‘Well, the burden never shifts, but it does’: celebrity, property offences and judicial innovation in Woolmington v DPP

Abstract In his 1935 judgment in Woolmington v DPP, Viscount Sankey declared the prosecution's burden of proving the accused's guilt was a ‘golden thread’, running ‘throughout the whole web of English criminal law’. This paper explores what Woolmington can tell us about the appeals process – and about the criminal law itself – less than 30 years after the first automatic right to appeal was created in English criminal law. It argues that the decision helps us understand the political pressures that could help to form – and make possible – legal decisions during this period. And it finds that the Woolmington decision itself – both in the text of the decision and in its immediate reception – was more universal than it was fundamental. Woolmington, I argue, has always been more about the high-level principles of English criminal law than about securing any kind of minimal procedural rights for a defendant.


Introduction
In 1935, the 21-year-old agricultural labourer Reginald Woolmington was convicted of the murder of his 17-year-old wife Violet. They had been married for a year when Violet had given birth to a son. Approximately a month later the two had separated, after he had started to threaten her whenever she left home without his permission. A month after that she had been shot dead by a gun he had been holding. Reginald Woolmington had for roughly one year been a member first of the Royal Marines and later of the Royal Field Artillery, but urinary incontinence had prevented him from serving on a long-term basis. 1 At his trial, 2 Woolmington claimed he had simply wanted to scare his wife, and that the gun had gone off by accident. Swift J directed the jury that in murder trials the defendant had the burden of proving this defence. This had been an orthodox direction for almost two centuries. But now the House of Lords held that a defendant should not normally be expected to prove their defence.
The decision in Woolmington has been widely regarded as being central to the idea that an accused person is innocent until proven guilty, and that it is for the prosecution to prove their guilt. One scholar writing on the case's significance has put it in the following way: This decision is widely regarded as one which introduced a new fundamental principle into the law of criminal evidence. According to that new principle, the accused should not carry the risk of nonpersuasion not only in regard to constituent elements of the crime, but, subject to a few exceptions, also in connection with any defensive issue which he raises. 3 In other words, the Woolmington decision is commonly understood to have established a fundamental principle that an accused does not have to prove their own innocence, and does not (in most cases) have to prove their defence. The decision is occasionally referred to in judgments of the Court of Appeal 4 and the Supreme Court 5 as establishing important principles around the presumption of innocence, but I will argue among other things that it is better understood as something symbolic than as a decision of great doctrinal significance. Lord Neuberger for example, discussing the case in 2015, noted that the principles in Woolmington are now 'taken for granted'. 6 In other words, they are simply part of the common law background. Another argument I will seek to make is that it is helpful in understanding this landmark case to seek to comprehend what these principles actually meant as they were being developed. For a start, were these ideas in any way new?
The Lord Chancellor, Viscount Sankey, argued in his judgment in the House of Lords that he was merely correcting an earlier error, and that the orthodox direction given to Woolmington's jury at the Bristol assizes had been a mistake. 7 There was, in other words, nothing new in the Woolmington decision. The Chief Justice, Lord Hewart, did not offer a dissenting judgment in the Lords, but did write in the Sunday Times three days later that the decision did in fact amount to a significant change in the law. 8 Homicide defences such as the one Woolmington had relied upon had previously carried what we now call a 'reverse burden'meaning it was for the defendant to prove their defence and not for the prosecution to disprove itand now they did not. Another possibility for understanding the case's significance can be found in scholarship which has argued that the Woolmington decision was not a departure from the ideas that had come before, but simply represented a change in these ideas' content. On this account, the main difference with what had come before rested on the specific distinction that should be drawn between justifications, which the prosecution must prove, and excuses, which must still be established by the defendant. 9 In this paper, I do not seek to argue that any of these claims are wrong. There was indeed a doctrinal tensionwhich we shall explore belowthat allowed Viscount Sankey to incrementally develop his way to the Woolmington position, even though on the face of it doing so meant abandoning centuries of homicide doctrine. Lord Hewart was also correct to note that the law of murder was now different to the version that had appeared in legal textbooks and in the practice of the courts before the Lords had 'clarified' the position. And more recent scholarship has tended to agree with the idea that Woolmington is best understood as reflecting a shift in the meaning of the ideas underlying the criminal trial, rather than a shift in any of the more fundamental structures. For Farmer, Woolmington is important because of its status as a case which was 'central to the establishment of a novel linkage between responsibility and guilt', which thereby created a firm link between proof of mens rea and a 3 A Stein 'From Blackstone to Woolmington: on the development of a legal doctrine ' (1993) 6 Lord Neuberger 'Reflections on the ICLR top fifteen cases: a talk to commemorate the ICLR's 150th anniversary ' (2015) https://www.iclr.co.uk/wp-content/uploads/media//2020/09/Transcript_Lord_Neuberger-anniversary-ICLR.pdf (last accessed 27 June 2022). wider commitment to the presumption of innocence. 10 For Loughnan, meanwhile, Woolmington formed part of a general move towards 'amplif[ying] the importance of voluntariness, will, or consciousness for criminal responsibility'. 11 Woolmington, then, is not simply about shifting doctrine. It is also about more fundamental questions of what the criminal law is and what it is for.
The question I want to ask is how it was that this decision was made in this way, at this time, and what those who read and reflected on the decision appear to have made of it. Without wishing to suggest there is anything wrong with other approaches, my instinct is to read Woolmington less as a doctrinal or a theoretical question, and more as a question of legal history. My hope is that shifting the perspective a little in this way can help us to better understand what the case was about, and what it was doing, in 1935. Perhaps doing so could even help us to understand its place for us today, almost a century on.
