BAILII [Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]

England and Wales Court of Appeal (Civil Division) Decisions


You are here: BAILII >> Databases >> England and Wales Court of Appeal (Civil Division) Decisions >> Odyssey Re (London) Ltd & Anor v OIC Run-Off Ltd [2000] EWCA Civ 71 (13 March 2000)
URL: http://www.bailii.org/ew/cases/EWCA/Civ/2000/71.html
Cite as: [2000] EWCA Civ 71

[New search] [Printable RTF version] [Help]


Neutral Citation Number: [2000] EWCA Civ 71
QBCMF 99/0512/A3

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
COMMERCIAL COURT
(Mr Justice Langley)

Royal Courts of Justice
Strand
London WC2A 2LL
Monday, 13th March 2000

B e f o r e :

LORD JUSTICE NOURSE
LORD JUSTICE BROOKE and
LORD JUSTICE BUXTON

____________________

(1) ODYSSEY RE (LONDON) LIMITED
(2) ALEXANDER HOWDEN HOLDINGS LIMITED
Claimants/Appellants
-v-
OIC RUN-OFF LIMITED
(formerly ORION INSURANCE COMPANY PLC)
Defendant/Respondent

____________________

Handed Down Judgment
Smith Bernal Reporting Limited
180 Fleet Street London EC4A 2HG
Tel: 0171 421 4040 Fax: 0171 831 8838
(Official Shorthand Writers to the Court)

____________________

Mr J Sumption QC and Mr N Calver (instructed by Messrs Clifford Chance, London EC1) appeared on behalf of the Appellant Claimants.
Lord Grabiner QC, Mr A Griffiths, Mrs L Lake and Mr D Jowell (instructed by Messrs Lovells, London EC1) appeared on behalf of the Respondent Defendant.

____________________

HANDED DOWN HTML VERSION OF JUDGMENT
SMITH BERNAL REPORTING LIMITED
180 FLEET STREET LONDON EC4A 2HG
TEL: 0171 421 4040 FAX: 0171 831 8838
(OFFICIAL SHORTHAND WRITERS TO THE COURT)
HTML VERSION OF JUDGMENT
____________________

Crown Copyright ©

    Lord Justice Nourse:

    Introduction

    The primary issue in this appeal is one of fact. Did Mr Leslie Sage commit perjury in giving evidence to Hirst J on 20th November 1989 at the trial of an action ("the first action") between Orion Insurance Co Plc and Sphere Drake Insurance PLC? Hirst J found that Mr Sage's evidence was truthful and his finding was upheld by this court (Lloyd and Mann LJJ, Stuart-Smith LJ dissenting); see [1990] 1 Lloyd's Rep. 465 and [1992] 1 Lloyd's Rep. 239. But in later proceedings to which another insurance company was also a party ("the Baloise action") Moore-Bick J found that Mr Sage did commit perjury in giving evidence to Hirst J; see [1998] 1 Lloyd's Rep. (Insurance & Reinsurance) 35. That has resulted in the present action, in which Sphere Drake claims that the judgment of Hirst J should be set aside on the ground that it was procured by fraud. Having found that Mr Sage did not commit perjury in giving evidence to Hirst J, Mr Justice Langley dismissed the action. Against that decision Sphere Drake now appeals. An appeal by Orion against the decision of Moore-Bick J was compromised by an order in the Tomlin form made shortly before the hearing of this appeal started.

    At first sight these different findings made by three different judges of the Commercial Court, the cost of which to Sphere Drake, we have been told, is something of the order of US $160m, are perplexing. But it is important to emphasise that the evidence was not the same at each of the three trials. By the time of the trial before Moore-Bick J Mr Sage was dead. On the other hand there was a limited amount of material then available which was not before Hirst J, in particular an affidavit sworn by Mr Sage in 1986 and the oral evidence of Mr M E Burtonshaw. Before Mr Justice Langley there was a vast amount of additional material which was not before Moore-Bick J, in particular the oral evidence of the solicitors and counsel who had acted for Orion in the first action and contemporary documents from the solicitors' files.

    Mr Justice Langley gave a very careful judgment running to some 210 pages. Although his decision of the perjury issue in favour of Orion was a sufficient basis for the dismissal of the action, he quite rightly went on to decide three further issues. The second, which he called the attribution issue, was mainly a question of law. He held that even if Mr Sage did commit the alleged perjury, the law required that at least one of those responsible for the management of Orion or the conduct of the litigation on its behalf should have procured him to do so, or at the least relied on his evidence to deceive the court in the knowledge that it was perjured. Having found that Sphere Drake had wholly failed to establish a case of such knowledge on the part of anyone, he decided the attribution issue, like the perjury issue, in favour of Orion. That part of the judge's decision is also attacked by Sphere Drake in this court.

    The third issue decided by Mr Justice Langley was the materiality issue. On the footing that he was wrong on the first two issues, he held that Sphere Drake would only be entitled to an order setting aside the judgment of Hirst J if the perjured evidence was such that it entirely changed the aspect of the case. He held, in favour of Sphere Drake, that that test was satisfied on the facts. That part of the judge's decision has not been questioned by Orion in this court. The same can be said of the fourth issue, the election/affirmation issue, which the judge also decided in favour of Sphere Drake.

    The facts of the case as they stood before Hirst and Moore-Bick JJ sufficiently appear from their reported judgments. Mr Justice Langley's judgment has not been reported, but the additional evidence before him is referred to in detail in the judgment to be delivered by Lord Justice Brooke. In the circumstances, a further extensive rehearsal of the facts and the evidence is unnecessary. I will confine myself to essentials. Although both Sphere Drake and Orion have changed their names since the hearing before Mr Justice Langley, it is convenient to refer to them by their old names.

    The 1975 agreement, the Russell Record and the Flint Note

    Mr Sage's evidence was given in reference to a meeting in London on 23rd April 1975 attended by representatives of Sphere and Drake (then separate companies), Sphere's holding company Alexander Howden & Swann Ltd ("Howden"), Orion and a Swiss insurance company which, for brevity, has been referred to simply as Baloise. The representatives of Orion included Mr Sage, who was then a director and its general manager. At that meeting an oral agreement ("the 1975 agreement") was made between Sphere and Drake, Orion and Baloise, its background and purpose being described by Mr Justice Langley as follows:

    "Between the mid-1950s and 1975 these companies had been concerned together in various capacities (as pool members, shareholders and reinsurers) in insurance business underwritten in London and by 1975 were engaged in the run-off of that business to which they were exposed in various proportions of some complexity. In 1975 it was believed that estimates could be made for the future run-off of all the business and the April 1975 agreement provided for contributions to be made to those estimates by each company to be paid to the company which was to be responsible for administering the run-off of the particular business in question."

    There are extant two documents which purport to record the whole or part of the April 1975 meeting. The first, known as "the Russell Record", is a typed record prepared by Mr C F Russell of Sphere and Drake, circulated some weeks after the meeting and signed by the representatives of the parties, including Mr Sage on behalf of Orion. The second, known as "the Flint Note", is a manuscript note made at the meeting by Mr R J Flint of Sphere and Drake, which was disclosed on discovery in the first action in 1987 but did not attract the attention of Orion's advisers until the beginning of May 1989, some six months before the trial was fixed to start.

    The principal issue in the first action was whether the 1975 agreement was made with the intention of creating legal relations and was thus a binding contract. Orion contended that it was not binding and Sphere Drake that it was. Hirst J found that it was not binding. Both before him and at every stage thereafter the Russell Record and the Flint Note have been documents of crucial importance. I refer to them in turn.

    The Russell Record is expressed in its heading to be a record of the April 1975 meeting. It starts by listing those present and in attendance who, in addition to Mr Sage, Mr Russell and Mr Flint, included Mr F Baumli of Baloise and Mr W N Rousell and Mr L J Heritage of Orion. Mr R C Comery of Howden is recorded as being in the chair. Paragraph 1 reads:

    "The Chairman referred to the meeting of 13th December 1974 which established the principles to govern the run-off of liability attaching to certain continuing accounts of 'Orion' and 'Sphere'. He said that the object of the present meeting was to agree the final figures involved and the basis for cash settlement."

    Having referred to an attached memorandum, already circulated, detailing the respective outstanding liabilities at 31st December 1974 and the basis of assessment thereof, paragraph 2 reads:

    "After discussion it was agreed by the interested parties:-

    (a) that the figures detailed in the schedule attached to the memorandum of 22nd April 1975, represented, subject to the general reservation in para. 3 below, the final liabilities of the parties involved."

    Sub-paragraphs (b), (c) and (d) of paragraph 2 provide for the currencies in which the liabilities are to be settled, for settlement to be made in specified percentages between 30th June 1975 and 30th June 1978 and for all settlements to be made through the agency of Howden. Paragraph 3 is in these terms:

    "It was understood that the present agreement, as recorded above, did not preclude the re-opening of discussion on the sharing arrangements if actual settlements increased to an extreme degree the commitment of any party as presently calculated and as evidenced in the memorandum attached hereto."

    Paragraph 4 notes that statistics recording the development of accounts involved in the agreement will continue to be exchanged if the previously agreed figures become seriously distorted.

    The Flint Note starts by recording the date, the time of the meeting (3.00 p.m.) and those attending. The main part of it is written partly in shorthand. The agreed transcription of that part, giving names where initials appear in the original, is as follows:

    "Sphere Termination Agreement

    LESLIE SAGE Goodwill agreement and not a legal contract

    R. C. COMERY (1) Bill Rousell/Charles Russell - Been fair in applying common yardstick. Have arrived at final figures. After this meeting a major factor affects the issue. It is fair to leave the spirit there.

