the PGM] and/or paternal great grandmother throughout the entire weekend and that had C been injured in that time, they would have heard.” The judge acknowledged that the PGM had minimised the seriousness of the fight between the father and his brother, but added that she had given a frank account to the police and previously reported the father about the windscreen incident. The judge concluded: “I am satisfied that had the father harmed C, [the PGM] would have said something at the time.”

32. The judge then set out her conclusion: “On the balance of probabilities, for all of those reasons, I find that the mother caused the fracture to C’s arm. I cannot say if that was before the children went to stay with their father on 16 November or after their return on 17 because there is no reliable evidence about his presentation before going to contact or following his return.” The judge added that “there was a failure by the mother, as the perpetrator, to seek prompt medical attention.”

33. At a further case management hearing on 12 September, the mother’s application for permission to appeal was refused. Permission was given for the mother and children to undergo a “Resolutions” style assessment by an independent social worker.

34. On 28 September, the mother’s solicitors filed a notice of appeal to this Court against the finding that the mother had inflicted the injuries. On 13 January 2026, permission to appeal was granted.

35. Two grounds of appeal were initially put forward with the appeal notice: (1) The judge was wrong to find on the balance of probabilities that the mother inflicted the partial fracture to C’s arm because she: (i) gave too much weight to inconsistencies or uncertainties in the evidence of the MGM when they did not assist in establishing when the injury occurred or considering sufficiently why there was uncertainty in the evidence; (ii) gave insufficient weight to the inconsistencies within the evidence of the PGM and the minimisation of risk factors therein. (iii) failed to sufficiently take account of relevant matters within the wider canvas of the evidence when considering the likelihood of each of the parents having inflicted the injury to C including factors which indicated a propensity for the father to cause harm and therefore the balancing exercise was flawed. (2) The finding in respect of the fracture to C’s arm would have had such a preponderant influence on the consideration of the evidence of both parents as a whole that there should be a full rehearing of all matters.

36. Shortly before the appeal hearing, Mr Aidan Vine KC was instructed to lead trial counsel Mr Richard Carroll at the hearing. An application to amend the appeal notice was made seeking to add two further grounds. The application was not opposed and permission was granted at the start of the hearing. The additional grounds were: (3) The judge misdirected herself in law and fact as to the absence of evidence of propensity relating to the father. (4) The judge failed to evaluate the evidence as a whole and failed to explain why the outcome she reached was selected in preference to the other possible outcomes.

37. In addition to Mr Vine and Mr Carroll, we also had the benefit of submissions from Ms Gemma Bower on behalf of the local authority, Ms Lucy Hendry KC, leading trial counsel Ms Tracey Hennessey on behalf of the father, and Mr Dylan Morgan on behalf of the children’s guardian. All opposed the appeal.

38. It is axiomatic that the assessment of evidence, and the apportionment of weight to be attached to each piece of evidence, are matters for the judge at first instance. An appeal court does not conduct a rehearing of the evidence and will not interfere with findings of fact by trial judges unless there is a very clear justification for doing so: see Lewison LJ in Fage UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5, paragraphs 114-5 and in Volpi v Volpi and another [2022] EWCA Civ 464, paragraph 2.This was a paradigm example of a case which turned on the judge’s impression of the witnesses and the reliability of their evidence. In respect of each witness, the judge summarised her view of the reliability of their evidence. Her comments, which I have recited at paragraph 25 above, were considered and nuanced. For example, she described how her impression of the mother’s reliability and honesty evolved in the course of her evidence. She also described her interpretation of the differences between the father’s written statements and his oral evidence as evolution of insight and acceptance rather than inconsistency. For my part, I can see no justification in interfering with the judge’s assessment of the evidence or her finding as to the perpetrator.

39. Mr Vine and Mr Carroll developed the points raised in the amended grounds of appeal in their succinct skeleton arguments and oral submissions. None of their arguments persuaded me that there was any reason for this Court to interfere. The judge’s observation about the MGM’s evidence concerning 15 November was of marginal relevance to her finding that the mother had perpetrated the injury. The importance of her evidence about 18 November was that it contributed towards the judge’s assessment of the unreliability of the mother’s evidence. I do not agree with Mr Vine’s suggestion that the judge in some way “blended” her assessment of the mother’s and MGM’s evidence. The judge was fully aware of the points raised on the mother’s behalf about the PGM’s evidence – including her minimisation of the seriousness of the fight between the father and his brother – and took them into account. Ultimately, for the reasons clearly set out in the judgment, she concluded that they should not weigh decisively in her assessment. They did not undermine the reliability of the PGM’s evidence about the overnight contact on 16 to 17 November. The judge was satisfied that, had the PGM heard or seen something that happened to C over that weekend, she would have said so. That was manifestly a matter which she, rather than an appellate court, was best placed to decide.

40. In his supplemental skeleton and oral submissions, Mr Vine sought to persuade us that the judge misunderstood, or failed to have regard to, the expert evidence as to symptomatology, which was not set out in the judgment. In particular, it was argued that the judge failed to take into account the expert evidence to the effect that, other than at the moment of infliction and its immediate aftermath, the nature of C’s incomplete fracture meant that its symptomatology was likely to have been variable and non-specific. Although there is no express reference to this in the judgment, it is implausible that the judge overlooked it when conducting her analysis of the evidence. She expressly took into account LP’s evidence that C was able to use his left arm, albeit not as much as his right, and the fact that the doctors did not diagnose any problem on the evening of 18 November or at any point before the x-rays were carried out. As I read the judgment, the important point was not that the symptoms were variable but that the mother said different things at different times about when she first noticed there might be something wrong with C’s arm. In assessing this evidence, the judge took into account the possibility that the mother had changed her story to avoid criticism for the delay in seeking medical attention. But the judge also took into account that the mother failed to follow LP’s clear advice to get the arm checked, failed to mention LP’s advice to the MGM, never contacted the father to ask whether anything had happened to his arm during his contact, and failed to alert the social worker, in contrast to what she had done only a few days earlier when C sustained a bruise to his chest. In that context, as Ms Hendry put it in her well-argued submissions, the judge was entitled to conclude that the inconsistencies in the mother’s evidence were not explained by nuanced symptomatology or a wish to deflect blame for the late presentation to hospital.