Prior to the mid-nineteenth century, it was difficult to say very much about the details of criminal doctrine, owing to the absence of any permanent appellate courts. 12 Handler has found that even after the Court of Crown Cases Reserved was established in 1848, little was done in the appellate courts to develop fundamental or universal principles, with much of the reality of criminal law still being left to trial judges to sort out. 13 By the 1930s it was still very unusual for the House of Lords to hear a criminal appeal, and before Woolmington it had never overturned a murder conviction. Engaging in a serious discussion of how the decision in Woolmington came about therefore adds to our understanding of the development of criminal law in England as an area of law with a properly developed doctrinal basis.
This leads on to a further element of what I am seeking to do here in understanding the Woolmington decision historically. Once the Court of Criminal Appeal had rejected Reginald Woolmington's appeal, the only hope he had was for the Attorney General to send his case up to the Lords. This was something that very rarely happened, and indeed this particular Attorney General had recently made clear he was at best ambivalent about what would soon become the Woolmington position. So there was no reason to expect that the Lords would ever even hear the appeal. And yet, of course, they did. In order to understand how the decision came about, we are therefore going to need to reckon with this serious improbability. We cannot do this without exploring Woolmington's growing celebrity, first locally in and around the Attorney General's home town of Bristol, and eventually nationally. In an article which considers Woolmington simply as a question of doctrine or of theory, the reporting of provincial newspapers and so on would be obviously irrelevant. But if we approach the case from this angle, it becomes essential.
I will suggest that situating this case historically means understanding three central issues. First, the significance of Reginald Woolmington's celebrity. Secondly, of the property offences doctrine that was central (without being expressly named) to Viscount Sankey's arguments in the House of Lords. And thirdly, of legal attitudes to judicial innovation in the decades after the first creation of a

Woolmington's trials
In this first substantive section, we will explore Woolmington's trial and retrial. This will be necessary for two reasons. First, it will allow us to see how the law of reverse burdens was understood and applied before Viscount Sankey's judgment was handed down in the House of Lords. Secondly, it will allow us to see how, from the very start, Reginald Woolmington became a celebrity figure during his trial for murdering his wife. This cultural and increasingly political context is essential for understanding how the case ever even ended up at the House of Lords. It is an important part of the history, which is necessarily overlooked by an exclusively doctrinal or principled exploration of the case.
Woolmington was immediately treated with great interest by the local press, and his two trials were widely reported on even before they gained significance as the backdrop to a famous appellate case. Very little of this coverage focused on the nature of the gendered violence the defendant stood accused of. 14 And as we shall see, the press's treatment of Woolmington may have ultimately saved him his life. Already at his first magistrates' court hearing, there were hints of the local prominence the case would soon acquire. According to one report, 'Pouring rain kept all but a few idle spectators from the precincts of the court, and inside not more than a dozen spectators listened to the brief proceedings'. 15 At Woolmington's second appearance before the magistrates, where there is no mention of rain, the court was full: Long before the court sat crowds of people lined the streets of Wincanton, trying to catch a glimpse of Reginald Woolmington … A hush fell over the tiny court room when Woolmington entered in company with a police sergeant. 16 Woolmington was repeatedly presented as a charismatic figure. Unlike in the national papers, which initially reported the case in a straightforward manner, 17 various local newspapers immediately focused on his brown suit, athletic build, and calm demeanour. 18 During his first trial at the assizes, held at Taunton, Woolmington claimed for the first time that the gun that had killed his wife had gone off by accident. On this account, 'the gun was accidentally discharged when he took it from underneath his overcoat where he had it concealed. As far as he was aware, he did not touch the trigger'. 19 This claim was apparently not directly challenged. In subsequent correspondence with the Home Office, the Director of Public Prosecutions explained that this defence had come as a surprise, and expert evidence had had to be commissioned between this trial and the subsequent retrial at the Bristol assizes which would form the basis of the Lords' famous appellate decision. 20 By the time this second trial was heard, expert evidence was able to be adduced by the 14 This is striking given the prevalence at this time of common narratives seeking to explain intimate violence by men against women: R Crites Husbands' Violence against Wives in England andWales, 1914-1939 prosecutionand it was not seriously contradicted by the defenceto the effect that the gun was not in perfect condition, but was unlikely to have gone off by accident.
Returning to the initial trial at Taunton, once Woolmington had offered his unexpected defence, Finlay J had to direct the jury of men 21 on its significance. Specifically, was it for Woolmington to prove the accident, or was it for the prosecution to rebut his claim? Who had the burden of proving this aspect of Woolmington's guilt? As we shall see, Finlay's direction here was quite different to the one subsequently given by Swift J at Bristol. Whether this was because of different judicial temperaments or philosophies, or simply because Woolmington's defence was not one the judge had had time to prepare for, is difficult to tell. Nonetheless, this first judge explained: The case for the prosecution is deliberate shooting. The Defence is: Not guilty of murder. They prove the killing, and in the absence of explanation that is murder. The Defence say: Excusable because accidental. Consider whether you entertain the slightest doubt that this was a deliberate killing … If you have no doubt, it is your duty to convict … If the result of a dispassionate survey is to leave a reasonable doubt in your minds, then your duty as well as your pleasure is to acquit. 22 Woolmington's counsel would later argue that this was the correct formulation. Once the defendant had raised a doubt, it was for the prosecution to dispel that doubt, and it was not for the defendant to fully prove the defence.