    LESLIE SAGE Agrees it is essential to spirit of meeting/agreement.

    FRANZ BAUMLI Agreed and will notify his successors (he retires shortly).

    R. C. COMERY (2) Simple agreement on payments, 35%, 25%, 20%, 20%.

    LESLIE SAGE Agrees in cash. Bill Rousell - 30th of June each year. 35%, 25%, 20%, 20%. R. C. Comery agreed 30th of June 1975.

    LESLIE SAGE Accounts in dollars. Should be fully funded in original currencies, ie Sterling, US $ and CAN $.

    C. F. RUSSELL Will draft record of meeting."

    Although it must always have been tolerably clear that the Russell Record was a record of an oral agreement and not an agreement in itself, it was only to be expected that once the dispute between Orion and Sphere Drake had begun to emerge at the end of 1983 a great deal of time and thought would be devoted to ascertaining the effect of the 1975 agreement on the footing that it was fully and correctly recorded in the Russell Record. The essence of the problem was, in the words of Mr Justice Langley, that there was an obvious tension between paragraphs 2(a) and 3, with paragraph 2(a) recording an agreement on the figures in the attached memorandum in terms which would usually be read as terms of final settlement, yet being expressly subject to the general reservation in paragraph 3 which suggested that any such agreement was qualified or conditional. The preponderance of judicial opinion, to which I would myself subscribe, has been in favour of the view that the agreement recorded in paragraph 2(a) was intended to be binding, while that recorded in paragraph 3, being at its highest no more than an agreement to discuss, could not be binding and, moreover, left the agreement recorded in paragraph 2(a) unimpaired. Accordingly, if Hirst J had found that the Russell Record fully and correctly recorded the 1975 agreement, judgment in the first action would have been given in favour of Sphere Drake. However, the agreement having been oral, it was open to either side to adduce evidence to the effect that the Russell Record did not fully or correctly record it.

    The Flint Note does not purport to be a full record of the April 1975 meeting. By far and away its most important feature is that it starts by attributing to Mr Sage the words "Goodwill agreement and not a legal contract". It is not disputed that he spoke those words. A critical question in the first action was whether they were spoken in relation to the 1975 agreement as a whole or only in relation to that part of it which is recorded in paragraph 3 of the Russell Record.

    Mr Sage's evidence

    In paragraph 65 of his witness statement before Hirst J Mr Sage said of the April 1975 meeting:

    "I made the point at the outset that this (meaning the agreement as a whole) was a goodwill agreement and could not be a legal contract. (Indeed, I could not override the earlier agreements ratified by the Orion Board except with its authority and in the proper form. So far as I was aware, the other persons present were in the same position.) I believe this was clearly understood and agreed by the other parties."

    The crucial part of Mr Sage's cross-examination took place on the morning of the fifth day of the trial, 20th November 1989. The four passages relied on by Sphere Drake are set out in full in the judgment of Lord Justice Brooke and I repeat the most telling exchanges below. In giving judgment, Hirst J said, under the heading "Mr Sage's evidence" [1990] 1 Lloyd's Rep. 501:

    "Mr Sage affirmed clearly that his testimony concerning his crucial statement at the outset of the April 23, 1975 meeting was based on his recollection, confirmed by the Flint note when he first saw it in the summer of 1989."

    At that point the judge referred to Mr Sage's answer to his "hey presto" question (see below) and continued:

    "The crux of Mr Mance's attack upon Mr Sage is that this evidence was an afterthought: that Mr Sage did not and cannot have had any independent recollection whatsoever, and that he embraced the Flint note when it first came into his hands, which is why the evidence in his written statement which I have quoted above ('I made the point at the outset . . .') is drawn almost verbatim from the Flint note."

    Having dealt individually and at some length with these and other points made by Mr Mance, Hirst J stated his conclusion as to Mr Sage's evidence at p. 503:

    "I therefore reject Mr Mance's suggestion, which is central to the attack on Mr Sage's evidence, that his testimony on this point was an afterthought, and I find that he has ever since 1975 held this in his memory, although of course, as he himself acknowledged, such recollection was confirmed and strengthened once his attention was drawn to the Flint note."

    These passages are of importance as showing Hirst J's understanding of what Mr Sage had been telling him, a topic to which I will return.

    The lie alleged

    Although in the early stages of this action it may not have been clear exactly what lie or lies it was alleged Mr Sage had told to Hirst J, there can be no doubt about how the matter stood by the close of the trial. In stating his conclusions on the perjury issue, Mr Justice Langley said:

    "The allegation of perjury is (as it must be) in precise and limited terms. It is that Mr Sage lied in saying he had a long-held recollection independent of the Flint Note of using the words the Note attributed to him at the April 1975 meeting. It is not alleged that he lied in saying that he had always believed that any agreement made at the April 1975 meeting was only a goodwill agreement nor in claiming a present recollection at the trial of using the words prompted by the Flint note."

    That is an important passage in the judge's judgment, in that it sets out not only the lie which was alleged but two different lies which might have been, but by that stage at any rate were not, alleged. The lie alleged was that Mr Sage had a long-held recollection, independent of the Flint Note, of using the words the note attributed to him at the April 1975 meeting. That has continued to be the lie alleged by Sphere Drake on this appeal.

    In that state of affairs Mr Sumption QC, in opening Sphere Drake's appeal, suggested that the issue of fact could be reduced to three essential questions which I have slightly reformulated as follows:-

    (1) Did Mr Sage in fact have a long-held recollection, independent of the Flint Note, of using the words the note attributed to him at the April 1975 meeting?

    (2) Did Mr Sage claim in his evidence to Hirst J that he did have such a recollection?

    (3) If he did make such a claim, did he know that it was false?

    I have found a consideration of these questions to be a most helpful means of arriving at a conclusion on this unusually difficult issue of fact, though I prefer to consider the second before the first.

    Did Mr Sage claim that he had the recollection?

    Mr Justice Langley did not find it easy to be sure exactly what Mr Sage was saying (or understood he was saying) in his evidence about his recollection about the April 1975 meeting and the relationship of that recollection to the Flint Note. With respect to the judge, who seems to have attached undue weight to what Orion's counsel and solicitors thought Mr Sage had been saying, I agree with Mr Sumption that that was a remarkable conclusion for him to arrive at in the light both of the passages in Mr Sage's cross-examination on which Sphere Drake relies and of the view which Hirst J evidently took of what Mr Sage had told him.

    The most telling exchanges as recorded in the transcript of the fifth day of the hearing are the following:

    1. "Q. Has it always been your recollection that there was specific agreement that the whole agreement was not to be legally binding, was not to be more than a matter of mere goodwill and not to be a legal agreement. Is that always your recollection?

    A. Yes, and for one specific reason . . ." (p. 17G).

    2. "Q. . . . it is your recollection, is it, that throughout you have always thought and understood and remembered that any agreement made on 23rd April 1975 was expressly agreed not to be a legally binding agreement?

    A. Yes." (p. 24A).

    3. "MR JUSTICE HIRST: What I think is being suggested to you is not that you made this up but that you never thought about it or remembered it as a goodwill agreement until you suddenly saw this document in August when you said 'Hey presto, of course that is what I said after all', is that correct?

    A. No, not in its entirety, my Lord, in the sense that I had it in mind - in my mind - and this did, perhaps, confirm - only confirm - my recollection was correct . . . it is not correct [that in the first sentence of paragraph 65 of my proof I was relying heavily on the Flint Note.] I remember the circumstances of that meeting and what was said. Both notes confirmed my thoughts on the matter - my remembrance of the matter - it was not in itself more than that." (pp. 27F to 28B)

    4. "MR JUSTICE HIRST: Are you saying quite categorically that your clear recollection is that you did open off with that remark?

    A. It is quite clear in my mind, my Lord." (p. 31B)

    Although the fourth exchange was in terms directed to Mr Sage's current recollection, it is clear to my mind, in the light of the previous exchanges and especially the third, that in his evidence before Hirst J Mr Sage was claiming that he had always recollected using the words of the Flint Note attributed to him at the April 1975 meeting.

    That that was the view of Mr Sage's evidence taken by Hirst J appears clearly from the passages I have read from his judgment. At p. 501 he said that Mr Sage had affirmed clearly that his testimony concerning his crucial statement at the outset of the meeting was based on his recollection confirmed by the Flint Note, at which point he referred to Mr Sage's answer to his 'hey presto' question. He then examined the crux of Mr Mance's attack upon Mr Sage, ie. that his evidence was an afterthought and that he did not and could not have had any independent recollection whatsoever. In rejecting that attack, the judge found, at p. 503, that Mr Sage had ever since 1975 held this in his memory, although such recollection was confirmed and strengthened once his attention was drawn to the Flint Note. So the judge's understanding was that Mr Sage had told him he had recollected making his crucial statement at the outset of the meeting ever since 1975 and that the Flint Note had confirmed, not generated, his recollection.

    In my judgment this question is one to be objectively determined on the transcripts of what was said by Mr Sage and Hirst J's understanding of it and not on what, some ten years later, Orion's counsel and solicitors thought he was saying. The material I have referred to convinces me that Mr Sage did make the claim that Sphere Drake say he made.

    Did Mr Sage in fact have the recollection claimed?

    I return to Mr Sumption's first question which, like the third, involves a consideration of the vast amount of additional material which was before Mr Justice Langley. In regard to both questions Lord Justice Brooke has set out in detail both the relevant evidence and the submissions of Mr Sumption for Sphere Drake and Lord Grabiner QC for Orion. I gratefully adopt his account of those matters.

