United Kingdom House of Lords Decisions

LORD CHANCELLOR (Viscount Kilmuir).—[His Lordship gave the narrative quoted supra, and continued]—

The Lord Ordinary (Lord Guest) and the learned Judges of the First Division held that the first respondents were in breach of duty at common law in failing, contrary to the proved practice of making them available, to provide a safety belt for the deceased, and that the second respondents were in breach of their statutory duty in the same respect. The Lord Ordinary and the Lord President (with whom Lord Carmont agreed) went on to hold that the appellant had failed to prove that the provision of a safety belt by the respondents would have prevented the accident, while Lord Guthrie, taking a different view of the onus of proof, held that the respondents had proved that it would not. The basis of these views was that, if safety belts had been provided, the deceased would not have been wearing one on the occasion of the accident.

Lord Guest held that the question of instructions to use a safety belt was academic. The Lord President took the view that exhortation would have been useless, and that, if instructions had been given, the deceased would have ignored them.

On the first point, the appellant's case before this House was fourfold. It was submitted that to hold that the appellant had failed to prove that the provision of a safety belt would have prevented the accident (and still more that the respondents had proved that it would not) was wrong on grounds of authority, the type of evidence required as compared with that called, the effect of the evidence before the Lord Ordinary and a theory of causation.

The first case to which reference was made was Roberts v. Dorman Long & Co. Ld. That was a case under Regulation 97 of the Building (Safety, Health and Welfare) Regulations, 1948, of which the material words are:—

"… and except for persons for whom there is adequate handhold and foothold … there shall be available safety belts … which will so far as practicable enable such persons who elect to use them to carry out the work without risk of serious injury."

At p. 946 Lord Goddard, C.J., is reported as saying:

"I think that if a person is under a duty to provide safety belts or other appliances and fails to do so, he cannot be heard to say: ‘Even if I had done so they would not have been worn.’"

At pp. 949 and 951 Birkett and Hodson, L.J. (as they then were) used language which, it was submitted to us, supported this view. In the case of Drummond v. British Building Cleaners Ld., Parker, L.J. (as he then was), at p. 1444, adopted Lord Goddard's words and reasoning.

With respect, I am unable to follow or accept this reasoning or its result. The necessity, in actions by employees against their employers on grounds of negligence, of establishing, not only the breach of duty, but also the causal connexion between the breach and the injury complained of, is, in my view, part of the law of both England and Scotland. It was emphasised in this House in the case of Wardlaw v. Bonnington Castings Ltd. [1956] AC 613, I refer to, without quoting, what was said by Lord Reid, Lord Tucker and Lord Keith of Avonholm. Their words made perfectly clear that the principle applied whether the claim was based on the breach of a common law or a statutory duty. This principle was, in my respectful view, correctly applied by Paull, J., in the first part of his judgment in Nolan v. Dental Manufacturing Co. Ltd. Counsel for the appellant was not able to suggest to us a principle of law which would not only place the onus of establishing that the breach of duty was not the cause of the accident on the employers but also, in Lord Guest's words, would preclude the employers from doing so.

It was urged on behalf of the appellant that to desiderate, in the case of a fatality, that the deceased would have used the safety device was to impose a burden on the pursuer which she could never discharge, since, without the deceased's evidence, it must be impossible to prove what the deceased would have done in a particular hypothetical situation which did not arise in fact. The argument, therefore, took the form that Lord Guthrie was right when he said that the onus was on the respondents, and that the onus could only be discharged by direct evidence from the deceased as to what he was going to do. I cannot see any grounds for such a limitation on the type of evidence. What the deceased was going to do can be a matter of inference from appropriate facts. Further, and with respect, I do not agree with Lord Guthrie that, in this case or generally, the onus is on the employers. I am not prepared to say that there can never be a case where the nature of the safety device and the obvious correlative duty of any reasonable employee to use it are such that in the absence of other evidence it should be inferred that he would have used it if it had been available: but that is not this case.