When the jury were unable to agree after 65 minutes, Finlay J took them at their word that further deliberation would be meaningless. It was not unheard of for an assize judge to seek an immediate retrial at the same place, in the same session; 23 but this did not seem right to this judge, in this context. The following day, Finlay J was able to confirm the retrial would be heard at Bristol, and that he would not be involved: 'for obvious reasons it is better that the same judge should not undertake the same trial, and I will ask Mr Justice Swift to take the case at Bristol'. 24 Swift had, as an advocate, developed a reputation for presenting simple and persuasive arguments to his juries. As a judge, he had also gained a reputation for giving his juries a clear steer. As his biographer put it in 1939: 'he always made up his mind about a case, and was never satisfied to leave it to the twelve jurors, indifferent which way they found'. 25 In this particular trial, Swift's directive approach led him to advise his jury (again, of men only) 26 that they should convict Woolmington unless he had proved that the gun had discharged by accident. The relevant passage was as follows: Once it is shown to the jury that somebody has died through the act of another, that is presumed to be murder, unless the person who has been guilty of the act which causes the death can satisfy a jury that what happened was something less, something which might be alleviated, something which might be reduced to a charge of manslaughter, or was something which was accidental, or was something which could be justified. 27 21 'Somerset jury disagree: killing of girl wife was an accident?' Western Daily Press and Bristol Mirror (Bristol 24 January 1935) p 7. We know that many women were present in court -'Tragedy of a border town: young Somerset wife shot dead: husband charged with murder: assize jury disagree at Taunton' Western Morning News and Daily Gazette (Plymouth 24 January 1935) p 4but it was common in the southwest of England for women to be excluded from the jury: K Crosby Once the prosecution had proved that Woolmington had caused his wife's death, he must be found guilty of murder unless he could prove his defence. This was a clearly very different direction to the one that had been given by Finlay J at Taunton a few weeks earlier. 28 This second jury convicted Woolmington of murder. He appealed, and was eventually released on the basis that Swift J had misdirected the jury. A large part of the remainder of this paper is about situating the events leading up to the famous decision of the House of Lords within a context that helps us to better understand this decision and the things that made it possible. There are few such contexts more obviously important than the pre-existing doctrine; and it is to this doctrine that we shall turn next, before we eventually start looking at the appeals themselves.

The law before Woolmington
Woolmington raised both doctrinal and principled questions. In addressing these questions, I will start to make out my wider argument that the decision in Woolmington is better understood as an attempt to universalise an existing principle that already existed in another part of the criminal law, rather than as an attempt to set out a fundamental procedural protection for defendants. There are two concurrent areas of the criminal law which need to be explored in order to make out the claim that the Woolmington decision simply extended a pre-existing principle. The first is a shifting attitude towards 'malice' in the law of homicide, and the aspects of a murder charge that were understood to require proof. 29 Second, and perhaps more surprising, is the law relating to the receipt of stolen goods. This second area of doctrine, I argue, made the decisive shift towards a presumption of innocence decades before the law of murder, and ultimately before the criminal law more generally. We shall see that this earlier use of the rule against reverse burdens was more likely to offer a fundamental procedural protectiona rule that if broken was likely to produce an unsafe trialthan we can see in later cases applying the more universal Woolmington principle.
The significance that the law of receiving had in laying the groundwork for the more famous principle in Woolmington was well understood by the lawyers in the case. While Viscount Sankey did not expressly state that he was being informed by these developments, only naming the receiving cases rather than mentioning that they all came from this other area of criminal law, three of the twelve cases cited in his judgment concerned receiving. Counsel for Woolmington also relied in their successful argument at the Lords on another five receiving cases. And in their argument at the Court of Criminal Appeal, they underlined the possibility that the law on burdens in receiving trials ought to be applied to the law of homicide. This pre-history of property offences is not something which is, to my knowledge, explored in any detail in the literature on Woolmington. 30 But its significance as part of the doctrinal backdrop to the decision underlines that an important part of what was going on this case was the universalising of a series of ideas that could already be found elsewhere.
But before we turn to property offences, what was the state of the law regarding reverse burdens in trials for murder? The classic statement here came from the high court judge Sir Michael Foster's 1762 treatise on criminal law. Foster had provided that: In every Charge of Murder, the Fact of Killing being first proved, all the Circumstances of Accident, Necessity, or Infirmity are to be satisfactorily proved by the Prisoner, unless they arise out of the Evidence produced against Him … The Defendant in this Instance standeth upon just the same foot that every other Defendant doth: the Matters tending to Justify, Excuse, or Alleviate, must appear in Evidence before he can avail himself of them. 31 This was subsequently confirmed in the 1837 case of Greenacre. 32 As Viscount Sankey would go on to explain in his decision in Woolmington, Foster and Greenacre were together the standard authorities on reverse burdens regarding defences to murder. 33 Farmer suggests that these homicide authorities should be understood as part of an earlier altercation-based model of criminal trials, 34 which preceded our notions of adversarialism. Even though the rule set out by Foster may have developed out of an older model of justice, which expected defendants to actively (and personally) fight back, 35 it survived well into the newer model in which everyone except the lawyers and the witnesses were effectively rendered silent. 36 As late as 1922, Russell on Crime provided that '[a]s a general rule, all homicide is presumed to be malicious, and murder, until the contrary appears, from circumstances of alleviation, excuse, or justification; and it is incumbent upon the prisoner to make out such circumstances to the satisfaction of the Court and jury'. 37 By the 1920s, as we shall see below, it is in truth possible to discern some judicial retreat even in homicide trials from the principle Foster had set out in 1762. But the turn against reverse burdens had developed earlier, and more decisively, in cases involving receiving stolen goods.