    In relation to the first question it is, as Mr Sumption submitted, essential to distinguish between Mr Sage's belief as to the effect of the 1975 agreement and his recollection of using the words attributed to him by the Flint Note. For my part, I am quite prepared to accept that Mr Sage always believed that the agreement was not legally binding. To put it in the simplest way, he could well have believed that that was the effect of the part of it which is recorded in paragraph 3 of the Russell Record. But the question whether, in 1989, he had always recollected using particular words, however important, at the outset of a meeting in 1975 is one which requires a far more rigorous enquiry. On the evidence before him Hirst J was satisfied that Mr Sage did have that recollection. He saw "nothing improbable in such an important matter for Orion sticking in his memory", see p. 503. But now we have extensive documentary evidence of the preparations for the trial on Orion's side.

    Our primary concern at this stage of the enquiry is to look for contemporary evidence of Mr Sage's recollection having been articulated in some way before the Flint Note first came to the notice of Orion's advisers in May 1989. As Lord Justice Brooke has demonstrated, not only is there no such evidence; the evidence that there is establishes that Mr Sage did not have the long-held recollection that he later claimed. That means that Mr Sumption's first question must be answered in the negative. Moreover, the fact that until the end of July 1989 Mr Sage had no independent recollection of what was said at the April 1975 meeting must be of great significance in any consideration of Mr Sumption's third and crucial question, to which I now turn.

    Did Mr Sage know that his claim to the recollection was false?

    On this question, after the most careful reflection, I have come to the same conclusion as Lord Justice Brooke. While conscious that such a view can only be safely formed on the basis of convincing evidence, I am satisfied, on the balance of probabilities that Mr Sage did commit perjury in the respect alleged against him. In general I am content to adopt the analysis and reasoning of Lord Justice Brooke. I subscribe in particular to the importance he attaches to the contemporary documents; to his observations on the significant differences between the hearings before us and in the court below; to his account of the determined efforts made by Orion's senior management to make Mr Sage feel that he was working as a member of a team for the purpose of winning the case, "the charm offensive" and so forth; and to his treatment of Mr Justice Langley's reasons for concluding that Mr Sage did not commit perjury. While it is unnecessary for any of this ground to be covered again, I must nevertheless state in my own words the essentials of the reasoning that has led to my conclusion.

    I can start with the meeting on 11th July 1989. I agree with Lord Justice Brooke that it is clear that the whole of Mr Sage's draft witness statement was considered at that meeting, and that the existence and content of the Flint Note were then drawn to his attention, even if the note itself may not have been produced. I also agree that it seems clear that Mr Sage was asked about the words "not a legal contract" in the Flint Note, and that he did not then recollect having used those words at the April 1975 meeting. The most probable explanation is that Mr Sage, having come to the meeting underprepared, was asked to go away and give detailed answers later to the specific questions raised in paragraph 36 of the draft statement.

    I have found it very difficult to reconcile Mrs Good's evidence, accepted by the judge without qualification, that she had a visual memory of Mr Henshaw coming down the corridor with the Flint Note, soon after the shorthand had been interpreted by his secretary on 10th May 1989, with her apparent unconcern for its significance before July of that year. This has caused me to wonder whether the event described by Mrs Good did not occur until much later. However, since the outcome of that particular speculation could not be decisive of anything we have to decide, I explore it no further. The important point is that the "vital" significance of the Flint Note was not drawn to Mr Sage's attention until he received Mrs Good's letter to him of 27th July 1989.

    It is clear that by the time that she wrote that letter, if not before, Mrs Good, no doubt encouraged by the view expressed by Mr Boyd QC in consultation on 24th July, was well aware of the significance of the Flint Note. In drawing Mr Sage's attention to it, in particular to his having said "Goodwill agreement and not a legal agreement", she said:

    "All of this is consistent with what you have told us about the intention of the meeting. It is vital to our case that we establish that what was agreed on 23rd April 1975 was a goodwill agreement only and not a legally binding full and final settlement. I trust this note may help to confirm what you have already told us."

    Although it is important in other respects, further quotation from the letter is unnecessary. In my judgment Mr Sage's response to it casts more light on the factual issue we have to decide than any other single feature of the case.

    Mr Sage's response consisted of an amended draft witness statement containing many manuscript deletions and additions, together with a manuscript note in which he said that the Flint Note "certainly supports Goodwill and not legal agreement. I have enlarged in my evidence." We know that between 31st July and 7th August, when the amended draft and the note were delivered by hand to Mrs Good, Mr Sage spent more than 20 hours in making the amendments. In answer to the specific questions raised in what was now paragraph 43, he noted in the margin, first that it was consistent with his recollection that he said words to the effect of "Goodwill agreement - not a legal contract"; secondly, that he was referring to the 1975 agreement as a whole, not just the "let-out" clause; and, thirdly, that the other persons at the meeting agreed. He then inserted the following words written on a separate sheet of paper:

    "I recall clearly that I made the point that this, meaning the agreement as a whole, was a goodwill agreement and could not be a legal contract. I could not override the earlier agreements ratified by the Orion Board except by their authority. (This also, I would suggest, applied to all the parties.) The reaction of other parties was one of agreement."

    Accordingly, the position at that stage was that Mr Sage was saying that he recalled clearly using the words the Flint Note attributed to him at the April 1975 meeting. But even assuming that the Flint Note had prompted some recollection in him to that effect, it plainly cannot have been a clear recollection. The evidence that any recollection that he had must have been unclear is overwhelming. Most recently, Mrs Good had said as much in the comments she made on Mr Sage's witness statement in her instructions to counsel of 14th July. Moreover, if, as is plain, Mr Sage's recollection was unclear, he must have known that it was unclear.

    The inescapable conclusion is that Mr Sage was prepared, some three and a half months before the trial, falsely to improve his evidence on a point which Orion's solicitors had told him was vital to its case. True, it is not alleged that he perjured himself when, in the fourth exchange above quoted, he gave similar evidence as to his current recollection at the trial. Indeed, it is possible that everything that had happened in between, especially in the last three weeks of August, had produced in him an honest belief that his recollection was then clear. But his earlier readiness falsely to improve his evidence must be of great importance in deciding whether he falsely improved it in a further and different respect at the trial.

    On one copy of the typed up sixth draft of Mr Sage's witness statement dated 27th August, against the words "I recall clearly", Mr Legg wrote the words "how clearly". In the seventh draft dated 30th August the words "I recall clearly that" had been deleted and, as appears above, they did not reappear in paragraph 65 of the final statement put before Hirst J. I agree with Lord Justice Brooke that at their proofing meeting on 30th August Mr Henshaw must have told Mr Sage that he could not say that he recalled clearly that he made the point. Mr Justice Langley did not think that it was of any significance that all four words had been deleted and not just the word "clearly". He said:

    "I can well understand why the words were considered inappropriate after 14 years and the effect of their deletion was not (as Mr Sumption seemed to suggest but Mr Henshaw rightly denied) to put the resulting draft outside the realm of recollection but only outside the realm of clear recollection. Mr Legg said he thought it was a drafting point. So do I."

    Even if the judge's view of the insignificance of the deletion of all four words is accepted, I nevertheless think that his adoption of Mr Legg's description of it as a drafting point is very curious. It shows, I believe, that he failed to appreciate the importance of what Mr Sage had been prepared to say as to the clarity of his recollection after he had received Mrs Good's letter of 27th July but before it had been questioned by Mr Legg. I remain of the opinion that that point must be of great importance in deciding whether Mr Sage falsely improved his evidence in a further and different respect at the trial.

    As to that crucial point, I see the matter in this way. At the beginning of August Mr Sage must have known not only that his recollection was unclear but also that it had only been prompted by the Flint Note. He must therefore have known that during the intervening period he had had no recollection at all. While it is possible that by the time of the trial he had come honestly to believe that his current recollection was clear, it is not in my judgment possible that in less than four months he had honestly forgotten his previous lack of recollection. I am convinced that Mr Sage was again falsely prepared to improve his evidence in a vital respect, and that on this occasion he did so.

    The legal consequences of Mr Sage's perjury

    My conclusion on the perjury issue, coupled with the judge's undisputed decision on the materiality issue, makes it necessary to proceed on the footing that the judgment of Hirst J in favour of Orion was procured by Mr Sage's perjured evidence. But it does not at all follow that Sphere Drake is entitled to have the judgment set aside on that ground. Orion, not Mr Sage, was the plaintiff in the first action. Accordingly, Sphere Drake must establish, as a matter of law, either that it is unnecessary for the perjured evidence which procures the judgment to be the evidence of a party to the proceedings or that Mr Sage's evidence can be treated as having been the evidence of Orion itself. Mr Justice Langley dealt with both alternatives under the description of the attribution issue, although it seems that it is only the second of them which can accurately be so described.

    Does the perjured evidence have to be that of a party?

    For the reasons given in the judgment to be delivered by Lord Justice Buxton, which I gratefully adopt, I agree with him that it is necessary for the perjured evidence which procures the judgment to be the evidence of a party to the proceedings. In his reply Mr Sumption, correctly as I believe, submitted that the real question on the authorities is whether the certiorari cases state the modern law as to the effect of perjured evidence in terms which apply, beyond the procedural boundaries of certiorari, to a collateral action such as the present. While I was much attracted by Mr Sumption's argument on that question, I have come to the conclusion that we in this court are precluded from acceding to it both by the origins of the fraud of a party rule as an exception to the doctrine of issue estoppel (see The Duchess of Kingston's Case (1776) 2 Smith LC 644, 168 ER 175) and by the subsequent authorities, in particular the decisions of this court in Abouloff v. Oppenheimer & Co (1882) 10 QBD 295 and Boswell v. Coaks (5th November 1892 - unreported).