The evidence demonstrates to a high degree of probability that, if safety belts had been available, the deceased would, in any event, not have worn one. On this aspect the Lord Ordinary and the learned Judges of the First Division found in favour of the respondents and rejected the appellant's contention. There were a number of witnesses called for the appellant and for these respondents with wide experience in structural steel operations, including, in some instances, work on tower cranes such as that in which the deceased was engaged. The combined effect of the evidence was that steel erectors never wear safety belts except in certain very special circumstances which do not include the erection of scaffolds for riveters on tower cranes. No witness deponed to having ever seen a safety beltworn in the course of such work, and there was ample evidence from these respondents' employees and from others that safety belts were not worn when such work was being carried out. One witness spoke of having seen the deceased wearing a safety belt on one or possibly two occasions, when doing an operation of a peculiar and special nature. The Lord Ordinary did not accept his evidence on that matter, which, in any event, was not corroborated. There was overwhelming evidence that the deceased did not normally wear a safety belt and, in particular, it was proved that he had been engaged in erecting riveters' scaffolds on the crane from which he fell, at heights greater than that from which he fell and at times when safety belts were available and that he had not on such occasions worn or asked for a safety belt.

In my opinion, it was clearly open to a Court to infer that the deceased would not have worn a safety belt, even if it were available.

Finally, it was submitted that, if the deceased's refusal to wear a safety belt must be recognised as the effective cause of his not wearing one, and hence of his death, the failure of the respondents to provide a safety belt should not be ignored as a causative factor. The answer, in my view, must be that there are four steps of causation: (1) a duty to supply a safety belt; (2) a breach; (3) that, if there had been a safety belt, the deceased would have used it; (4) that if there had been a safety belt the deceased would not have been killed. If the irresistible inference is that the deceased would not have worn a safety belt had it been available, then the first two steps in the chain of causation cease to operate.

On the submission that the first respondents should have exhorted or instructed the deceased to use a safety belt I have considered carefully the argument based on the extent of the danger. I have, however, come to the conclusion that it fails. There was a strong feeling among steel erectors that safety belts were certainly cumbersome and might be dangerous, except in very special circumstances, which did not obtain here. This being so I bear in mind the words of Lord Radcliffe in Qualcast (Wolverhampton) Ltd. v. Haynes :

"One [word of caution] is that, though indeed there may be cases in which an employer does not discharge his duty of care towards his workmen merely by providing an article of safety equipment, the Courts should be circumspect in filling out that duty with the much vaguer obligation of encouraging, exhorting or instructing workmen or a particular workman to make regular use of what is provided."

Moreover, I cannot exclude from my mind the fact that Regulation 97 to which I have referred contains the words "such persons who elect to use them." It would, in my view, be wrong to say that, in such circumstances, there is a duty on the employers to exhort or instruct.

I would dismiss the appeal.

VISCOUNT SIMONDS .—The appellant claims to recover damages from the respondents jointly and severally in respect of the death of her husband in an accident which occurred on 27th January 1956. Her claim is based on the negligence of the first respondents—a common law claim—and a breach by the second respondents of their statutory duty under section 26 (2) of the Factories Act, 1937. She has failed in her claim before the Lord Ordinary and the First Division of the Court of Session and, however much one may sympathise with her in her loss, I think that there can be no doubt that she must fail again.

The deceased was a steel erector of many years' experience and was employed by the first respondents in erecting a steel tower for a crane for the use of the second respondents in their shipbuilding yard at Port Glasgow. Whilst so employed, he fell from a height of about seventy feet and was fatally injured. I need not describe in detail the nature of his work. It was dangerous work, as all such work must appear to the layman, but it was not specially dangerous work. It was the work which he had been accustomed to do for many years. On the day of the accident he was not wearing a safety belt. It was proved that a safety belt was not on that day available for his use, if he had wanted to use it. A belt had been available until two or three days before the accident, but then had been removed together with the hut in which it had been stored to another site. It is a matter of conjecture whether the deceased knew that it had been removed.

In these circumstances, the simple case was made that the respondents were in breach of their duty to provide a safety belt for the use of the deceased; he was not wearing one when he fell to his death; if he had been wearing one, he would not or at least might not have fallen: therefore, the respondents are liable.

To this simple case the respondents make answer. Let it be assumed that they were in breach of their duty in not providing a safety belt on the day of his accident, and, further, that, if he had then been wearing one, the accident would not have happened. Yet there is a missing link: for it was not proved that the deceased was not wearing a belt because it was not provided; and, alternatively, if any question of onus of proof arises, it was proved that, if one had been provided, the deceased would not have worn it.