Wigmore, writing from an American perspective in 1905, had complained of the 'troublesome and fruitless' 38 debate caused by the longstanding principle, derived from English law, that a person charged with receiving stolen goods only had to have possession proved against her. The defendant would then be found guilty unless she could show she did not know the goods were stolen. Wigmore cited with approval the 1902 decision in State v Brady, which had held that while a defendant in such circumstances does have to offer some evidence, it is not right to describe her as having the burden of proving her own innocence. 39 This decision was expressly drawn on by the English Court of Criminal Appeal in its 1909 decision in Stoddart, establishing that in a case involving an allegedly fraudulent football lottery 'the jury should have been told that if they accepted the explanation given by and on behalf of Stoddart, or if that explanation raised in their minds a reasonable doubt as to his guilt, they should acquit him, as the onus of proof still lay upon the prosecution'. 40 This principle, not yet set out in anything like the poetic terms of the Woolmington decision, and as yet leaving homicide laws untouched, was next adopted in the law of receiving, with the 1914 decision in Schama and Abramovitch taking its lead from the earlier American jurisprudence. 41 Here, the Old Bailey Commissioner had directed his jury that the defendants were required to 'explain how it came that they were dealing with these stolen goods, and to give an explanation that will satisfy twelve reasonable men'. 42  All this was doubtless a reflection of the earlier commitment within the criminal law to an idea that intention could be inferred from conduct, and did not require our modern idea of a largely subjective mens rea requirement: Handler, above n 13, p 268. 34 Farmer, above n 10, p 62. if the jury think that the [defendant's] explanation may reasonably be true, the prisoner is entitled to an acquittal, because the Crown has not discharged the onus of proof imposed upon it of satisfying the jury beyond reasonable doubt of the prisoner's guilt. That onus never changes, it always rests on the prosecution. That is the law; the Court is not pronouncing new law, but is merely restating it. 43 This decision was made by a panel involving Avory and Atkin JJ, who in the Woolmington appeals would find themselves on different sides. This was not yet a principle which could be confidently applied throughout the criminal law.
It was also not a principle that was consistently applied at first instance, even in trials for receiving. The habit of reversing the burden was a difficult one to shake, with the Court of Criminal Appeal having to reaffirm the principle at least four times over the next half decade. 44 In one such case, even Crown counsel on appeal conceded that the Recorder had 'clearly' failed to engage with Schama and Abramovitch. 45 In another, the Old Bailey Commissioner had attempted to resolve his confusion between what we would regard as evidentiary and legal burdens by stating in court that 'I suppose the phrase "the burden shifts" is too strong, but it is a nice distinction which I do not understand. Well, the burden never shifts, but it does'. 46 By the end of the decade, the Court was regularly finding receiving convictions reached in the absence of a Schama and Abramovitch direction to be unsafe. In this sense, the rule could be regarded as something fundamental: a minimum procedural protection for a defendant seeking to make use of this particular defence. But while the idea was settling into a fundamental principle in the Court of Criminal Appeal, it was at the same time struggling to take hold in the practice of the lower courts. Textbooks, too, did not always grasp the significance of these doctrinal developments. As late as 1922, one textbook still ignored Schama altogether, giving the old principle as good law and simply citing a case from 1803 as authority. 47 How, if at all, did these developments inform the Law Lords' work before 1935? Two years before Woolmington was decided, the judges of the House of Lordssitting in their imperial capacity as the Judicial Committee of the Privy Councilhad made it clear that they were content with these general developments in the Court of Criminal Appeal regarding the burden of proof. In Lawrence v The King, the Attorney General Sir Thomas Inskip (whose views on these matters will be important for us later on) had argued that an improper reverse burden direction was fixed by Crown and defence counsel advising the jury that the prosecution had the burden of proof. There was nothing fundamental about this particular procedural protection, and any breach of the principle could simply be worked around. Lord Atkin disagreed, holding not only that a jury must be adequately directed that the prosecution has the burden of proof, but also that a judge failing to do so will inevitably produce 'a substantial miscarriage of justice … unless it can be predicated that properly directed the jury must have returned the same verdict'. 48 In the context of property offences at least, a judicially imposed reverse burden would virtually automatically create an unfair trial. Again, there was an argument to be had here that what was developing was a fundamental principle, in the sense of a minimum procedural protection for a defendant seeking to rely on one of these defences.
The law of homicide was much slower to incorporate this principle. In his 1913 decision in Davies, for example, Lawrence J held that the defendant did not have to prove his defence in a trial for a gunshot wound to a gamekeeper's shoulder. Slightly frustratingly for our purposes, Lawrence ignored Crown counsel's argument that this would create an illogical tension with the law of murder, where the defendant would be expected to prove his defence. 49 The possibility of a reverse burden 43 Ibid, at 49. 44 R v Aubrey (1916)  would therefore depend on whether the gamekeeper survived, rather than any difference in the defendant's conduct. Two years later, in Hopper, it was held that the jury in a murder trial should have also been directed on manslaughter in circumstances where 'there was some evidence [for the defence]we say no more than that'. 50 This suggests the defendant need not prove her defence, but only had to offer some evidence. In our legal language, this is an evidential, and not a full legal burden. By 1915, the Court of Criminal Appeal had started to hint that the eighteenth century rule in Foster may need to give way. But compared to its judgments elsewhere, its rulings on reverse burdens in homicide trials were noticeably more tentative.
