Mackenzie v Alcoa Manufacturing (Gb) Ltd

This appeal raises an issue about the circumstances in which it will be appropriate to draw an inference adverse to the operator of a factory in industrial deafness cases where there was noise in the workplace and no noise survey is available. It involves a consideration of the proper effect of the judgment of the Court of Appeal in Keefe v Isle of Man Steam Packet Company Limited [2010] EWCA Civ 683.

This appeal is from the judgment of Mr Justice Garnham dated 31 January 2019 sitting in the High Court of Justice at the Cardiff Civil and Family Justice Centre. Garnham J. had heard and allowed an appeal from the judgment of His Honour Judge Vosper QC dated 17 August 2017 sitting at the Swansea County Court. HHJ Vosper QC had heard a trial of the claim for damages for personal injury for noise-induced hearing loss brought by Brian Mackenzie (“Mr Mackenzie”) against, among other defendants, Alcoa Manufacturing (GB) Limited (“Alcoa”). HHJ Vosper QC dismissed the claim for damages. Garnham J. allowed the appeal and ordered that damages for Mr Mackenzie’s hearing and consequential losses should be assessed.

Mr Limb QC submitted that Keefe could not be applied to this case because there was in this case undisputed expert engineering evidence from which the judge could find that the noise levels had not exceeded the limits of 90 dB(A) Lepd at the relevant times. Mr Limb also submitted that, if there was a duty to carry out a noise survey there was no evidence that a noise survey was not carried out. The fact that a noise survey could not be found some 39 years after Mr Mackenzie had ceased working at the factory should not come as a surprise and it was therefore wrong to draw a Keefe adverse inference. Further Mr Limb submitted that any duty to carry out a noise survey did not arise until at least about 1974, and not 1970 as found by Garnham J. Garnham J. had been innocently misled by the misdescription of the revised version of the third edition of “Noise and the Worker” dated 1971, as the second edition of “Noise and the Worker” dated 1968. This meant that at the least the appeal should be allowed by restricting any finding of breach of duty to the period from 1974 to 1976. Mr Limb also submitted that Garnham J. was wrong to accept Mr Mackenzie’s evidence because the trial judge had found that, for very understandable reasons, Mr Mackenzie’s recollection of working times and conditions was not reliable.

Mr Huckle QC submitted that Garnham J. was right to apply Keefe because although there was uncontested expert engineering evidence it related to a survey of a different factory operated by a different party. Further there was nothing to suggest that a noise survey had been carried out because no noise survey document was listed in any part of the list of documents, including under the heading of documents which existed at one time but which were no longer in the possession of Alcoa, and the factory had operated for a long time after Mr Mackenzie had ceased working there and Alcoa had had a registered office until about 2015. It was agreed that Garnham J. had been innocently misled by the misdescription of the third edition of the “Noise and the Worker” as the second edition published in 1968. This did mean that there was no breach of duty on the part of Alcoa before about 1974 (allowing for 2 years to consider the guidance, arrange for a survey and implement the changes) and it was therefore common ground that insofar as Garnham J’s order had provided for Mr Mackenzie to recover damages from 1963 to 1974, then it required amendment on appeal. Garnham J. had been right to rely on Mr Mackenzie’s evidence.

I am very grateful to both Mr Limb and Mr Huckle and their respective legal teams for their helpful written and oral submissions. By the conclusion of the submissions it was apparent that the main issues on the appeal were: (1) whether Garnham J. was wrong to apply Keefe to the present case. This will involve consideration of: when the common law duty to carry out a noise survey arose, relevant authorities on inferences, and whether a noise survey had been carried out by Alcoa at any material time; (2) whether Garnham J. should have accepted the whole of Mr Mackenzie’s evidence where some of the evidence had been found by HHJ Vosper QC to be innocently exaggerated as a result of the passage of time.

Mr Mackenzie was born on 15 March 1946 and is now aged 73 years. Mr Mackenzie started his working life aged 16 in 1963 and was employed by Holliday Hall Limited to work in Alcoa’s factory at Waurnarlwydd, Swansea (“the factory”). Aluminium smelting was carried out at the factory and there was a foundry and extrusion mill. Mr Mackenzie was an apprentice electrician and he worked installing plant and maintaining and repairing machines. Mr Mackenzie’s case was that he worked 7 days each week with overtime, although on one day he attended college and the machines were shut from 6 am on Saturday morning until 6 am on Sunday morning and that he was exposed to noise throughout his working day.