I do not doubt that it is a part of the law of Scotland, as it is part of the law of England, that a causal connexion must be established between a breach by an employer of his duty at common law, or under a statute, and the damage suffered by his employee—see, e.g., Wardlaw v. Bonnington Castings . If a contrary principle is thought to be established in Roberts v. Dorman Long & Co. Ld. I cannot reconcile that case with Wardlaw. It may, however, be said that, where the employer is in breach of his duty, there is in that fact some prima facie evidence of a causal connexion between the breach and the subsequent damage. So far in this case I would go with the appellant. It is the next step that I cannot take. For, it having been found as a fact by the Lord Ordinary, and their Lordships of the First Division having unanimously concurred in that finding, that it would be totally unrealistic to hold that the failure to provide a belt was the cause of the accident, the learned counsel for the appellant was driven to the argument that the evidence on which that finding was based was inadmissible, or at any rate of no weight. This argument I cannot accept.

The evidence showed conclusively that the deceased himself on this and similar jobs had, except on two special occasions (about which the evidence was doubted by the Lord Ordinary), persistently abstained from wearing a safety belt and that other steel erectors had adopted a similar attitude. Nor was their attitude irrational or foolhardy. They regarded belts as cumbersome, and even dangerous, and gave good reason for thinking so. It was, however, urged that on this single occasion the deceased might have changed his mind and that the respondents did not and could not prove that he had not done so.

I would agree that, just as a claim against a dead man's estate must always be jealously scrutinised, so also an inference unfavourable to him should not be drawn except upon a strong balance of probability. But there is justice to the living as well as to the dead, and it would be a denial of justice if the Court thought itself bound to decide in favour of the deceased, because he might, if living, have told a tale so improbable that it could convince nobody. That is this case and, in my opinion, the Courts below were amply justified in receiving the evidence given (not only by the respondents' witnesses) as to the attitude adopted by the deceased and other steelworkers to the wearing of belts, and acting upon it.

Another aspect of the appellant's case deserves mention. It was urged that the respondents were at fault in that they not merely did not provide a safety belt for the use of the deceased on the fatal day, but also that they had not instructed him to use it. It was pointed out by the Lord Ordinary that any instruction on that day was out of the question, for the belt was not there. The failure must, therefore, relate back to an earlier period. The only meaning that the contention can bear is that, if during his period of service the deceased had been instructed to wear a belt, he would have done so, and, in such a case, a belt would or should have been available for his use on the fatal day. My first difficulty in accepting this contention lies in the ambiguity of the word "instruct" which learned counsel did little to resolve. If the word means "teach how to use" instruction was superfluous, for no one could know better than the deceased how to use a belt. If it meant "direct" or "order," there was certainly a failure to do so; but counsel did not advance this meaning. In face of the language of Regulation 97 of the Building (Safety, Health and Welfare) Regulations, 1948, and of the widely, if not universally, established practice of steel erectors, it would have been impossible to contend that it was the duty of the respondents to order the deceased to wear a belt. We are left with some intermediate meaning which can best perhaps be conveyed by the word "exhort." But I cannot see that anything is involved in that beyond pointing out the possible advantage of wearing a belt, which needed no pointing out, and decrying the disadvantage of it, with which the deceased and his fellow workmen did not agree. As the Lord President said in his opinion, with which I agree:

"Indeed some of the witnesses stated they would have had something to say to anyone who insisted on their carrying safety belts about on these jobs."

They might have listened more politely to exhortation than direction. The result would have been the same.

I have not thought it necessary to deal separately with the case of the second respondents. What I have said as to the necessity of a causal connexion between the breach of duty and the subsequent damage applies mutatis mutandis to them.

The appeal should, in my opinion, be dismissed.

LORD REID .—The appellant is the widow of William M'Williams, a steel erector, who was killed on 27th May 1956, when he fell from a steel tower which was being erected in a shipyard occupied by the second respondents. The first respondents were his employers. M'Williams was setting up a working platform for riveters on the outside of the tower about seventy feet from the ground. This had to be placed on "needles," which are battens projecting some four feet from the tower. They were secured to the tower by lashings. A lashing of one of the needles was not properly fixed so that when the deceased put his weight on this needle it tilted, and he fell to the ground. It is not clear whether he was responsible for not fixing it properly, or not inspecting it, and in this action no fault is alleged against the respondents with regard to the needle.