By the time Woolmington was decided, then, the appellate courts had made clear their commitment to a general idea that the prosecution had the burden of proof. But it was well established, and confirmed as late as 1910, that this would not apply to insanity defences for example; 51 and there was little indication beyond a hint in Hopper that any of this might be used to undermine the centuries-old rule set out by Foster. The underlying concepts were there for the Woolmington decision, but it was still far from clear that we were about to be given a universalising principle like Viscount Sankey's 'golden thread'. So how exactly did we get there?

The Court of Criminal Appeal
The early stages of the Woolmington appeal, as opposed to the judgment of the House of Lords, have not previously been discussed in the literature. But it is my contention that they are vital for helping us to understand how exactly it was that Viscount Sankey's landmark judgment came about, and how it was that it took the form it did.
The Woolmington trial and appeal existed in two parallel settings, both as a series of legal decisions and as a moment in popular culture. Woolmington was convicted at Bristol on 14 February, and by 20 February his mother, Maria, had sent the Home Office a petition, signed by over 7,000 people, asking for mercy for her son. 52 At this stage the campaign appears to have been limited to the area in which the Woolmingtons had lived, with a preliminary letter sent on 19 February describing the petition as coming 'from the countryside where my boy lived'. 53 Two days later, the magistrate that Woolmington had initially appeared before also wrote to the Home Office requesting leniency, and referring both to the Taunton jury's disagreement, and to his sense of the general feeling within the local community. 54 By 2 March, an article on Woolmington's case had appeared in the populist John Bull magazine, arguing that retrials must never take place in murder trials unless the prosecution had substantial new evidence: in the absence of such new evidence, any retrial necessarily undermined the idea that a prisoner's guilt must be proved beyond reasonable doubt. 55 In a little over two weeks, Woolmington's celebrity, and the popularity of his cause, had started to shift from the local to the national.
When Woolmington's lawyer, JD Casswell, put in his notice of appeal, he was quickly called in for a lunch with Swift J. Recalling the events a quarter of a century later, Casswell remembered feeling 'a little apprehensive', 56 but Swift's biographer insists that the judge was fairly good-natured about the situation. 57  at one point even independently of any solicitors' instructions. And as we have already seen, there was by this stage a clear shift on the level of principle regarding the legitimacy of reverse burdens. But this shift had never before been definitively represented in the law of murder.
Avory J, who gave the judgment of the Court of Criminal Appeal, was widely regarded as the judge with the greatest expertise in criminal law. 59 He was described in Swift's biography as 'that doyen of criminal judges'; 60 and one of Avory's biographers stated 'Horace Avory, perhaps more than any other figure in English legal history, represented the lay conception of the perfect legal machine, just, impartial and a little inhuman'. 61 According to the same author, the Chief Justice Lord Hewart had once described him as 'a legal encyclopaedia'. 62 Casswell was less impressed by the judge, recalling 25 years later that Avory had never been a judge favourably disposed towards defendants. Now in his eighty-fourth year and, as it transpired, within months of his death, he was implacably pro-prosecution. He would never admit that Archbold, commonly called 'the prosecutor's bible,' could possibly be wrong. 63 And in fact Avory had himself authored the section in Halsbury's Laws of England which gave substantially the same account of the law as that which was provided in Archbold. 64 Avory's temperament, and settled view, were both against extending the law on reverse burdens from property offences to murder.
But whatever their principled views, the doctrine posed clear problems both for Casswell and for Avory. For Casswell, it was clear that the rule stated in Foster was the law, in the sense of being the rule consistently applied by the courts. Swift himself had previously had directions along the lines he had given in Woolmington upheld by the Court of Criminal Appeal (in fact, as recently as 1927, nobody in one such appeal had even thought it worthwhile to pause on the legitimacy of this part of his direction). 65 Nonetheless, Avory must have been aware that the decisions of the courts were increasingly hostile to reverse burdens. Indeed, he had been one of the judges in Schama and Abramovitch. Frankly, a judgment either for or against Swift's direction at Bristol was possible at this stage, given the principles and rules then available.
How, then, did Caswell seek to persuade this pro-prosecution encyclopedia? His first argument was that the Court should hear additional evidence regarding the gun. No evidence had been offered on this point at Taunton, and at Bristol the defence had not seriously objected to the prosecution evidence. This evidence had provided that the gun was not in excellent condition, and was perhaps a little sensitive, but was essentially safe. The prosecution had only advised Casswell they would be calling this evidence two days before the trial, and so, he now argued, he had not had adequate time to prepare a response. 66 Now he sought to adduce new expert evidence suggesting that the gun was completely unsafe. This expert, an international shooting champion, had contacted Casswell immediately after the trial, on seeing reports of the other expert's evidence. The new expert was present in the vicinity of the court, but having briefly conferred the three judges concluded it would not be appropriate for them to hear from him. 67  questions of fact; although the Lords would subsequently have a look at the gun (in the absence of any expert guidance) before reaching their decision.
Casswell's main doctrinal argument was that Swift had misdirected his jury. Rather than directing that Woolmington had to prove his defence, he should instead have directed that there was a rebuttable presumption of malice. Once Woolmington had adequately raised his defence, the burden should have reverted to the prosecuton to prove his guilt beyond reasonable doubt. Casswell drew in his argument on Stoddart, Schama and Abramovitch, Aubrey, Grinberg, and Hagan. 68 These cases, as we have already seen, are suggestive of a growing judicial unease regarding reverse burdens, but were restricted at this stage to property offences. Casswell sought to justify the leap to murder by pointing to the relative stakes: 'If that is the law, where the liberty of the subject only is involved, it cannot be otherwise where the life of the subject is involved'. 69 He also argued that the question of what a properly directed jury would have done was not a hypothetical one: that the correctly directed jury at Taunton had been unable to conclude that Woolmington was guilty. 70 Of course this jury had not had the benefit of the expert testimony regarding the gun. In truth, it seems that failures of disclosure on both sides had made the facts about the gun difficult to ascertain, with neither side ever really having had an opportunity to mount a satisfying courtroom response to the other side's claims.