    Can Mr Sage's evidence be treated as the evidence of Orion?

    It is clear both in principle and on authority (see eg. McIlkenny v. Chief Constable of the West Midlands [1980] QB 283, 333, per Goff LJ) that the fraud of a party rule applies where a judgment is procured by the perjured evidence of a party to the proceedings, being a natural person. It is also clear in principle and undisputed by Orion that the rule applies where the party is a corporation, provided that the perjured evidence can be treated as having been that of the corporation itself. That raises a question of company law on which the authorities cited to us cast no direct light.

    Orion's arguments on the legal questions were presented primarily by Mr Griffiths. After exchanges between the court and Mr Griffiths, followed by Lord Grabiner, and a period for reflection, Lord Grabiner formulated Orion's submissions on this question in six propositions, the first five of which were to the following effect. (The sixth is not directly material and need not be stated.) First, in order that the perjured evidence may be treated as that of the company, the company must either procure it to be given or adopt it with knowledge of the perjury. Second and in consequence, the rank or status within the company of the person who gives the perjured evidence, is not necessarily decisive. Third, in order to determine whether the company has the requisite knowledge for it to procure or adopt the perjured evidence, the directing mind and will test (see Lennards Carrying Co Ltd v. Asiatic Petroleum Co Ltd [1915] AC 705, 713, per Viscount Haldane LC) is applicable. Fourth, the knowledge of a person to whom the conduct of the litigation on behalf of the company is delegated is probably (I would think certainly) sufficient, even though that person is not part of its directing mind and will. Fifth, if the person who gives the perjured evidence is part of the directing mind and will of the company, albeit not a person having the conduct of the litigation, the rule would in principle also apply.

    In my judgment those five propositions, some of which are unobjectionable in themselves, do not adequately address the problem which confronts us in this case. I am of the opinion, shortly stated, that, for the purposes of the fraud of a party rule, it is in certain circumstances possible for perjured evidence to be treated as that of a company, even where it is neither procured or knowingly adopted by the company nor given by someone who is part of the company's directing mind and will or a person to whom the conduct of the litigation has been delegated.

    It is necessary to state or repeat three further propositions, which I take to be obvious. First, the company cannot give evidence itself. Second, a person who gives evidence on behalf of a company does not do so as its agent. Third and on the other hand, there would be an unacceptable shortcoming in company law if evidence given on behalf of a company was incapable of being treated as the evidence of the company itself. So the question is when and in what circumstances it ought to be so treated.

    A good starting point is an illuminating passage in the judgment of the Privy Council delivered by Lord Hoffmann in Meridian Global Funds Management Asia Ltd v. Securities Commission [1995] 2 AC 500, 507, where, after referring to a company's primary rules of attribution (generally found in its constitution) and the general principles of agency, he said:

    "The company's primary rules of attribution together with the general principles of agency, vicarious liability and so forth are usually sufficient to enable one to determine its rights and obligations. In exceptional cases, however, they will not provide an answer. This will be the case when a rule of law, either expressly or by implication, excludes attribution on the basis of the general principles of agency or vicarious liability. For example, a rule may be stated in language primarily applicable to a natural person and require some act or state of mind on the part of that person 'himself', as opposed to his servants or agents. This is generally true of rules of the criminal law which ordinarily impose liability only for the actus reus and mens rea of the defendant himself. How is such a rule to be applied to a company? . . . . there will be many cases . . in which the court considers the law was intended to apply to companies and that, although it excludes ordinary vicarious liability, insistence on the primary rules of attribution would in practice defeat that intention. In such a case, the court must fashion a special rule of attribution for the particular substantive rule. This is always a matter of interpretation: given that it was intended to apply to a company, how was it intended to apply? Whose act (or knowledge, or state of mind) was for this purpose intended to count as the act etc. of the company? One finds the answer to this question by applying the usual canons of interpretation, taking into account the language of the rule (if it is a statute) and its content and policy."

    Although Lord Hoffmann's language was no doubt influenced by the particular task of interpreting a statutory provision whose operation depended on a company's knowledge of a certain state of fact, his statement of principle is of general application. Thus in the present case we are concerned with a substantive rule of judge made law which applies to a company, and the question, in Lord Hoffmann's words, is:

    "Whose act [ie. the giving of perjured evidence] was for this purpose intended to count as the act of the company?"

    In answering that question we must take into account the policy of the substantive rule, which is that a person who procures a judgment to be given in his favour by fraud ought not to be allowed to profit from the fraud.

    The question whether perjured evidence should, for the purposes of the fraud of a party rule, be attributed to a company depends on the facts of the particular case. Here they must be unusual. On the one hand, Mr Sage's evidence was neither procured or knowingly adopted by Orion nor was it given at a time when he was still part of its directing mind and will. On the other hand, Mr Sage, as a director and the general manager of Orion, was both part of its directing mind and will at the time and personally responsible for the transaction about which he gave the evidence in support of its case. Moreover, in 1989, although no longer an officer or servant of Orion, he was paid £50 an hour as a consultant to it for the purposes of the first action, and as a member of the team which took decisions as to how its case was to be presented at the trial.

    On those facts Mr Sumption submitted that Mr Sage's evidence should be attributed to Orion. The basis of his submission was that Mr Sage was a person who had, in relation to the first action, a connection with Orion that was so close as to make it unjust that it should be entitled to disavow the evidence even though it had no knowledge that the evidence was perjured.

    Although it makes no difference in the result, I do not myself feel comfortable with a test whose satisfaction depends on the injustice of the company's being entitled to disavow the perjured evidence. The question is one of fact and degree which must be determined on a realistic assessment of the particular circumstances. I prefer a test I have found helpful in the past, by reason of its requirement that the acts of the natural person should be identified as the acts of the company. That is the test which was adopted by Eveleigh J in delivering the judgment of the Criminal Division of this court in R v. Andrews-Weatherfoil Ltd [1972] 1 WLR 118, 124:

    "It is necessary to establish whether the natural person or persons in question have the status and authority which in law makes their acts in the matter under consideration the acts of the company so that the natural person is to be treated as the company itself."

    On that view of the matter the question, ultimately, is this. For the purposes of the fraud of a party rule, did Mr Sage have the status necessary to make his evidence the evidence of Orion? I refer only to his status, because the concept of authority adds nothing in the present context and, moreover, is potentially misleading where the natural person is neither the agent of the company nor someone who has been suborned to commit the perjury.

    The principal objections to the test embodied in Lord Grabiner's five propositions are that it is insufficiently flexible to reflect the policy of the fraud of a party rule and that it can in certain circumstances be capricious in its results. Thus, under the first objection, in the case where nobody but the perjurer knows of the perjury the company would be allowed to profit from the fraud unless, at the time that the evidence was given, the perjurer happened to be part of the company's directing mind and will or a person to whom the conduct of the litigation had been delegated. Under the second objection, though the rule would have applied if the trial had preceded Mr Sage's retirement in 1980 or if his retirement had been deferred until 1990, it would not have applied if the trial had taken place shortly after, even one day after, his retirement or if he had retired shortly before, even one day before, the trial took place in 1989. These objections are fatal to the acceptability of Lord Grabiner's test and confirm the need for a more pragmatic test such as I have suggested. It is also to be noted that in the Meridian case the Privy Council were of the opinion that the directing mind and will test had not been expounded by Viscount Haldane LC as "a general metaphysic of companies", and that the difficulty of applying it in some of the subsequent cases had been caused by concentration on the particular phrase rather than on the purpose for which it had been used; see [1995] 2 AC, at pp. 509-511.

    Applying the test suggested, I am satisfied that, at the time that he went into the witness box in November 1989, Mr Sage did have the status necessary to make his evidence the evidence of Orion. In my judgment the two most important considerations are, first, that he was the witness, above all others, on whose evidence the success of Orion's case had come to depend. He was its "vital" witness. There is nothing fanciful, adopting the words of Eveleigh J, in treating Mr Sage as having been, for the purposes of the trial, Orion itself. Of equal importance is the consideration that he had acquired that status not simply because his evidence related to a transaction for which he had been personally responsible as part of Orion's directing mind and will at the time, but also because in the six months or so before the trial he had been a committed member of the team which took decisions as to how Orion's case was to be presented. The evidence established that Orion deliberately sought, as Mr Sumption put it, to make Mr Sage feel part of a team which was helping to row it to victory. Whatever the rights and wrongs of that may have been, Orion succeeded in identifying him with its own interests and thus with itself.

    While I have expressed discomfort with a test whose satisfaction depends on the injustice of the company's being entitled to disavow the perjured evidence, in following the Privy Council's precept that we must fashion a special rule of attribution for the particular substantive rule I fervently agree with Mr Sumption that technicalities cannot be allowed to defeat substantial justice. On a view of the case as a whole, it would in my judgment be unjust if Mr Sage's perjured evidence could not be treated as the evidence of Orion. If that requires an extension of the principles of company law as they are at present understood, so be it. In support of such an extension Mr Sumption offered the interesting and valuable analogy of the development of the principles of vicarious liability in order to meet the justice of cases such as Lloyd v. Grace, Smith & Co [1912] AC 716.

    Conclusion

    I would allow Sphere Drake's appeal and set aside Hirst J's judgment in the first action.

    Lord Justice Brooke:

    Like Nourse LJ, I will call the claimants by their original name "Sphere Drake" and the defendants by their original name "Orion".