In 1968 (when he was aged 22 years) Mr Mackenzie then worked for a few months at Port Talbot when he was not exposed to noise. Mr Mackenzie was then employed by Industrial Needs Limited and returned to work at the factory operated by Alcoa until about 1976 (when he was aged 30 years).

From 1976 to 2002 Mr Mackenzie worked at Ford Motor Company Limited and was also exposed to noise. His claim against Ford was compromised on 4 August 2017, shortly before judgment was given by HHJ Vosper QC on 17 August 2017. In 2002 Mr Mackenzie returned to work for a short period at the factory and he was provided hearing protection and his hearing was tested.

At the trial Mr Mackenzie gave evidence. He said he became aware in the summer of 2012 that he had a hearing problem. Mr Mackenzie described the noisy working conditions at the factory and the fact that he could not be heard on occasions. He said he had not seen any noise surveys being carried out.

By the time that the claim was made in 2015 it was 39 years after the relevant time of what was said to be the wrongful exposure to noise at the factory. There were no noise surveys available from the factory, which had since closed down. Alcoa’s list of documents did not list any documents relating to noise surveys as having been carried out but no longer retained.

At the trial there was evidence given in the form of a report and answers to CPR Part 35 questions by Mr Kevin Worthington, a consulting engineer working for Acoustic & Engineering Consultants Limited as a single joint instruction by the parties.

Mr Worthington noted that up to the Noise at Work Regulations 1989 it was understood that protection was necessary at levels of noise exposure above 90 dB(A) Lepd averaged over an 8 hour working day.

Mr Worthington noted that Mr Mackenzie described working on the installation of machinery at the factory and noted that “such machinery would not be operational for the majority of the period of installations, with operational conditions only occurring during testing/adjustment of machine operation. Hence, any exposure to noise during such work would be limited to background noise in the mill in general”.

Mr Worthington reported that he had addressed noise levels at the Newport premises of British Alcan, which had contained furnace areas, hot and cold mills, and finishing areas based on a survey carried out by Sound Research Laboratories Limited in 1989. He noted varying noise levels across the Newport premises and recorded that noise levels in the vicinity of the furnaces was typically in the range of 84-90 dB(A) Leq, other noise measurements were higher for 66 Tandem Mill where levels at one operator location were in the range of 97-106 dB(A). A noise measurement for the 66 tension levelling machine when running was 85 dB(A) Leq, and was noted to be 79 dB(A) Leq with the machine shut down “which would be appropriate for a person engaged in the electrical maintenance of such a machine”.

Mr Worthington concluded “…whilst it is accepted that the premises referred to are not those in which the Claimant actually worked, the indication is that whilst there are some areas of such a mill where noise levels could exceed 90 dB(A), the average level for a maintenance/installation employee would be unlikely to regularly exceed such a level. Hence, without observation of contemporaneous noise surveys/measurements from the premises at which the Claimant worked, it is not possible to demonstrate, on the balance of probability, that his average daily noise exposure would have reached or exceeded 90dB(A) during these periods of employment.” Mr Worthington concluded “Based on the limited information currently available it is not possible to demonstrate, on the balance of probability, that the Claimant’s average daily exposure level would have reached or exceeded 90 dB(A) during his employment with the 1st Defendant and/or at the premises of the 2nd Defendant. Hence, substantiation of this claim, on engineering grounds, would be very difficult.”

He said in answer to Part 35 questions “I have pointed out that the data referred to [(the Alcan Survey)] [is] of limited assistance, although if the processes and/or operations were similar to those at the relevant premises, then such information would assist. Essentially, this information is what is currently available. Without this, the conclusion would have been that there is no evidence available to show what levels of noise would have existed”.

HHJ Vosper QC set out the relevant background, noting that Mr Mackenzie had finished working at Alcoa’s factory when he was aged 30 years. Having found that the

HHJ Vosper QC then addressed the submissions made by counsel for Mr Mackenzie to the effect that no noise surveys had been produced and that the engineering expert had been unable to provide relevant evidence because of Alcoa’s breach of duty in failing to carry out a noise survey, and that a Keefe adverse inference should be drawn. HHJ Vosper QC noted that the duty to carry out a noise survey was said to arise by reason of the Code of Practice issued in 1972 and that the duty had only therefore arisen in the last few years of Mr Mackenzie’s employment. HHJ Vosper QC noted that Mr Mackenzie had said that noise surveys were not carried out but recorded that Mr Mackenzie “would not have been involved in the carrying out of noise surveys”. HHJ Vosper QC recognised the force of the submission that in the intervening years documents are likely to have been lost and concluded in paragraph 33 that “it is not possible to make a finding that the Second Defendant is in breach of duty in failing to carry out noise surveys”.