JG Trapnell KC, responding for the Crown, sought first to argue that Swift's direction had not truly imposed a reverse burden. The judges understandably met this point with some scepticism. 71 Trapnell next developed an intricate argument for conceptual differences between property offences and trials for murder. He argued that intent required positive proof in trials for receiving, and it therefore made sense to require the prosecution to prove it; but in trials for murder intent would be presumed, and therefore it made sense to require a defendant to disprove that presumption. 72 Finally, Trapnell pointed to the fact there were significant differences between the two trials, notably the absence of the expert evidence at Taunton, meaning the Court could not rely on the first jury's verdict when considering the consequences of a differently worded direction. 73 Giving the judgment of the Court, Avory J started by explaining ('in his dry, crackling voice') 74 that the first verdict must be disregarded, as it was impossible to know why the jurors had been unable to agree. 75 He next suggested '[i]t may have been better' if Swift J had directed the jury in the way Casswell had wanted. 76 At the same time, Swift's direction had been in accordance with the settled doctrine, supported by 'ample authority'. 77 This represented a clear tension. But this tension was sidestepped by the judge, who noted that the Criminal Appeal Act required the Court to only uphold an appeal if they believed a 'substantial miscarriage of justice has actually occurred'. 78 Having surveyed the evidence, Avory concluded that 'even if any such warning had been given to the Jury about a reasonable doubt which is said to have been omitted, they must in the circumstances of this case have inevitably come to the conclusion that the Prisoner was guilty of the crime of murder'. 79 In other words, even if there was a rule against reverse burdens in the context of a murder trial, it was not a fundamental principle: whether a conviction was actually unsafe would depend on the wider circumstances of the trial. This, of course, was the argument that the Attorney General had made, and the Privy Council had rejected, two years earlier in the context of a property offence. Avory concluded that the Court did not have to decide whether the reverse burdens principle that had been developing for decades in the context of property offences also applied in trials for murder, as the result for this defendant would have been the same either way. This conclusion echoed the preference for the certainty of a jury's verdict over the precision of well worked doctrine that had long been a feature of appellate decisions in English criminal law. 80

The House of Lords
Woolmington's trials were from the outset the subject of great local interest. Five days after the Bristol trial, Maria Woolmington had sent the Home Office a seven-thousand strong petition for mercy from the countryside near Milborne Port. 81 Six weeks later, as the Lords heard the case, she presented a petition signed by another seven thousand people, this time from Bristol. 82 Meanwhile, ER Pursey, an international clay pigeon shooter who the Court of Criminal Appeal had refused to hear from, had written to the Home Office, setting out his view that the gun was 'most definitely unsafe, and it would be very inaccurate to describe it as reasonably safe. It is in fact a gun in such condition that if it got into the hands of a gunsmith he would never let it leave his premises again'. 83 The DPP privately conceded the strength of this evidence. 84 The Illustrated Police News reported that, on realising his evidence would never be heard in court, Pursey set off to collect signatures in his home town of Bristol. 85 By this time members of the public from further afield had also started to lobby the Home Office for mercy for Woolmington. 86 In order for a case to proceed from the Court of Criminal Appeal to the House of Lords, the Attorney General's authorisation was required. We have met the Attorney General, Sir Thomas Inskip, once before in this paper. In Lawrence, decided two years before Woolmington, Inskip had conceded that the prosecution had the burden of proof, but argued that directions failing to advise the jury of this jurisprudential principle would not necessarily be appealable. Inskip was not on the face of it a committed proponent of reverse burdens. It was also unusual for the Lords to hear murder appeals, and The Times observed they had never previously upheld an appeal against a murder conviction. 87 Finally, Inskip's own advisers were clear that Woolmington's case was hopeless: There can be no doubt that the learned Judge set out to the jury the present state of the law, and in order therefore to justify you in issuing your certificate that an Appeal should be taken to the House of Lords you will no doubt require some authority that the textbooks, and the cases cited in support thereof, are wrong. 88 The best Inskip's adviser could offer was that there was a vague principle against reverse burdens; and that in 1910 the Court of Criminal Appeal had had regard to an earlier jury's disagreement in an appeal concerning a retrial. 89 80 So if Inskip's advisers were clear with him that Woolmington's case could not succeed, and if Inskip himself was unconcerned about the principled sanctity of the burden of proof, how exactly did the case reach the Lords? Why was it that Inskip's advisers framed their memo in terms of 'justifying' his decision that the appeal should go ahead? One factor was doubtless Woolmington's lawyer, Casswell, who insisted on meeting Inskip to discuss the case, despite the fact the appeal was over and he was therefore no longer being instructed by solicitors. 90 I have been unable to find any description of their conversation, but perhaps Casswell was simply very persuasive. More likely, I would suggest, is the growing public support for Woolmington at Bristol. Inskip was born and had grown up in the city. He had between 1918 and 1929 been one of the city's three MPs. 91 His half-brother had been the city's Lord Mayor in 1931. 92 It seems unlikely that the growing public pressure in the city, as evidenced by the seven-thousand strong petition being gathered as he made his decision, did not form part of his highly improbable decision to send the case to the Lords.