41. The focus of Mr Vine’s challenge in oral submissions was on the judge’s treatment of the father’s propensity. It was submitted that there was plenty of evidence that he had a propensity to anger and violence generally, including to family members and one of his children. Mr Vine contended that the judge failed to consider the evidence of propensity or, in so far as she did consider it, looked at each incident in isolation rather than “in the round”. He cited case law about similar fact evidence, in particular R v Mitchell [2016] UKSC 55, [2017] AC 571, as applied in R v P (Children: Similar Fact Evidence) [2020] EWCA Civ 1088, [2021] 1 FLR

652.

42. The issues around similar fact evidence raised in those cases have no bearing on the point under consideration on this appeal. The law about similar fact evidence is concerned, first and foremost, with admissibility. In the family jurisdiction, the rules about admissibility are generally less strict than in other areas. The crucial question is as to the weight to be attached to the evidence. Here, there was plenty of evidence – covering the various incidents involving A, the mother, his brother, the PGM’s car, and the traffic warden – that the father had the propensity to be angry and aggressive. The judge was fully aware of all of this evidence, analysed it in some detail, and took it into account in her overall evaluation. It was plainly evidence which was capable of supporting a finding that the father had inflicted the injury to C’s arm, but on the totality of the evidence the judge reached a different conclusion. In my view she was entitled to do so. As submitted by Ms Hendry, the father had acknowledged his wrongdoing in all of those incidents, and his acknowledgment of wrongdoing enabled the judge to form the view that he was generally an honest witness.

43. Finally, it was submitted by Mr Vine that the judge failed to apply the procedural rigour required by case law to the exercise of identifying the probable perpetrator, failed to evaluate the evidence as a whole, and failed to explain why the outcome she reached was selected in preference to the alternative. There is no merit in these submissions. I accept Ms Hendry’s submission that, as the judge reached the conclusion on a balance of probabilities that the injury was inflicted by the mother, she was not required to engage further with the principles about a pool of perpetrators. As for her reasoning, this Court must apply the approach articulated by Lewison LJ in Fage, supra, at para 115: “The primary function of a first instance judge is to find facts and identify the crucial legal points and to advance reasons for deciding them in a particular way. He should give his reasons in sufficient detail to show the parties and, if need be, the Court of Appeal the principles on which he has acted and the reasons that have led him to his decision. They need not be elaborate. There is no duty on a judge, in giving his reasons, to deal with every argument presented by counsel in support of his case. His function is to reach conclusions and give reasons to support his view, not to spell out every matter as if summing up to a jury.” In this case, Judge Magee’s judgment plainly met the standard required when explaining why she concluded that the injury was inflicted by the mother.

44. For those reasons I reached the clear conclusion that the appeal should be dismissed.

45. It was plain when we announced our decision at the end of the appeal hearing that this outcome is deeply distressing to the mother. I am therefore pleased to end on a note of optimism. Since the hearing, the mother and children have been subject to a “Resolutions” style assessment by an independent social worker. Although the final report was not available at the date of the hearing, we were told that the assessor intends to recommend that the children remain in the mother’s care and that, if so, the local authority, father and guardian were likely to accept the recommendation, subject to consideration being given as to the terms of the final order. If this turns out as everyone hopes, it will mean that the mother has worked hard to address the problems which were affecting her capacity to care for the children and that the professionals will have concluded that it is safe and in the children’s interests to remain in her care. Wherever possible, all children should be brought up by their parents. For my part, I hope that that these three young boys are able to remain safely with their mother. LORD JUSTICE ARNOLD

46. I agreed that this appeal should be dismissed and I am in full agreement with the reasons given by Baker LJ for that decision. I would nevertheless add a brief comment. In this case the events of 14-18 November 2024 were central to the issues before the court. Each of the witnesses gave an account of those events. In addition, there was a limited amount of documentary evidence in the form of text messages exchanged between the mother and the father. The judge’s approach was carefully to review the evidence of each witness and then to consider the overall effect of the evidence based primarily on her assessment of the credibility of the witnesses. That approach and the findings based on it were fully open to the judge. In my view, however, it would have been advantageous to analyse the evidence chronologically: who was where when, what messages were exchanged over that time period and what other events occurred. As can be seen from paragraph 130 of her judgment (quoted by Baker LJ in paragraph 32 above), the judge concluded that, on the balance of probabilities, the mother caused the fracture, but the judge was unable to say whether this occurred before the children went to stay with their father on 16 November or after their return on 17 November. The judge’s findings suggest that it was more likely to have happened between about 7pm on 17 November (when the children were returned to the mother’s care) and 9 to 9.30am on 18 November (when LP arrived). One of the advantages of a chronological analysis of the evidence is that it would have assisted the judge to determine both the probable perpetrator and the probable timing of the injury, which are inter-related questions. LORD JUSTICE DOVE

47. I agree with both judgments.