HHJ Vosper QC then distinguished Keefe on the basis that there was a noise survey at a similar factory and there was expert engineering evidence. He asked himself “whether the expert engineering evidence is sufficiently cogent to enable a finding about the level of noise to which the Claimant was exposed” and recorded that even if there had been a breach of duty in failing to carry out a noise survey, he was not persuaded that Alcoa was not entitled to rely upon the opinion of Mr Worthington. He then assessed the “cogency” of the expert engineering evidence. He noted that Mr Worthington had been told on behalf of Mr Mackenzie that the machines operated 7 days a week but that Mr Mackenzie’s evidence was not to that effect. At paragraph 44 HHJ Vosper QC recorded: “As I have said, the Claimant’s evidence did not fully accord with this information. It is likely that the information given to Mr Worthington exaggerates the Claimant’s noise exposure.”

In paragraph 47 of the judgment HHJ Vosper QC noted that because up to the 1989 Regulations 90 dB(A) Lepd been regarded as the relevant level of exposure it might be inferred that the noise levels in the Newport factory had not changed since the 1960s and early 1970s and were comparable to levels to which Mr Mackenzie had been exposed at the factory.

Having analysed Mr Worthington’s evidence and set out some paragraphs from his report HHJ Vosper QC said at paragraph 56 of the judgment that those paragraphs showed that Mr Worthington’s opinion was not based solely on the data from the Newport survey “he has had regard to the nature of the work which the Claimant was carrying out and the circumstances in which it is likely that that work was done, based upon his own engineering experience”. HHJ Vosper QC said Mr Worthington’s evidence was not simply that Mr Mackenzie had failed to discharge the burden of proof, rather it was his opinion “that as a maintenance employee the Claimant is unlikely to have been regularly exposed to levels of noise in excess of 90dB(A) when working at the Second Defendant’s factory”.

HHJ Vosper QC held in paragraph 64 of his trial judgment: “…the Claimant’s evidence could not be regarded as sufficiently precise for me to reject the engineering evidence in favour of it. That is not to criticise the Claimant. It is simply unrealistic for him to be

Garnham J. allowed the appeal from the judgment of HHJ Vosper QC. Having set out parts of the report from Mr Worthington and the judgment of HHJ Vosper QC, Garnham J addressed the argument based on Keefe.

Garnham J. considered the terms of the Ministry of Labour publication in June 1963 called “Noise and the Worker” and analysed the questions set out under the heading “Have you got a noise problem?”. The guidance provided that if the answer to several (emphasis added) of the questions was “yes” there may well be a problem of excessive noise and efforts should be made to reduce it or provide protection. It was also noted that the first steps were to carry out a noise survey and obtain specialist advice. Garnham J. found that there was no duty to carry out a noise survey before 1968 having regard to the terms of the 1963 edition of “Noise and the Worker” and held that “it is a complete answer to the appeal in respect of the period until publication of the second edition of the document in 1968”, as appears from paragraph 22 of the judgment.

Garnham J. noted “there was evidence before the Judge, however that the 1963 Guidance was amended in 1968. In the later version of the guidance, employers were advised to take such action if any one of the answers to any one (emphasis added) of the questions posed was “yes”. (As appears above it was common ground on appeal before us that the “Noise and the Worker” edition before Garnham J. was in fact the 1971 edition which had been innocently misdescribed as the 1968 edition).

Garnham J. found that there was a duty to carry out noise surveys after 1972, as appears from paragraph 25 of his judgment. He addressed the intervening period in paragraph 26 of his judgment noting that Mr Worthington had said that prior to 1972 a reasonable and prudent employer would reasonably be expected to have been aware of the existence of the “Noise and the Worker” guidance and its recommendations about noise surveys if several questions were answered yes. Garnham J. considered what he was told was the terms of the 1968 version of “Noise and the Worker” and allowed a two year period to appreciate the effect of the 1968 edition, identify specialist advice and act upon the resulting report. Garnham J. therefore held that Alcoa “was under a duty to conduct noise surveys from 1970”, as appears from paragraph 28 of the judgment.