Setting out their cases in the Lords, Casswell (now working with TJ O'Connor after the Lord Chancellor's insistence that there should be a KC on each side) 93 and Trapnell did not drastically change their arguments from the ones offered at the Court of Criminal Appeal a little under a month earlier. O'Connor's central claim was both that the burden of proof never normally shifted to the defendant, and that, contrary to what Avory J had decided at the Court of Criminal Appeal, any misdirection on this point was sufficient to ensure that 'the rest of his summing-up was vitiated'. 94 The presumption of innocence was both a universalisable, and a fundamental principle. Trapnell, meanwhile, contended that the law as set out by Foster and applied by Swift was correct, 95 and that even if it wasn't then no jury could have failed to convict Woolmington on the evidence submitted. 96 No mention was made by either side about the expert evidence (although the Lords did take a look at the gun). 97 These arguments were heard across 4 and 5 April, and Viscount Sankey immediately announced that the appeal would succeed. His reasons would be provided some seven weeks later.
Early in his judgment, Sankey asked whether the statement of the law in Foster's treatise was correct. Not whether it was still valid, but simply 'Is that statement correct law?' 98 The first part of Sankey's answer to that question was to note that the Court of Criminal Appeal was not established until 1907: while the judges had previously considered matters of criminal law, these earlier discussions either engaged with issues that could be distinguished from the matters under consideration here, or failed to secure for the law of evidence very much more than a 'fluid condition'. 99 There were also important questions about the accuracy of the reporting in the various nominate reports, drawn up before the advent of professional law reporting. 100  Woolmington v DPP [1935] AC 462, at 475. 99 Sankey did recognise that the Court of Crown Cases Reserved had been established a few decades earlier, but argued that even then most pronouncements regarding the criminal law were to be found in trial judges' directions to their juries, and in textbooks. Neither source, of course, had the authority of an appellate court decision. Ibid, at 476-478 (quote at 478). 100 Ibid, at 478-479.
1909 had a weight that cases such as Greenacre, decided almost a century earlier, could not have, even though the former concerned property offences and the latter concerned homicide. This was a shrewd way of establishing that the principles found in the modern receiving cases, rather than the principles found in the older murder authorities, had the greater practical significance.
Having explained that the earlier doctrine on murder was of dubious significance compared to the recent property offences doctrine, Sankey next went on to consider the recent work of the Court of Criminal Appeal, starting with Stoddart, which had directly, reliably, authoritatively engaged with questions of the burden of proof. These authorities relating to property offences, having been carefully set up as the only authorities that were really available, inevitably led to the conclusion that Swift J's direction at the Bristol assizes had been wrong.
But did this matter? The principle against reverse burdens was universalisable, because there were no valid authorities other than those supporting the principle, but would a breach of this principle necessarily produce a miscarriage of justice? Here, Sankey was less clear, concluding '[w]e cannot say that if the jury had been properly directed they would have inevitably come to the same conclusion'. 101 This was less the establishment of a clear test, and more a rebuttal to Avory J's use of the word 'inevitable' at the Court of Criminal Appeal. But as we shall see in the next section, Woolmington was immediately understood as offering very little by way of procedural protection to a defendant.
One striking aspect of the judgment in the House of Lords is that there was in fact only one judgment, with the Lord Chancellor's reasoning simply being assented to by his colleagues. As we shall see later on, this unity was at best only apparent, with one judge taking to the pages of the Sunday Times only a few days later to offer a very different account of what had happened in the case. But despite the low numbers of criminal cases that it heard during this period, it can be seen that it was not unusual for the Lords to give a unitary judgment in this sort of case. In its seven reported decisions concerning English criminal law during the interwar period, the Lords delivered a unitary judgment on four occasions, 102 with this judgment being delivered by the Lord Chancelloras it was herethree times. Separate judgments were only given in order to demonstrate disagreement, with all or all but one Law Lord on the panel giving a full judgment either when there was a dissentient, 103 or where a particularly complex question of statutory interpretation was involved. 104 Generally speaking, the House of Lords wanted in its criminal decisions to project certainty, with any minor disagreements ordinarily being smoothed out in public.
But what did that certainty entail? The principle against reverse burdens was now understood as a universal principle, rather than one derived from an application of an academic critique of the American law of evidence to English property offences. But there were exceptions built into the judgment. Sankey allowed that a defendant would still have to prove a defence of insanity, in part because the extrajudicial, Victorian M'Naghten principles had been confirmed and explained in a 1910 decision of the Court of Criminal Appeal. 105 And he briefly endorsed the concession made by Woolmington's lawyers that reverse burdens would be possible when created by statute. 106 But what he did not do is conclude that any judicial failure to properly direct on the burden on proof would necessarily lead to a miscarriage of justice. No attempt could be made to whittle down the broad principle, but what the judgment ultimately did was limited to confirming the principle's breadth. Finally, the case was not creating new law, because it was impossible to find an authoritative statement on reverse burdens in homicide cases prior to 1907. Hewart's article was simply an accessible account of the issues explored in Woolmington. But he concluded by underlining that the decision did constitute a clear change in the law, quoting with approval the following words of the jurist Henry Maine: When a group of facts comes before an English Court for adjudication, the whole course of the discussion between the judge and the advocates assumes that no question is, or can be, raised which will call for the application of any principles but old ones … Yet the moment the judgment has been rendered and reported, we slide unconsciously or unavowably into a new language and a new train of thought. We now admit that the new decision has modified the law. 121 This, no doubt, is the passage that Inskip had found 'inconvenient'.