    This is a most unusual piece of litigation. It is centred on the events which took place at a fairly short meeting at Orion's offices in the City of London on the afternoon of 23rd April 1975. In November 1989 Hirst J conducted a 13-day trial and made findings which were determinative of the legal effect of what took place at that meeting: see the report of his judgment at [1990] 1 Lloyd's Rep 465. His judgment was upheld by this court, after a 12-day hearing, in August 1991: see [1992] 1 Lloyd's Rep 239. In October and November 1996 Moore-Bick J conducted a 27-day trial, in which a different party was involved, and at which he received evidence that was to some extent different from that tendered to Hirst J. He came to precisely the opposite conclusion to Hirst J: see [1998] Lloyd's Ins Rep 35. We were scheduled to hear an appeal by Orion against Moore-Bick J's judgment during the fortnight which preceded the hearing of the present appeal, but that appeal was settled on undisclosed terms in September of this year shortly before it was due to be heard. Sphere Drake were, in effect, the losers on both occasions. We have been told by Mr Sumption QC that the estimated cost to them of these mutually inconsistent findings is in the region of US$160 million.

    They brought the present action in an endeavour to rectify in their favour what they see as the unhappy situation created by these two conflicting judgments. Lord Grabiner QC discouraged us from having too much sympathy with Sphere Drake on this account, since it would have been open to them to appeal against Moore-Bick J's judgment and they chose not to do so. Mr Sumption riposted that his clients did not appeal against that judgment because they considered that it was correct.

    It was the evidence of Mr Leslie Sage, who was the general manager and an executive director of Orion in April 1975, which persuaded Hirst J to find in their favour in the first action. Mr Sage had died before the second trial took place, and Moore-Bick J, for his part, not only declined to put the same weight on his evidence but expressly held that Mr Sage had lied in a material respect. In the present action Sphere Drake seek an order setting aside the judgment of Hirst J on the grounds that it was procured by fraud. After a trial lasting 17 days Langley J found that Mr Sage had not lied. He went on to hold that even if he had lied, he was a mere witness and his lie could not be imputed to Orion so as to deprive them of the fruits of their judgment. He also dismissed Sphere Drake's contentions that two senior representatives of Orion, along with their leading counsel and two of the solicitors who had acted for them in connection with the first action, had been parties to the alleged fraud, and there is no appeal against that part of his judgment. The lie attributed to Mr Sage is that he told Hirst J that it had always been in his recollection that he had said at the outset of the critical meeting that any agreement made at that meeting would be a goodwill agreement only and that the other parties had assented to this proposition. This came to be described as "the goodwill agreement exchange".

    The judge held that this part of Mr Sage's evidence was material, in the sense that it entirely changed the relevant aspect of the case, for four reasons. First, the documents, and in particular the documents described in the earlier judgments as the Flint Note and the Russell Record, were not conclusive and contained material on which each side could probably rely. Nourse LJ has quoted the salient parts of these two documents. Secondly, the probabilities seemed to him to have been nicely balanced. Thirdly, he thought that the oral evidence at the trial before Hirst J, apart from Mr Sage's evidence, was either neutral or self-cancelling. After considering the judgments of Hirst J and the majority in the Court of Appeal the judge believed that the focus of the trial, and even more the focus of the appeal, was centred on the reliability of the evidence Mr Sage had given to the effect that it was understood at the meeting that the discussions were not to result in a legally binding agreement. This was the issue to which the allegation of perjury was directed, and the length and existence of Mr Sage's recollection became an important part of the attack on his reliability. Finally, the consideration of the documents and the probabilities by Hirst J and the majority of the Court of Appeal not only led them to different conclusions about whether the parties were driving towards finality but were substantially conducted in the context of testing and reaching an assessment of Mr Sage's reliability.

    In these circumstances the judge considered that it would be artificial to suggest that a lie by Mr Sage about the length of his recollection would not have affected an assessment of the fact of his recollection itself. Furthermore, whilst the point was not conclusive, it was persuasive that Moore-Bick J, having found that Mr Sage had been untruthful about his recollection, had reached the opposite conclusion to Hirst J as to the binding nature of the agreement.

    We did not hear argument from the parties on the way in which the judge expressed his reasons on this point, since Orion did not seek to challenge his conclusion that Mr Sage's evidence about the length of his recollection was material in the requisite sense. They merely contented themselves by submitting a short written note explaining why Hirst J and the majority of the Court of Appeal were willing to accept that, whatever might have been the position in December 1974, by 23rd April 1975 it might no longer have been the case that the parties were set on finality.

    The background facts of this dispute have been extensively set out in the reported judgments of Hirst J, the Court of Appeal and Moore-Bick J at [1990] 1 Lloyd's Rep 465, 466-481; [1992] 1 Lloyd's Rep 239, 240-252; and [1998] Lloyd's Ins Rep 35, 39-48.

    They are more briefly summarised in the judgment under appeal, which has not yet been reported. Nourse LJ has quoted a paragraph of that judgment which goes to the heart of the matter. I consider that it is not necessary to repeat them again here. It was common ground that the meeting on 23rd April 1975 was concerned with an effort, started the previous November, to bring to an end the outstanding loose ends that were left over from a number of unsuccessful insurance ventures in which the parties represented at that meeting (Orion, Sphere, Drake, Howden and Baloise), had been involved, in one capacity or another, in the 1950s and early 1960s. The dissenting judgment of Stuart-Smith LJ in the appeal in the first action and the judgment of Moore-Bick J set out cogent reasons, based on many contemporaneous documents, in support of their conclusions that those who had participated in the April 1975 meeting had intended their agreement to be legally binding, even though they had also committed themselves to reopening their discussions if the situation subsequently deteriorated to an extreme degree. Apart from Lloyd LJ, all the judges who have considered the effect of this ancillary commitment have held that it was binding in honour only.

    The points that are material to the view of the evidence set out in Orion's short written note are to be found in the judgments of Hirst J at pp 496-7, Lloyd LJ at pp 269-70 and Mann LJ at pp 295-6 and do not call for any great elaboration here. They rely in particular on some of the contents of a six-page memorandum prepared for Orion by Mr Heritage and Mr Rousell in February or March 1975, which Mr Rousell had discussed with Mr Russell on 21st April 1975, and on some of the language used by Mr Russell of Sphere Drake (who were then separate companies) in his memorandum of 22nd April 1975 which formed the agenda for the critical meeting the following day. For example, phrases like "cut-off", "full and final settlement" and extinguishing liabilities", which had appeared in Mr Russell's record of the meeting on 13th December 1975, no longer appeared, and the word "premiums" was replaced by the word "cash payments".

    These points did not carry the same weight with Stuart-Smith LJ and Moore-Bick J, and they do not explain why Mr Baumli of Baloise, who was not a party to the discussion between Mr Rousell and Mr Russell, remained oblivious of the fact that on 23rd April the parties did not reach the finality he was seeking if Hirst J's view of the matter was correct. It is not, however, necessary for this court to express its own opinion on these matters other than to say that the judge's unchallenged finding as to the materiality of Mr Sage's evidence about the length of his recollection was clearly right.

    On the hearing of the present appeal the two leading counsel encouraged us to adopt widely differing approaches to the evidence before Langley J. This evidence had consisted in part of contemporaneous documentary evidence, and particularly that relating to the years 1985, 1986 and 1989, when Orion was preparing for the trial which eventually took place before Hirst J in November 1989, and in part of oral evidence by Orion's senior executives and their lawyers who were engaged in telling the judge towards the end of 1998 what they then recollected of events which had taken place for the most part more than ten years earlier. Lord Grabiner QC encouraged us to rely quite heavily on this oral evidence, particularly because all Orion's witnesses had impressed the judge as being witnesses on whose evidence he could safely rely. Mr Sumption QC, for his part, argued that however much these witnesses might be attempting in good faith to recollect these events that had occurred so long ago, the judge ought to have recognised that their evidence was inherently unsafe. He advised us to concentrate most of our attention on the contemporary documents. He cited in this context a passage in the speech of Lord Goff of Chieveley in Grace Shipping Inc v CF Sharp & Co (Malaya) Pte Ltd [1987] 1 Lloyd's Rep 207, 215-6, PC.

    In that case the trial judge had been faced with the task of assessing the evidence of witnesses about telephone conversations which had taken place over five years earlier. In such a case, Lord Goff said, memories might well be unreliable, and it was of crucial importance for a judge to have regard to the contemporary documents and to the overall probabilities. Lord Goff referred with approval to what he had himself said in The Ocean Frost [1985] 1 Lloyd's Rep 1, where he had said that it was his experience that it was essential in cases of fraud, when considering the credibility of witnesses, always to test their veracity by reference to the objective facts, proved independently of their testimony, in particular by reference to the documents in the case, and also to pay particular regard to their motives and to the overall probabilities.

    Mr Sumption suggested that there were three essential questions at the heart of this appeal:

    (1) Did Mr Sage in fact have a prior recollection independent of the Flint Note of the goodwill agreement exchange?

    (2) Did Mr Sage claim in his evidence that he had such a recollection?

    (3) Did he appreciate, when making this claim, that it was unjustified?

    He said that the answers we should give to these questions are: (1) No; (2) Yes; (3) Yes.

    He first set out to show us why he contended that up to the end of July 1989 Mr Sage had never suggested to anyone that he possessed any such recollection. He said that the evidence on this point fell into five self-contained stages in the development of Mr Sage's views on what took place in 1975.

    The first of these stages occurred between 1980 and 1984, and arose out of the evidence given by Mr Burtonshaw to Moore-Bick J

    In the late summer or early autumn of 1980 Mr Sage had been asked for the first time to recall the arrangements that had been made in April 1975. In 1978 Mr Burtonshaw had joined the defendant company in the capacity of joint managing director with Mr Sage. In that capacity Mr Sage had remained responsible until his retirement in September 1980 for the London market side of Orion's affairs, and the existence and nature of the 1975 arrangements had cropped up when Mr Burtonshaw was being briefed by Mr Sage about that side of the company's business before taking over as sole managing director.