Garnham J. then noted that there was no evidence of any noise survey at trial, nor any evidence to explain the absence of surveys. He referred to the inference drawn in British Railways Board v Herrington [1972] AC 877 and the decision in Keefe. Garnham J. noted that “…when engineering evidence is produced which provides positive evidence as to the level of noise to which workers were in fact exposed at the relevant time and the relevant premises, it is likely to mean the absence of noise surveys is of little significance. But where, as here, the engineering evidence serves simply to explain why it was not possible now to establish to what level of noise workers would have been exposed, that does not make the absence of noise surveys irrelevant. To the contrary, it serves to underline the significance of the absence of such surveys.”

Garnham J. recorded that HHJ Vosper QC had distinguished Keefe on a number of grounds but held that “in my judgment, and with great respect to the judge, none of

Garnham J. also found that the judge had not addressed the argument that after 1972 there was a duty to avoid any exposure at or exceeding 90 dB(A), which was to be contrasted with the pre-1972 duty to avoid exposure which exceeded an average dose of 90 dB(A) Lepd. Garnham J. rejected the argument that this issue was disposed of by the finding of fact that it was unlikely that Mr Mackenzie was exposed to noise about 90 dB(A). Garnham J. held that the judge should specifically have addressed the point.

Garnham J. allowed Mr Mackenzie’s appeal. The order giving effect to the judgment of Garnham J. dated 26 March 2019 provided “the case is remitted to the County Court for the assessment of damages”.

It does not appear from the order made by the Court giving effect to Garnham J.’s judgment that there was any limitation on the period for which Mr Mackenzie could claim damages notwithstanding Garnham J’s finding that there was a complete answer to the appeal up to the publication of the second edition of “Noise and the Worker” in 1968, and the fact that Garnham J. did not find a relevant breach of duty until 1970.

Keefe was a case where a seaman had been exposed to noise working in ships’ galleys. The trial judge found that Mr Keefe had been exposed to excessive levels of noise, sometimes in excess of 90dB(A) but because he could not say that this had occurred with any regularity, he dismissed the claim. The trial Judge had found that the shipowners were in breach of duty in failing to make noise assessments and that this had caused a witness for the shipowner to be visibly ill at ease when asked about the breach of duty, but the trial judge had given little weight to that breach of duty. On appeal, the Court of Appeal noted that there was a “potent additional consideration that any difficulty of proof for the claimant has been caused by the defendant’s breach of duty in failing to take any measurements” and that the judge’s judgment had become “something of a muddle” on this main issue.

The Court of Appeal held that an inference adverse to the employer should have been drawn and Longmore LJ said in Keefe at paragraph 19: “If it is a defendant’s duty to measure noise levels in places where his employees work and he does not do so, it hardly lies in his mouth to assert that the noise levels were not, in fact, excessive. In such circumstances the court should judge a claimant’s evidence benevolently and the defendant’s evidence critically. If a defendant fails to call witnesses at his disposal who could have evidence relevant to an issue in the case, that defendant runs the risk of relevant adverse findings see British Railways Board v Herrington [1972] AC 877,

In order to consider whether Keefe should have been applied to this case it is therefore necessary to consider: when the common law duty to carry out a noise survey arose in relation to this factory; whether there was a failure to carry out a noise survey; and in any event whether the judge was entitled to rely on the engineering evidence.

It was common ground that there was a common law duty to avoid exposing workers to significant noise, which was in excess of 90dB(A) Lepd. As was noted by HHJ Inglis sitting as a Judge of the High Court in the case of Parkes and others v Meridian and others Nottinghamshire District Registry 14 February 2007 “to put it in the context of Swanwick J’s judgment [in Stokes v Guest, Keen and Nettleford (Bolts and Nuts) Ltd [1968] 1 WLR 1776], complying with 90dB(A) Lepd [average] as the highest acceptable level was I think meeting the standards of the reasonable and prudent employer …”. 90dB(A) Lepd was an exposure to noise at 90 dB(A) for an 8 hour period. Every increase of 3 dB from the 90 dB(A) limit would involve a doubling of noise, so that an exposure of 93 dB(A) for a period of 4 hours equalled 90dB(A) Lepd.