The case made by the appellant is that both respondents were at fault in not providing safety belts. These belts have about fifteen feet of rope attached to them, so that the end of the rope can be tied to some convenient part of the structure near where the man is working; then, if he falls, the rope prevents him from falling more than its length. It is not denied that, if M'Williams had been wearing a safety belt, when he fell, he would not have been killed. The employers do not deny that it was general practice to provide such belts, but they do not admit any duty to provide them. The Courts below have held that they had this duty, and also that, by reason of the shipyard being a factory within the meaning of the Factories Act, 1937, section 26 (2) of that Act required the second respondents to provide these belts. I need not consider whether this was right, because the main defence of both respondents is that, if such belts had been available on the day of the accident, M'Williams would not have worn one and, therefore, any failure to provide a belt was not the cause of his death. I shall assume in the case of both respondents that they were in breach of duty in not providing belts.

There can be no certainty as to whether the deceased would, or would not, have worn a belt on this day, but the defenders maintain that it is highly probable that he would not. Work on this tower had been proceeding for many weeks, and, at least for a good part of that time, he had been doing work similar to that which he was doing when he fell. Throughout this period, safety belts had, to his knowledge, been available in a hut nearby, and it is clear that it was not his practice to wear a belt. Steel erectors were neither required nor exhorted to wear belts, and several witnesses with long experience say that they had never seen any steel erector wear a belt, and, in particular, that they had never seen M'Williams wear one. And there is evidence that the condition of the belts showed that they had seldom, if ever, been used. But one witness says that he saw M'Williams wearing a belt on two occasions when working in an exposed position. The Lord Ordinary thought this extremely doubtful, but I am prepared to assume in the appellant's favour that this evidence can be accepted. It was left to the discretion of each man to decide whether to wear a belt, and it appears that the reason why belts were not generally worn was not mere prejudice against them. They are cumbersome, and some witnesses say they might be dangerous in certain circumstances.

For some reason the belts were taken away to another site two or three days before the accident. So, after that, the defenders were in breach of their duty to provide belts. We do not know whether the deceased knew that they had been removed, and there is nothing to suggest that, during those two or three days, he may have considered changing his normal practice not to wear a belt. So it appears to me to be a natural, and indeed almost inevitable, inference that he would not have worn a belt on this occasion, even if it had been available. And that inference is strengthened by the general practice of other men not to wear belts.

It was argued that the law does not permit such an inference to be drawn because what a man did on previous occasions is no evidence of what he would have done on a later similar occasion. This argument was based on the rule that you cannot infer that a man committed a particular crime or delict from the fact that he has previously committed other crimes or delicts. But even that is not an unqualified rule—see, for example, Moorov v. H. M. Advocate —and there are reasons for that rule which would not apply to a case like the present. It would not be right to draw such an inference too readily because people do sometimes change their minds unexpectedly. But the facts of this case appear to me to be overwhelming.

I would have had much more difficulty, if the only evidence had been that there was a general practice not to wear belts. One would assume, in the absence of evidence to the contrary, that the deceased was a reasonable and careful man, and it may be that, if the evidence proved that a reasonable and careful man would not have worn a belt on such an occasion, that would be sufficient. But I would reserve my opinion about a case which merely depended on evidence of general practice. I regard the evidence about general practice in this case as corroborating the inference to be drawn from M'Williams' own past conduct.

The appellant founded on the case of Roberts v. Dorman Long & Co., Ld. There, a steel erector who was not wearing a safety belt was killed during the erection of a steel building to which the building regulations of 1948 applied. They required that belts should be available which would "so far as practicable enable such persons who elect to use them to carry out the work without risk of serious injury." The employers did have belts, but they were kept so far from the site that they were held not to be available. One question in the case was whether the employers' breach of statutory duty could be founded on in face of evidence of a general practice to elect not to use such belts. The evidence is not fully reported, and it is not clear whether the deceased himself had ever had an opportunity to use such belts, or whether the evidence merely related to the practice of other men not to use them at other sites where they were available. Lord Goddard, C.J., said (at p. 946):

"It may well be that the Judge could form the opinion on the evidence that it was unlikely that if safety belts had been available the deceased would have used one."