    Mr Burtonshaw did not give evidence before Hirst J. He gave evidence on behalf of Orion at the trial of the second action before Moore-Bick J. In his witness statement in that action, signed in March 1996, he described how Mr Sage had discussed with him the complex arrangements and working relations which had evolved over several years between Orion, Sphere, Drake, Baloise and Howden. His witness statement continued:

    "In particular I remember him stating that some liabilities of Baloise had been capped under a cut off agreement which I understand to be the agreement entered into in April 1975 between Orion, Sphere, Drake and Baloise in relation to the parties' respective liabilities arising out of a marine insurance pool ... and ... a special excess of loss account known as the Sphere X Account.... This agreement had been accompanied by a gentleman's agreement to the effect that should claims paid subsequently escalate far beyond the levels contemplated at the time of the cut off Sphere Drake and Baloise would continue to contribute to those claims. I do not remember seeing any written instrument recording this gentleman's agreement. I do not remember any more specific detail."

    In his oral evidence Mr Burtonshaw told Moore-Bick J that he thought the words "capped" and "cut off agreement" were his, not Mr Sage's, but that Mr Sage had used words to that effect. It was Mr Sage who had used the words "gentleman's agreement". He added that he had not discussed in terms with Mr Sage whether the whole agreement was a gentleman's agreement or whether it was only the agreement to reopen the matter in the event of the escalation of claims that was so characterised. He said that Mr Sage had mentioned the way in which the parties had estimated their liabilities, how some money had changed hands as a funding operation in order to ensure the easy handling of the run-off, and how this was accompanied by a gentleman's agreement to say that should the claims experience deteriorate very significantly the parties would come back to the whole matter.

    On 21st August 1984 Mr Burtonshaw discussed this matter again with Mr Sage at a lunch meeting which was also attended by Mr Rousell. This was at a time when the claims were escalating, and the defendants were considering whether there were any effective steps they could take to limit their losses.

    Mr Burtonshaw could not recall what Mr Sage had said at this meeting, and said that he was no wiser after the lunch than he had been before it. It is clear, from a contemporary note by Mr Rousell, that Mr Sage suggested to Mr Burtonshaw that he should study the board minutes of the 1966-7 period to see if they contained any helpful material.

    Mr Burtonshaw did not give evidence before Langley J. Orion did not apparently wish to call him, and his evidence at the second trial was adduced by Sphere Drake pursuant to the provisions of the Civil Evidence Act.

    Mr Sumption derived from this evidence, such as it was, the contention that when Mr Sage was first invited to recollect what had happened in 1975 he did not say that he remembered that it had been expressly agreed at the meeting that anything that was decided then would not be legally binding. On the other hand he appeared to be saying that what came to be described as "the let-out clause" represented a gentleman's agreement along the lines recalled by Mr Burtonshaw in his evidence.

    The next time Mr Sage was invited to recall what had happened in April 1975 was on 1st March 1985 when he attended the offices of Orion's solicitors, Linklaters and Paines. Orion had sought advice from that firm in September 1984 and this meeting formed part of the evidence-gathering exercise which had to be carried out before the solicitors could put the papers before counsel. Mr Harris of Linklaters had told Mr Rousell that Mr Sage's recollection of the agreement might be crucial if there was a dispute as to its construction. Mr Tapsfield and Mrs Good were by now handling the matter on behalf of Linklaters, and Mr Sage was accompanied to the meeting by Mr Rousell and Mr Heritage, who had also attended the 23rd April 1975 meeting on behalf of Orion. In March 1985 Mr Rousell (as well as Mr Heritage) was still employed by Orion and was acting on their behalf in giving instructions to their solicitors. Mr Tapsfield was a litigation partner in Linklaters, while Mrs Good had been an assistant solicitor in their litigation department for only two and a half years at this time. When Langley J saw her 13 years later, he described her as careful, impressive and obviously intelligent.

    When she gave evidence to him, Mrs Good told him that Mr Sage had done most of the talking at the meeting. She said that he was a coherent, intelligent and articulate man but one with a tendency to ramble on. He was 66 years old at the time and still held a number of consultancy positions with insurance companies. She added that the meeting lasted a couple of hours.

    Mrs Good's notes of the March 1, 1985 meeting include references to:

    stopping the flow of paper; coming to an arrangement adjustable in certain events; principle: no one wins or loses; if 10% too much/too little then paid back; trusted each other to account, trust re accurate accounts and reserves; they trusted us to re-open if we were wrong, can't now suggest a fixed settlement, go back to November 1974 participation.

    Mr Tapsfield's notes include :

    reserve very prudent (over egged). Stop mass of paper. ?Trusted each other to account. Arrangements adjustable in certain events. Remembers variation by 10% mentioned. Fairly sure. Rousell think £10,000; don't want to hurt each other, if proves inequitable we'll re-open; re-negotiate - friendly - no one to lose or gain. You'd go back to the formula on the sheet. We'd re-open all accounts and go back to a cash basis; why have para. 4? Sage can't recall

    and a comment by Mr Tapsfield to himself:

    Keeps going back to 65/66/67 principles.

    After the meeting Mrs Good prepared the first draft of Mr Sage's statement, which was to be sent to counsel in due course after Mr Sage had approved it. She sent it to him on 6th March, and he returned it, heavily annotated with three riders written in his own handwriting, on 12th March. He said he had looked carefully at the papers Mrs Good had sent him, and he hoped that the alterations he had made would give a clearer picture. Mrs Good had included in her draft references to the pool being based on an underlying intention that no single party was to make a profit or loss which was not shared by the other parties in the agreed proportions, and to the same spirit underlying the discussions which began in December 1974. Mr Sage left those references untouched, but he altered a reference to the parties seeking a "settlement agreement" in 1974 to seeking "a method to avoid unnecessary administrative procedures and cash".

    The April 1975 agreement was dealt with in paragraphs 9-11 of this six-page draft. As drafted, one sentence read "I cannot remember precisely but I have a recollection that a figure of 10% was mentioned ..." Mr Sage redrafted this so as to read: "I recall that the trigger amount (to cause readjustment) was discussed. I am reasonably certain that a figure of 10% was mentioned ...".

    After she received Mr Sage's suggested amendments, Mrs Good spent a day at Orion's Folkestone offices going through old files, and on 19th March she wrote to Mr Sage suggesting she should give him a call to discuss his statement once he had had the chance to consider some relevant documents she had bespoken. She had in fact found nothing which shed any light on the nature and status of the arrangement which had been entered into in April 1975, but there has survived a draft of the statement which must be attributed to late March 1985 which contains a number of changes and additions to the wording approved or suggested by Mr Sage on 12th March, no doubt following the telephone call foreshadowed in Mrs Good's letter.

    On 2nd April 1985 Mrs Good sent Mr Sage a new clean draft of his statement, which now ran to nine pages. She told him that the purpose of the statement was to inform both counsel and her firm as fully as possible in respect of his recollection. She added that it was most important that they had a record of his recollection in the absence of full documentation.

    Mr Sage responded two days later. He said he had gone through the statement very carefully with Mr Rousell and made various comments. He told her that if she wished to see him he was sure they could arrange a suitable time.

    Notwithstanding the work that had already been done, Mr Sage made a large number of suggested alterations to the wording on the first six pages of the draft. He left the description of the April 1975 agreement unaltered. It was now contained in paragraphs 12-14 of the draft and read:

    "I attended a meeting at Orion's offices on 23rd April 1975 a note of which (prepared by Mr Russell of Drake) is attached hereto as Annex 2. At that meeting an agreement was reached (the "Settlement Agreement") whereby cash payments were to be made on a deferred basis to avoid any material advantage to Orion through holding cash before settlement became due. This agreement was made subject to a reservation that if actual settlements increased the agreed commitment of any party "to an extreme degree", then the sharing arrangements would be re-opened. I remember that there was a discussion as to what would constitute "an extreme degree". It was the clear intention of all parties at that meeting, in line with the spirit of the Agreements, that if the then estimates proved to be wrong then no single party should suffer because too much or too little had been paid under the cash adjustment. Some risks underwritten were very long-term and it was understood that the agreed figures were best estimates only and that the run-off might prove to be substantially less or more onerous than anticipated. The meeting agreed that in that event the contributions of parties to the run-off would have to be adjusted. It was for this reason that the general reservation in paragraph 3 of the note of the meeting was included and that paragraph 4 provided for exchange of statistics if the previously accepted figures became "seriously distorted". I recall that the level which would call for re-adjustment was discussed. I am reasonably certain that a figure of 10% was mentioned, namely that if the outstandings proved to be more than 10% out of line (plus or minus) with the agreed figures then the position would have to be re-opened.

    Mr Flint, the Secretary to Sphere, subsequently circulated the note of the meeting (Annexe 1) together with the memorandum (Annexe 2) which each party then signed by way of agreement.

    In my experience as an underwriter the amount of the losses which are now being incurred in the course of the run-off in relation to product-related hazards, and in particular asbestosis, are of an extreme degree whatever measure is applied. The losses presently being incurred were wholly unforseeable by the Market in 1974. The Settlement Agreement thus requires adjustment as was the original intention and the three parties should now continue the run-off in accordance with the original percentage contributions under the Agreements on the basis of good faith and trust."

    Mrs Good was still in the course of incorporating the suggested amendments when Mr Sage telephoned her on 10th April 1985. She noted that he told her:

    "that in his view the matter was very simple and that the [Russell Record] was merely the practical application of a principle which had been established and operated for some long time between the three parties with their cross shareholdings, the assets (sic) being that the agreement was to share losses pro rata to the shareholdings."