It is also well known that hearing loss may manifest itself many years after the relevant exposure to noise. This is such a case. It is common ground that Courts should not have an over ready resort to the burden of proof to dismiss hearing loss claims, see Mustill J. in Thompson v Smiths Shiprepairers [1984] QB 405. It is always necessary to assess the relevant evidence. It is also well known that in many historic cases there are no direct surveys of the Claimant’s workplace and reference is sometime made to other surveys by expert engineers “and there is a degree of the application of judgment based on the expert’s experience in arriving at least a bracket of likely exposure”, see paragraph 133 of Parkes and others v Meridian and others.

There was no statutory duty to carry out surveys until the coming into force of the Noise at Work Regulations 1989 on 1 January 1990, but there was guidance which pre-dated the Regulations which it is common ground gave rise to a common law duty to carry out a noise survey in certain circumstances.

The first guidance available was from the Ministry of Labour publication, “Noise and the Worker” the first edition of which was published in June 1963 and the second edition of which was published in June 1968. The third edition was published in 1971.

The “Noise and the Worker” edition in 1963 asked six questions to enable an employer to discover whether they had a noise problem. One of the questions was “do workers find it difficult to hear each other speak while they are at work in a noisy environment”. It was common ground that Alcoa would have answered that question “yes” but could reasonably have answered the other questions “no”. The 1963 and the 1968 edition then provided “If the answer to several of these questions is `yes’, there may well be a problem of excessive noise” noting that first steps would be to carry out a noise survey. It was common ground at the hearing before us that Garnham J. had been inadvertently given the 1971 edition and was told that it was the 1968 edition. He was therefore misled into believing that it was in 1968 (as opposed to sometime in 1971) when a “yes”

In 1972 the Department of Employment published The Code of Practice for Reducing the Exposure of Employed Persons to Noise (“the Code”). The “Noise and the Worker” and the Code were guidance seeking to avoid long-term exposure to noise. If a two year period was allowed to consider the guidance, obtain a noise survey and act on it (which is the approach taken in some cases) this suggests that a common law duty to carry out and act upon a noise survey arose around 1973 or 1974 (depending on the date of publication of the third edition in 1971, which was not provided to us on this appeal).

I note that this conclusion accords with the finding about the common law duty to carry out noise surveys in noisy working environments made by the Court of Appeal in Baker v Quantum Clothing Group Ltd [2009] EWCA Civ 499; [2009] PIQR P19. Part of the judgment of the Court of Appeal was reversed on appeal by the Supreme Court in Baker v Quantum Clothing Group Ltd [2011] UKSC 17; [2011] 1 WLR 1003, but the analysis of when the common law duty arose was based on the judgment at first instance by His Honour Judge Inglis (when the case was called Parkes and others v Meridian and others) was approved by the Supreme Court. The best summary of what HHJ Inglis had decided appears in paragraph 5 of Keefe which recorded that outside those industries such as shipbuilding, where noise had always been recognised as being a problem “(ii) … precautions in relation to noise could not be expected to be taken before the publication in 1972 of the Department of Employment’s Code of Practice …; (iii) that Code of Practice required employers to measure the noise in their premises and if it was 90dB(A) leq or above, they had to take steps to reduce the noise or provide ear protectors; (iv) employers with any noise processes should by mid-1973 have been turning their minds to the problem of noise and were thus initially obliged to measure noise levels; (v) such measures should have been completed by about mid-1974 and, if noise levels above 90dB(A) leq were found to exist, protective measures should be taken.”

This means that it became common ground at the hearing of the appeal that, to the extent that the order remitting Mr Mackenzie’s case to the County Court for an assessment of damages had covered the whole period of his work at the factory, namely from 1963, it should be amended to provide that damages should be assessed for the period from 1973 or 1974 to 1976 only.

The law about inferences has been considered on a number of occasions. As long ago as 1721 in Armory v Delamarie (1721) 1 Strange 505 the court approved a direction to a jury to infer and find the highest value consistent with the jewel as described in the evidence. The jewel had been given to a pawn broker by a chimney sweep but the pawn broker refused to return it.