But he went on to say:

"I think that if a person is under a duty to provide safety belts or other appliances and fails to do so he cannot be heard to say: "

‘Well if I had done so they would not have been worn.’"

In my view, this is not correct. "He cannot be heard to say" suggests to me personal bar or estoppel: indeed I know of no other ground on which a defender can be prevented from proving a fact vital for his defence. If I prove that my breach of duty in no way caused or contributed to the accident I cannot be liable in damages. And if the accident would have happened in just the same way whether or not I fulfilled my duty, it is obvious that my failure to fulfil my duty cannot have caused or contributed to it. No reason has ever been suggested why a defender should be barred from proving that his fault, whether common law negligence or breach of statutory duty, had nothing to do with the accident.

Hodson, L.J. (as he then was) put the matter rather differently. His view was that there was no possibility of finding out whether the man would have exercised his election one way or another. If my noble and learned friend meant that, if a man is dead, you can never prove what he would have done, I would not agree with him. Proof in civil cases depends on probability, and I think that the ordinary man would be surprised, if told that you can never say that it is probable that, in certain circumstances, a deceased man would have done one thing and not another. But if his observation was directed to the facts of that particular case, I am not prepared to say that it was wrong, without fuller knowledge of the evidence which had been led. I have already said that I wish to reserve my opinion about a case where the only evidence relates to the practice of other men engaged on other work: much may depend on the precise nature of that evidence.

The appellant also founded on Drummond v. British Building Cleaners Ld. There a window cleaner had a safety belt, but there was nothing obvious to which to fix it. It might have been fixed to a transom, but that might have caused damage to the paint work. The employers never got permission to do this, or told the man he might do it: so he had no real opportunity to use his belt. So the case depended on such evidence as there was, to the effect that he would not have used his belt if he had been told how he could use it, and I need not express any opinion as to whether it was rightly decided.

It has been suggested that the decision of this House in Wardlaw v. Bonnington Castings Ltd. lays down new law and increases the burden on pursuers. I do not think so. It states what has always been the law—a pursuer must prove his case. He must prove that the fault of the defender caused or contributed to the damage which he has suffered. But proof need not be by direct evidence. If general practice or a regulation requires that some safety appliance shall be provided, one would assume that it is of some use, and that a reasonable man would use it. And one would assume that the injured man was a reasonable man. So the initial onus on the pursuer to connect the failure to provide the appliance with the accident would normally be discharged merely by proving the circumstances which led to the accident, and it is only where the evidence throws doubt on either of these assumptions that any difficulty would arise. Normally, it would be left to the defender to adduce evidence, if he could, to displace these assumptions. So, in practice, it would be realistic, even if not theoretically accurate, to say that the onus is generally on the defender to show that the man would not have used the appliance, even if it had been available. But in the end, when all the evidence has been brought out, it rarely matters where the onus originally lay: the question is which way the balance of probability has come to rest.

The appellant had also a second line of attack. It was maintained that the employers had a duty not only to provide the deceased with a safety belt, but also "to instruct him to wear it" (I quote from condescendence 3). "Instruct" is an ambiguous word. It may mean provide information or it may mean give orders. Here, it is admittedly intended to mean exhort or encourage. It was quite unnecessary to give the deceased information: the device was not new, he was no novice, and the work which he was doing was his ordinary work. He was well aware of the advantages and disadvantages of wearing a safety belt. It was not maintained that he should have been ordered to wear it. For one thing, there is some evidence that the practice of providing belts, even where regulations do not require it, arose from the requirement of the building regulations that belts shall be available so that men can "elect" to use them. The practice is to give men that choice whether or not the regulation applies. There is certainly no practice to order or exhort men to use belts. The appellant's case is that there ought to be a practice to exhort them to do so.

In answer to that I would adopt what was said by Lord Radcliffe in Qualcast (Wolverhampton) Ltd. v. Haynes, at p. 753:

"Though indeed there may be cases in which an employer does not discharge his duty of care towards his workmen merely by providing an article of safety equipment, the Courts should be circumspect in filling out that duty with the much vaguer obligation of encouraging, exhorting or instructing workmen or a particular workman to make regular use of what is provided. Properly to measure that obligation as a legal duty requires a fuller knowledge of the circumstances of the factory and of the relations between employers and workmen and their representatives than was available at any rate in the present case."