    Mr Sumption derives from this evidence (which represents the second stage of the history he was inviting us to consider) the contention that when Mr Sage was being invited to recall the April 1975 meeting ten years after it had taken place he had no memory that anything had been said in express terms at the meeting about the discussions being on a goodwill basis only, or that there would not be a legal contract. His recollection then was to the effect that the parties all trusted each other and were dealing with the problem of the run-off in accordance with the principles they had agreed between 1965 and 1967. The judge himself noted that there was no suggestion at all in the draft, as sent back by Mr Sage (in fact, on two separate occasions), that it had been expressly agreed at the meeting that any deal was to be only a gentleman's agreement and not legally binding.

    Mr Sumption's third stage of the history occurred in July 1985.

    On 4th July 1985 a conference took place with junior counsel (Mr Roderick Cordara). Mr Tapsfield and Mrs Good attended, along with Mr Heritage and Mr Rousell. Mr Sage did not attend the conference, because he had retired and counsel were not then permitted to discuss a case with witnesses.

    Mr Cordara advised that Orion's prospects of recovery were less than 50%. He mentioned possible arguments that could be raised, but in general he took the view that the April 1975 agreement was legally binding and that Clause 3 of the Russell Record was too vague to be enforced. On 5th July Mrs Good sent Mr Sage a note of this conference, and on 8th July Mr Sage rang her in response. He told her that he felt very strongly that the agreement constituted an accounting exercise only, this being one of the possibilities which Mr Cordara had considered. On 9th July Mrs Good wrote again to Mr Sage because counsel had asked her to check with him whether his reference to a 10% drift had been agreed by the other parties present at the meeting. She asked whether he could recall any agreement, whether verbal or even indicated by a nod of the head.

    On 12th July Mr Sage prepared his own written comment on this letter and the conference note. His account of the agreement was to the same effect as he had described it before. He also referred to the underlying principle always being proportionate sharing with no advantage or disadvantage to any one party. He said that there had indeed been a discussion between the parties of a 10% variation plus or minus on the outstanding figures This was put forward as an acceptable basis for triggering a readjustment and there were no dissenters. On 16th July Mrs Good asked for further details of this discussion. Mr Sage responded on 22nd July to the effect that after they had discussed in the Orion Board Room the matter outlined in the [Russell Record], most, if not all, the participants moved into his office, perhaps to allow the Board Room to be prepared for lunch. He believed he then raised the subject of the "trigger point" and he did not recall any adverse reaction. He was sure there was none.

    Mr Sumption submitted that if Mr Sage had recollected (before he saw the Flint Note) that he had said at the April 1975 meeting that this was a gentleman's agreement only, these exchanges would have provided him with a golden opportunity to say so.

    Mr Sumption's fourth stage related to the preparation of the affidavit which Mr Sage eventually swore in May 1986. The history of this affidavit runs along the following lines.

    During the course of 1985 counsel advised that proceedings should be issued on the basis that Sphere Drake would not be able to rely successfully on the April 1975 agreement, although they were not particularly sanguine about the prospects of success. Because of concerns about Mr Sage's age and state of health it was decided that he should be invited to swear an affidavit setting out his recollection of events, in case he was unavailable to give evidence at the trial. At an Orion board meeting on 9th October 1985 it was reported that Mr Sage had been most co-operative and was willing to swear the suggested affidavit.

    Mrs Good and her assistant Mr Reid met Mr Sage on 22nd November to discuss the proposed affidavit, and on 17th December 1985 he was sent a draft affidavit on which he was asked to comment. This draft was based on the statement prepared earlier in the year, supplemented by what had been said at the November meeting.

    Mr Sage responded on 18th December. He told Linklaters he had gone through the draft and had made alterations "which I believe truly reflect the position as I recall it". The draft had been composed in much the same terms as the earlier statement, and Mr Sage made very few amendments to it.

    Following a further meeting in March 1986, and some further exchanges with Mr Sage, the draft affidavit was approved by Mr Sage in April. In this draft the April 1975 agreement was described as a Termination Agreement rather than a Settlement Agreement. The third, fourth and fifth sentences of what had become paragraph 20 now read as follows:

    "This agreement was made subject to a reservation that if actual settlements increased "to an extreme degree", then the sharing arrangements would be re-opened. I remember that there was a discussion as to what would constitute "an extreme degree". All parties at that meeting agreed in line with the spirit of the Agreements, that if the then estimates proved to be significantly wrong then no single party should suffer."

    Apart from this change the three paragraphs describing the 1975 agreement were reproduced in substantially the same terms as they appeared a year earlier, except that the last two sentences of the main paragraph were completely excised. The documents before the court show that these words were removed at a late stage in the drafting process after a paragraph which had been inserted the previous autumn had also been removed. This had said: "I remember that the meeting, which had been very friendly, ended shortly before lunch. We all had drinks and then lunch in the Orion Board Room".

    Once again, Mr Sage said nothing during this renewed drafting process about recollecting that it had been expressly agreed at the meeting that any agreement that might be made would be of a goodwill nature and would not constitute a legal contract.

    Shortly before he approved his affidavit in these terms, Mr Sage had reiterated in writing to Linklaters his belief that the original agreement in 1966 to "save" Sphere and Drake was the underlying reason for all that happened later. The affidavit was now 18 pages long, and Mr Sage had suggested quite a large number of changes and additions at a fairly late stage of the drafting process.

    Before Mr Sage swore his affidavit on 6th May 1986 the word "Arrangement" was substituted for the word "Agreement" on the advice of junior counsel. In the instructions to counsel, Linklaters had explained that the affidavit was fuller and more detailed than might otherwise have been the case because of the concerns for Mr Sage's health.

    This evidence reflected the way in which Mr Sage recollected, at different moments between 1980 and 1986, the events of 1975. The next time he was invited to turn his mind to the circumstances of the April 1975 agreement was in 1989, nearly 14 years after the critical meeting. Mrs Good told the judge that after 1986 the action moved slowly because there was no real impetus behind it, or instructions from the client, for a long period.

    It is convenient to explain at this point that between 1984 and 1989 there had been a number of changes in Orion's senior personnel. Mr Burtonshaw left the company in September 1985, and Mr Fordyce, who had spent over 30 years in the insurance industry, joined Orion as Managing Director in May 1986. During that year Mrs Atkins, a young chartered accountant with Coopers and Lybrand, had spent a lot of time with Orion helping them to resolve their computer problems, and eventually she joined the company as Assistant General Manager (Accounts) in July 1987. In that capacity she was given day to day responsibility for liaison with Linklaters in connection with this litigation, reporting to Mr Fordyce. Mr Heritage retired in December 1986 and Mr Rousell retired as Orion's Assistant General Manager in October 1987.

    One consequence of all these changes was that nobody in Orion's senior management now had any direct knowledge of the events with which the litigation was concerned. They were heavily dependent on Mr Sage, who celebrated his 70th birthday in April 1988. Mrs Atkins described in her witness statement how Orion's organisation was so hierarchical that it would have been regarded as inappropriate for her to contact Mr Sage directly. She first met him at a lunch in Orion's boardroom to which the chairman, Mr Smith, had invited him in late November 1988. Although Mr Sage had previously asked for payment from Orion if he was to go on helping them, Mrs Atkins told counsel in December that Mr Sage appeared to be very pro-Orion and keen to help, and that he did not discuss money. Mr Fordyce, for his part, described in his witness statement how relations between Mr Sage and Orion had not been particularly good since his departure. He added that it was clear that preparing the matter for trial would need a lot of work from him. He recalled giving Mr Sage lunch at the City of London club two or three times during 1989 as part of a public relations exercise.

    The fifth stage in Mr Sumption's history is set against this background.

    The trial of the action was fixed to start in November 1989, and in late January 1989 Mrs Good set about the task of preparing witness statements in advance of the trial. On the morning of 13th March she had a long telephone conversation with Mr Sage in this connection. She recorded that:

    "[Mr Sage] is very spirited and alive to the issues and is anxious and pleased to help Orion. He tells me that he has no doubt whatsoever that the intention of the arrangement entered into in April 1975 was that if extreme degree was reached, the arrangement ... would fall away and the parties would go back to square one .... He is satisfied that the parties left open the opportunity to negotiate but that was negotiation from pool liability on existing percentages rather than from no liability for Sphere Drake. He is quite convinced that the whole spirit and intention of the pooling agreement (which was carried through to the arrangement in April 1975) was that no party would lose out and that accordingly in the event of extreme degree there would be recourse to Sphere Drake."

    Once again Mr Sage made no reference to any express agreement being made at the April 1975 meeting to the effect that any agreement they made would be only a goodwill agreement.

    Mrs Good sent him a file of documents and a copy of his affidavit on 17th March, and in early April 1989 she gave a thorough briefing on the case to Mr Henshaw. He had qualified as a solicitor in October 1986, and he was now rejoining Linklaters after a period in the United States. Mrs Good herself was due for transfer to the firm's Brussels office as soon as she could be released from London. The judge described Mr Henshaw as careful, straightforward and professional in the best sense of the word.

    On 19th April 1989 Mr Henshaw met Mr Sage for the first time, in the company of Mrs Good and of Mr McDonnell, who was another junior solicitor in the office. This meeting was planned to start at 10.30 and continue all day, if necessary. Mr Henshaw's notes of the meeting include references to Mr Sage saying that the 1966 arrangements were not overridden and that it was implicit that "we won't let each other down", that Sphere Drake had an obligation to pay the original proportions which were the starting point if and when clause 3 came into effect, that it was not such a big deal because nothing had changed and there was no abrogation of agreements, and that if it had been a final termination the Orion Board would have had to approve it and solicitors would have been used.