In Herrington the House of Lords confronted a situation where the defendant did not call any evidence about a fence separating the railway from the meadow which had been in a poor state of repair for several months. Noting that not calling evidence was a legitimate tactical move in the adversarial system of litigation, Lord Diplock said “but a defendant who adopts it cannot complaint if the court draws from the facts which have

Gibbs v Rea [1998] AC 786 was a case about the malicious procurement of a search warrant where no evidence explaining why the warrant had been sought was adduced. The Privy Council noted that “It was of course open to the defendants to elect to give no evidence and simply contend that the case against them was not proved. But that course carried with it the risk that should it transpire there was some evidence tending to establish the plaintiff’s case, albeit slender evidence, their silence in circumstances in which they would be expected to answer might convert that evidence into proof…”.

In Wiszniewski v Central Manchester Health Authority [1998] PIQR P324 the Court of Appeal held that the silence or absence of a witness might justify drawing an inference adverse to the party, but there must be some evidence which raised a case to answer before an inference could be drawn, and if an explanation for the absence was given, even if it was not wholly satisfactory, the potentially detrimental inference may be reduced or nullified.

In Shawe-Lincoln v Dr Arul Chezhayan Neelakandan [2012] EWHC 1150 Lloyd Jones J. distinguished Keefe and said at paragraphs 81-82 that “Keefe is concerned with the weight which is to be attached to evidence and the circumstances in which the court may draw inferences. This is how Longmore LJ explained it… Whether it is appropriate to draw an inference at all and, if so, the precise nature and extent of such an inference will depend on the particular circumstances of each case. Relevant considerations will include the proximity between a breach of duty and the nonavailable evidence, the effect of the other evidence before the court and what other evidence might have been available but which is not before the court.”

In Garner v Salford City Council [2013] EWHC 1573 (QB) Keith J stated at paragraph 28: “The case is unlike Keefe v The Isle of Man Steam Packet Co Ltd [2010] EWCA 683 (Civ), in which the Court of Appeal held that the defendant could not assert that it had not been proved that the noise levels on its boats were excessive when in breach of duty it had failed to measure those levels. There was no duty on the company in 1978 to check the lagging for asbestos, only guidance, and it was sufficient compliance with that guidance for the company to proceed on the assumption that it contained crocidolite”.

In Petrodel Resources Limited v Prest [2013] UKSC 34; [2013] 2 AC 415 Lord Sumption commented on BRB v Herrington at paragraph 44 noting that the courts had recoiled from parts of the statement in Herrington recording they might “convert openended speculation into fact” He noted that there must be a reasonable basis for some hypothesis in the evidence or the inherent probabilities, before a court can draw useful inferences from a party’s failure to rebut it. Silence of one party in the face of the other party’s evidence may convert that evidence into proof in relation to matters likely to be within the knowledge of the silent party.

It seems therefore that it is possible to state the following propositions. First whether it is appropriate to draw an inference, and if it is appropriate to draw an inference the nature and extent of the inference, will depend on the facts of the particular case, see Shawe-Lincoln at paragraphs 81-82. Secondly silence or a failure to adduce relevant documents may convert evidence on the other side into proof, but that may depend on

In this case HHJ Vosper QC did not draw the inference adverse to Alcoa because he accepted that documents relating to the noise survey may have been lost rather than being lacking because no noise survey was carried out, and because he found that Mr Worthington’s engineering evidence did not support the case that Mr Mackenzie had been subjected to tortious levels of noise.

It is apparent that HHJ Vosper QC took into account the fact that there were no documents showing that a noise survey had been carried out and the fact that Mr Mackenzie had not seen a noise survey carried out, but was not prepared to find that no noise survey had in fact been carried out stating “I conclude that it is not possible to make a finding that the Second Defendant is in breach of duty in failing to carry out noise surveys”.

Garnham J. found that there was an obligation to carry out a noise survey in 1970 but for the reasons which are set out above this was because he had been misled by the misdescription of the 1971 edition of “Noise and the Worker” as the 1968 edition. Garnham J. when distinguishing Herrington stated that in this case “the absence of noise surveys was more obviously explainable, given the passage of time …”. However Garnham J. later found “whilst it might be reasonable here to accept that the passage of time might explain the absence of noise survey reports, it remains surprising that none at all were produced, despite the development of relevant regulatory requirements during the ensuring period”. Garnham J. found that in these circumstances from 1970, it did not “lie in the defendant’s mouth to say that noise levels were not excessive”.