I can find nothing to justify holding, either that there ought to have been a general practice to exhort skilled and experienced steel erectors to use these belts, or that this man ought to have been specially urged to use a belt when doing work on this tower. I am therefore of opinion that this appeal should be dismissed.

LORD MORRIS OF BORTH-Y-GEST .—I have had the privilege of reading in advance the speech which has been delivered by the noble and learned Lord Chancellor and I am in agreement with it.

I would dismiss the appeal.

LORD DEVLIN .—The appellant is the widow of a steel erector, who was killed by a fall in the course of his employment by the first respondents. She alleges that the employers were in breach of their duty at common law to the deceased in that they failed to provide him with a safety belt and failed to instruct him to wear it. The Courts below have held that the employers were in breach of their duty in failing to provide a safety belt, but that that was not the cause of the deceased's death, since he would not have worn it, if it had been provided. They have held also that there was no duty on the employers to instruct the deceased to wear it. On this second matter, I cannot usefully add anything to the opinions already expressed by your Lordships, with which I agree.

On the first matter, three points have been taken. First, whether the employers can be heard to say that the deceased would not have worn what they, in breach of their duty, failed to provide. Secondly, as to the burden of proof: whether it is for the appellant to prove that the deceased would have worn a safety belt, if provided, or for the employers to prove that he would not. Thirdly, whether, in order to establish as a matter of probability that the deceased either would, or would not, have worn a safety belt, if provided, inferences can be drawn from the deceased's habits and past conduct and from the habits of his fellow employees.

On the first of these matters Mr Stott for the appellant based his case upon the proposition that the failure to provide the safety belt was the cause of the deceased's death. In my opinion, this proposition is incomplete. There is a missing link. The immediate cause of the deceased's death was the fact that at the time of the fall he was not wearing a safety belt. The cause or reason why he was not wearing a safety belt may have been the fact that one was not provided, but the failure to provide operates only through the failure to wear. The correct way of stating the appellant's case is, I think, as follows. The immediate cause of the deceased's death was that at the time of the fall he was not wearing a safety belt; but for the fault of his employers, he would have been wearing a safety belt: therefore the fault of his employers was an effective cause of his death. So stated, it is plain that the reason why the deceased was not wearing a safety belt must be a proper subject for inquiry.

Mr Stott relied upon the decision of the Court of Appeal in Roberts v. Dorman Long Ld. This also was a case in which the death of a steel erector was caused by his not wearing a safety belt and his employers were in breach of duty in not making one available. The duty relied upon was created by Regulation 97 of the Building (Safety, Health and Welfare) Regulations, 1948, which provided that "there shall be available safety belts … which will so far as practicable enable such persons who elect to use them to carry out the work without risk of serious injury." The Court held that it was no answer for the employers, in such circumstances, to say that, if they had made safety belts available, the deceased would not have used them; the fact that they were not available gave him no opportunity of exercising his election. It is clear that this reasoning was based upon the words of the regulation; the Court held the employers to be guilty of a breach of the regulations, but not a breach of their duty at common law. The case does not therefore, in my opinion, assist Mr Stott in his argument against the first respondents; although I shall consider it again, when I consider the case against the second respondents, which is based on the breach of statutory duty.

The second point raises the question of the burden of proof. The proposition, as I have stated it above, appears to put on the appellant the burden of showing why the deceased was not wearing a safety belt. She must prove her case, and it is part of her case that he was not wearing a belt because of the fault of his employers. But, since ex hypothesi a prudent employer would provide a belt, it may follow that a prudent employee would wear it when provided. Any inquiry of this sort starts from the presumption that the pursuer or the defender, as the case may be, has done what is reasonable and prudent; and it is for the opposite party to displace that presumption by pleading and proving negligence or contributory negligence, as the case may be. So, if there were no evidence at all to show why the deceased was not wearing a safety belt, it would be proper to conclude that the reason was that the employers had failed to provide one.