The Woolmington decision had an immediate impact as the expression of an abstract idea. The legal press praised the decision for favouring underlying principles over longstanding doctrine. 122 American and English commentators noted that the case showed how quickly established doctrine can be undone. 123 In Australia, Sankey's judgment had 'greater eloquence than historical accuracy', 124 but set out a humane principle which should be further developed so as to include the law on insanity. 125 Comparative lawyers, meanwhile, expressed their hope that the Woolmington principle might be extended into colonial legal codes which still contained the old formula. 126 Judges also occasionally saw Woolmington as the start of a discussion about high-level principles. One lawyer drew on the case both judicially and extra-judicially as part of his more general thinking about the relationship between burdens and standards of proof. 127 Similarly, in Carr-Briant, where the Court of Criminal Appeal faced conflicting authorities on the burden an accused must bear when faced with a statutory reverse burden, it was held that the logic of the Woolmington principle required the defendant to have something less than a criminal burden. 128 Woolmington was immediately understood as expressing important points of principle.
But how was the case received within the courts in its more immediate doctrinal setting, in the first few years after 1935? Tracing things through for the first decade, until the end of 1945, I have found two occasions where Woolmington was used as a simple homicide decision, rather than a source of any great principle. 129 In five cases, counsel are reported as drawing on Woolmington, but then the judgment of the court declines to engage with it directly. 130 We have already seen that there were suggestions in some places that the Woolmington decision should have gone further, also placing the legal burden on the prosecution in the context of the insanity defence, for example. Smith points out that even after Woolmington defendants were still expected to prove the common law defences of selfdefence and of duress. 131 Beyond this, there was also the question of how significant the case actually was, when it only referred to reverse burdens created at common law. By 1946, the Law Journal noted 121 that Woolmington was not a case with very much to say about the growing number of reverse burdens created by statute. 132 And I have only found two cases to the end of 1945 where a direction was found to have breached the Woolmington principle in ways that actually mattered to the outcome of an appeal. 133 The one way in which Woolmington was engaged with doctrinally regarded the question of how s 4 of the 1907 Act should be interpreted. When Sankey had talked of another jury 'inevitably' reaching the same conclusion even if properly directed, had he meant that almost all breaches of the Woolmington principle would result in a quashed conviction? As early as 1936, Lord Hewart held that the absence of a Woolmington direction was unimportant where the jury would have convicted anyway. Woolmington was 'a case to which far too much importance had been attached'. 134 Also in 1936, the Privy Council twice decided that breaches of the Woolmington principle did not require a conviction to be quashed. 135 Five years later, in Mancini, the House of Lords explained that a Woolmington direction would not be required in every case. 136 In 1944, the Court of Criminal Appeal (subsequently confirmed by the Lords) 137 added that s 4 had since 1909 meant that a conviction would only be overturned if a jury could have 'reasonably' voted to acquit. 138 This was not changed by Woolmington. 139 The Woolmington principle was therefore not fundamental: a judge who failed to direct his jury accordingly would probably not thereby create a miscarriage of justice. This was a change from how the principle had been understood in its earlier, more narrow doctrinal setting as part of the law of receiving stolen goods.

Conclusions
Woolmington, then, was immediately deemed significant, but in limited ways. The popular pressure which had secured Woolmington his hearing at the Lords continued for a time. But when he failed to become a star, interest waned. The case had a more lasting significance as an exemplar of principles contained within the criminal law, even if commentators were reluctant to acquiesce to Sankey's insistence that nothing the Lords had done was new. The case, understood at the level of principle, was taken up in developing arguments about what the law stood for, and where it should go next. This is the kind of significance that the Woolmington decision arguably still has for us today. As a precedent that could actually be relied upon by those seeking to appeal against their convictions, however, the decision was less longlasting. Its breadth did not extend to statutorily created reverse burdens, nor to the common law defence of insanity. And it was not fundamental in the sense of offering a minimum procedural guarantee: it is very difficult to find the courts quashing convictions on Woolmington grounds in cases decided in the first decade after the decision, even where it was conceded that the jury was actually misdirected.
Woolmington was immediately a case with deep symbolic importance, laying down a universal principle within English criminal law. But at the same time it was doctrinally narrow, going no further than extending a rule that had developed primarily in a context of property offences to a new context of defences to murder. It was, in this sense, blandly incremental. It set out a principle that, for all its universalisable breadth, never offered anything fundamental, in the sense of a minimum set of standards that all correct judicial directions must abide by. It was a decision that was never intended to, and never did stop the growth of statutorily created reverse burdens. It was a decision that could not 132 'Burden of proof and statutory offences ' (1946)  have been predicted in advance, abandoning as it did centuries of settled doctrine. It was also a decision that probably never would have reached the Lords, if not for the arbitrary, contingent fact that an Attorney General who was at best ambivalent about the importance of what would become the Woolmington principle was nonetheless sensitive to popular and political pressures coming from his hometown, a hometown in which Woolmington had happened to have been tried.
Woolmington, ultimately, is a decision that matters less for its practical consequences than for the ideas it represents. But by tracing through what those ideas actually meant at the time they were set out, it is clear that they were both ambitious, universalising points of principle, while at the same time being carefully constructed as something less than fundamental. Woolmington was, depending on your mood, either a way of producing quite mundane and unimportant theoretical consistency, or a way of setting out an ambition for where the criminal law might go next. But what it was not was an expression of a fundamental procedural right.