It appears that Garnham J. made a positive finding that there was a failure to carry out a noise survey rather than a finding that a noise survey had been carried out but was not available, whereas HHJ Vosper QC had made a finding that it was not possible to find as a fact that Alcoa were in breach of duty in failing to carry out a noise survey. It is established that appellate courts should be very cautious in overturning findings of fact made by a first instance judge. This is because first instance judges have seen witnesses and take into account the whole “sea” of the evidence, rather than indulged in impermissible “island hopping” to parts only of the evidence, and because duplication of effort on appeal is undesirable and increases costs and delay. Judges hearing appeals on facts should only interfere if a finding of fact was made which had no basis in the evidence, or where there was a demonstrable misunderstanding of relevant evidence, or a demonstrable failure to consider relevant evidence so that the decision could not reasonably be explained or justified. In my judgment there was no sufficient basis available for Garnham J., sitting on appeal, to overturn the finding of fact made by HHJ Vosper QC that there was no breach of duty on the part of Alcoa, because noise surveys might have been lost. This was especially so where despite the facts that the factory had continued to operate and Alcoa had had a registered office until recently, it was common ground that the absence of noise surveys was explicable because of the passage of time. In future cases where it is relevant to determine whether a noise survey has been undertaken in the past it would be helpful if both parties addressed that in pretrial questions about the existence of documents or in the evidence at trial. This would help to avoid a situation where the trial judge is left to deal with the factual finding

A principal reason why HHJ Vosper QC did not draw the inference against Alcoa was because he accepted Mr Worthington’s evidence that it could not be shown that Mr Mackenzie had been “regularly exposed to noise levels in excess of 90dB(A)”. HHJ Vosper QC found in paragraph 56 of his judgment that Mr Worthington had regard to the nature of the work done, the circumstances in which it was done, his own engineering experience, and the results from a comparable factory carrying out comparable processes. This was much more than a dismissal of the case because on the balance of probabilities it was not possible to say what was the exposure to noise. In my judgment HHJ Vosper QC was entitled to accept this engineering evidence and avoid resort to inferences, even if they might otherwise have been drawn. The approach taken by Garnham J. to the adverse inference risked elevating the decision in Keefe to a rule of law, rather than an example of the proper approach to finding facts in a particular case where the evidence showed that the defendant had failed in its duty to carry out noise surveys, and the claimant had been deprived of the opportunity to prove his case.

In these circumstances the ground of appeal relating to Mr Mackenzie’s evidence does not arise. I should say as a matter of fairness to Mr Mackenzie that there was no suggestion that he had done anything other than attempt to give his honest recollection but he, like Alcoa, was disadvantaged by the passage of time.

I should also note that we were not addressed on the other ground of appeal dealt with by Garnham J relating to the duty to avoid any exposure exceeding 90dB(A) from 1972. This was probably because of the recognition that unless the approach taken by Garnham J. to drawing the adverse inference was upheld, this point would not have made any difference to the result at trial. I can confirm that I agree with Garnham J. that a party should understand from a judgment why a point was accepted or rejected, or did not need to be addressed. As it is in the light of the findings by HHJ Vosper QC I agree that this point does not form a basis for ordering damages to be assessed, because there was no basis in this case for making a finding of fact that there had been an exposure exceeding 90dB(A) which caused loss.

For the detailed reasons given above I would allow this appeal, and restore the order made by HHJ Vosper QC.

I agree that the appeal should be allowed for the reasons given by Dingemans LJ. I have also had an opportunity to read in draft the judgment to be delivered by Bean LJ, with which I also agree.

At the trial before Judge Vosper QC the report of the single joint expert engineer, Mr Worthington, stated his opinion that as a maintenance employee the Claimant is unlikely to have been regularly exposed to levels of noise in excess of 90dB(A) when working at the Second Defendant’s factory between 1963 and 1976. The trial judge was entitled to accept that opinion and the evidence and reasoning on which it was based. He was also entitled to conclude that, given the passage of time and the closure of Alcoa’s Swansea plant, he should not draw adverse inferences from the Second Defendant’s inability to produce any noise surveys. I do not think that anything in the decision of this court in Keefe v Isle of Man Steam Packet Co Ltd prevented him from reaching that conclusion. In my view this claim was properly rejected on the facts.

Accordingly, for these reasons and those given by Dingemans LJ, I too would allow the appeal and restore the order of Judge Vosper dismissing the claim.