This question of the burden of proof is frequently important when what is in issue is what a dead workman in fact did. Without his evidence, it may be difficult to prove that negligence by the employers was an effective cause of the death. Once negligence is proved, the fact that the workman cannot be called to account for his actions often defeats the proof of contributory negligence. But in the present case, the question is not what the deceased actually did but what he would have done in circumstances which never arose. Whether the workman is alive or dead, this cannot be proved positively as a matter of fact, but can only be inferred as a matter of likelihood or probability. Even when the workman himself is perforce silent, there may be plenty of material, as there is in this case, from which an inference can be drawn one way or the other; and then the question of burden of proof is unimportant.

That brings me to the third of Mr Stott's submissions. He submits that what the deceased would have done cannot be proved by inference. It must be proved, he says, by direct evidence, such as a statement by him that he never, in any circumstances, wore a safety belt. The fact that in the past the deceased never wore a safety belt is not admissible to show that he would not have worn one on the material occasion. Such is the argument.

There is here, I think with respect, confusion of thought. The fact that a man under certain conditions on Monday, Tuesday and Wednesday (I take this example from Mr Stott's argument) drove carelessly may be inadmissible to prove the fact that he drove carelessly under the same conditions on Thursday. Inglis v. The National Bank of Scotland, on which Mr Stott relied, may take him that far. The proposition laid down is "that it is not evidence against a party of having committed a delict to show that he has committed delicts of the like description against other persons on other occasions." I express neither approval nor disapproval of that decision; the rule embodied in it is certainly one generally applied in the criminal law. But here, the question is not what the deceased did, but what he would have done. That is a matter that is incapable of direct proof; it must be a matter of inference. His statement about what he would have done, if he were alive to make it, is only one of the factors which the Court would have to take into consideration in its task of arriving at the correct inference. A man's actions in the past may well be a safer guide than his own forecast of his actions in the future.

In my judgment, the Courts below were right to receive and consider the evidence that the deceased had never used a safety belt in the past when it was available. That is material from which it is permissible to draw the inference that he probably would not have used one, if it had been provided on the day of his death. I think also, although with more hesitation, that the Courts below were right in considering, for what it was worth, the evidence of the general practice of steel erectors, although, without some evidence of the deceased's own attitude towards safety belts, I do not think it would have been worth much.

Undoubtedly a Court should be very careful about finding what one may call hypothetical contributory negligence. A defender, whose negligence has prevented the matter in issue from being put directly to the proof, must expect that a Court will be very careful to make sure that it is acting upon legitimate inference and not upon speculation. But, in the present case, the evidence, even if it were confined to the deceased's own past acts, is, in my opinion, conclusive. If he had been injured only by the fall, and could have gone into the witness-box, and if he had there sworn that he would have been wearing a safety belt, if one had been available that morning, I do not see how he could have been believed.

The second respondents are sued, as the occupiers of the "factory" in which the deceased was working at the time of his death, for a breach of the Factories Act, 1937, section 26 (2), in that they failed to provide means, so far as is reasonably practicable, for ensuring his safety. The means suggested was the safety belt; and, as in the case of the first respondents, the Courts below found that there was a breach of duty, but that it was not the cause of the deceased's death. On the issue of causation the only question that remains is whether the obligation imposed by section 26 (2) is different in nature from that imposed by Regulation 97 of the Building Regulations, so that Roberts v. Dorman Long & Co. Ld. can be distinguished; and if not, whether that case ought to be followed. I was at one time disposed to think that the obligations were different because of the words in Regulation 97 "enable such persons who elect to use them," and because the breach in Roberts v. Dorman Long & Co. Ld. was expressed as a failure to give the workmen the opportunity to elect. But on consideration, I do not think that anything turns on these words. Some statutes may require an employee to use the means provided; others may leave him free to use them or not as he chooses. In either case, it must be open to the employer to show that the means would not have been used or that the opportunity to elect would not have been taken; it is substantially the same thing. I therefore agree with Davies, L.J., giving the judgment of the Court of Appeal in Wigley v. British Vinegars Ltd, that Roberts v. Dorman Long & Co. Ld. cannot be distinguished from the ordinary case; and that, since it can no longer be supported on the authority of Vyner v. Waldenberg Brothers Ltd., that, case having been overruled by Wardlaw v. Bonnington Castings Ltd. it can no longer be regarded as good law.