    Mr McDonnell's notes include references to

    "'relationship was a gentleman's club and not hurt any one'; 'appears legal agreement but friendly termination - all friendly or all legal'; 'what's said we not hurt each other. Implicit as 66/67 Agreement'; 'Any Agreement has the force of law'; 'nothing changed - administrative matters'; 23/4 meeting fairly short; 'Subject to Board Approval'; no visual recollection of the meeting; 'discussion of extreme degree: can recall nothing at meeting - only 10% afterwards'."

    Once again, there is no indication that Mr Sage had yet told his solicitors that he remembered saying anything expressly at the April 1975 meeting about the discussions being on a goodwill basis only.

    After that meeting Mr Henshaw took over the work of preparing Mr Sage's witness statement for exchange in advance of the trial.

    He was also involved in considering the documents that had been disclosed on discovery by Sphere Drake two years earlier, and on about 4th May he came across "the Flint Note" for the first time. Since it was written in shorthand, he had to ask his secretary to interpret it. Her interpretation of the first line of the note was that it read:

    "Leslie Sage: Goodwill agreement: not a legal contract."

    The discovery of the note did not cause Mr Henshaw to alter the wording of the current draft of Mr Sage's witness statement. Instead, he asked a number of questions about its significance in the relevant section of the draft, which ran to 31 paragraphs in all. (He also raised questions in the text about other matters). This draft read, so far as is material (starting at paragraph 27):

    " We discussed [Mr] Russell's memorandum dated 22nd April, and agreed on the figures and a series of payments to be made on a deferred basis .... we hoped and believed that the estimates of outstanding liabilities, on which the figures were based, would prove to be reasonably adequate. At the same time all parties recognised, as any prudent insurer would, that it was possible that actual liabilities might turn out to be seriously out of proportion to the estimates, and it was understood and agreed that, if this proved to be so, no single party should suffer. Consequently, we agreed that one party would not ask another party, pursuant to the Pooling agreements, for further contributions to pool business claims, unless the claims situation deteriorated to an extreme degree. There was some discussion of the meaning of "extreme degree", although the concept was not precisely formulated as between the parties. In the event that the claim situation did so deteriorate, then the liabilities under the Pooling agreements would continue. [Similarly, the contribution arrangements made in relation to the sale of shares in Sphere, and through the Manor reinsurance, would come back into play.] In this way, it would be ensured that no party would be out of pocket, consistent with the essential general principle underlying all these arrangements, including the Pooling agreements themselves, and the Sphere termination arrangements.

    28. Neither I nor any of the other Orion representatives at the meeting agreed to abrogate the Pooling agreements. We did not accept, or give any indication of accepting, that the "settlement" was in any way a final, irrevocable termination of the parties' underlying obligations to share liabilities in the appropriate proportions. In the event that "extreme degree" was reached, the parties would be obliged to meet those liabilities: this would be the starting point of any discussions of the kind referred to in paragraph 4 of [Sphere's] note of the meeting (referred to in paragraph 30 below).

    29. [Qs - the first line of the contemporaneous manuscript notes made at the meeting (Exhibit 39) (see also L&P's retyped version, immediately following the copy of the one in the file) appears to indicate that you said words to the effect of "Goodwill agreement - not a legal contract". Is that consistent with your recollection? If so, (a) were you referring to the termination agreement as a whole, or only the "let-out" clause? (b) what was the reaction of the other persons at the meeting? Are you able to reconstruct the approximate sense of the remainder of the note, (where the interpretation of the shorthand is unclear)? (This point is significant, because Orion, in its pleadings, does not admit that the arrangements made at the meeting were intended to be legally binding. It would be of great assistance if we could establish, with the note as corroboration, that the parties actually accepted this point at the meeting.)]

    30. Mr Flint, the Secretary of Sphere, subsequently circulated a note [the Russell Record] of the meeting together with the memorandum (Exhibit 40), and asked each party to initial them. [I] initialled the note for Orion...."

    Ten further lines of questions then followed.

    On 17th May Mrs Good sent Mr Sage a copy of this 29-page draft for his comments. She also enclosed with it forty exhibits. The Flint Note was No 39. She warned him that the draft statement would require a great deal more work and consideration and described it as a "working draft for further discussion". She described the present state of the litigation in some detail and invited him to a team meeting on 26th May when many different aspects of the litigation would be discussed, and at which Mr Smith, Mr Fordyce and Mrs Atkins were all likely to be present.

    On 18th May Linklaters sent instructions to counsel. Mr Henshaw prepared these instructions. They stated, as regards the April 1975 meeting, that Orion's position was that

    " if any binding agreement was reached, it was subject to a definite reservation that the pre-existing sharing arrangement would revive if "extreme degree" was reached."

    They continued:

    "The recollection of the Orion representatives at the meeting, Messrs. Sage, Rousell and Heritage, is that it was understood and agreed that if the agreed figures for outstanding liabilities did prove to be seriously under estimated, then no single party should suffer - ie, further contributions under the sharing arrangements would be made. Mr Sage's firm view is that the settlement reached at the meeting was not a final, irrevocable abrogation of the Pooling agreements, and if the claims situation did deteriorate to 'an extreme degree', Sphere and Drake would be obliged to make further contributions in accordance with the sharing arrangements.

    In terms of reconciling this with the note of the meeting, counsel will note that the part of the note, paragraph 2(a), which records the alleged agreement that the agreed figures represented the parties' final liabilities, is expressly made subject to the reservation in paragraph 3 of the note ....

    Orion has, in its pleadings, not admitted that a binding agreement was reached at the meeting. The documents produced on discovery include a manuscript note, apparently made at the meeting by Mr Flint of Sphere Drake. (Counsel's papers include, immediately following the copy of the note, a re-typed version prepared by instructing solicitors which attempts to translate the shorthand parts of the note.) The first line indicates that Mr Sage stated that the 'settlement', or possibly some part of it, was not intended to be legally binding. Instructing solicitors will be reviewing this note with Mr Sage as part of the work to be done on his witness statement, including the question of whether the comment referred to the arrangements as a whole (as opposed to, for example, the 'extreme degree' reservation only), and whether it was concurred in by the other persons present at the meeting.

    No reference to the arrangements or prior negotiations appears in the board minutes of Orion, Sphere or Drake."

    During the course of May, after taking counsel's advice, Orion agreed to pay Mr Sage, Mr Rousell and Mr Heritage at agreed hourly rates for all the time they devoted to this case in the lead-up to trial. There had been some quite acrimonious exchanges before this issue was resolved. Mr Sage was to be paid at the rate of £50 per hour. I will return to this matter later in this judgment.

    On 26th May 1989 Mr Sage attended the team meeting, along with Mr Rousell and Mr Heritage, Mr Smith, Mr Fordyce and Mrs Atkins on behalf of the present management of Orion, and the three Linklaters solicitors he had met the previous month. This meeting lasted seven hours. The judge found that it was most improbable that the Flint Note was referred to. He said it was not that type of meeting. This finding was not disputed on the appeal. Mr Sage told Linklaters again that Mr Grob, the chairman of Howdens, had had lunch with them after the meeting on 23rd April 1975 (he had mentioned this on 19th April), and this detail was included when his draft statement was next revised.

    Following the meeting Mr Sage sent Mr Smith, Mr Fordyce and Mrs Good a "memorandum" dated 29th May which said nothing about the Flint Note. He said he had not been able to stop thinking about the matter since the meeting. The memorandum repeated and amplified the point about administrative simplification. He also said that it had just come to his mind that "the most important and relevant point" which he should have recalled earlier was the verbal agreement between Philip Gilbert (in his capacity as Chairman of Howdens, Sphere, Drake and Orion) and Jan Van der Velden, a director of Orion and later the President of its Dutch parent company, in the 1960s to the effect that the shareholders of Sphere and Drake would bear losses incurred in proportion to their shareholdings, and that this verbal agreement "had been the underlying basis on which all decisions thereafter were made".

    On 14th June there was another team meeting, which was attended by almost everyone who had attended the previous one. Mr Sage's statement was the last of the seven items on the agenda, and Mr Henshaw appears to have reached paragraph 19 of the draft before time ran out. Once again the judge, basing himself on a note made by Mr McDonnell, found that nothing immediately material to the present issues was recorded as having been discussed. Again, this finding was not challenged on the appeal.

    Mr Legg, a senior assistant solicitor, had joined the Linklaters team in May, and on 11th July 1989 he and Mrs Good met Mr Sage to continue the preparation of his witness statement. He had with him a work-book in which he wrote notes which covered about a page and a half of the book. He also made a number of notes in the 4th July draft of Mr Sage's witness statement, which now ran to 32 pages. Paragraph 29 of the May 1989 draft had become paragraph 36 of the new draft.

    The events of this day assumed much greater significance on the appeal than appears to have been the case before the judge, and it is necessary to consider Mr Legg's notes in some detail.

    The annotations on the witness statement appear to contain a mixture of Mr Legg's notes for his internal purposes and items of fact which Mr Sage seems to have told them. For instance, on page 1 of the statement the name of the other joint general manager of Orion is added. Page 2, on the other hand, which contains a description of the original pooling arrangements, has a note against it: "agreed Statement of Facts?". Page 3, similarly, contains a note: "Effect - matters of law. Intention - he cannot know - do we want this anyway?" Some of the other notes reflect drafting changes to improve the flow of the text, or to correct typing errors.

    Some of the questions which Mr Henshaw had inserted in the text of the statement received answers and some did not. For instance, detailed questions on documents which came into existence in 1967 and 1970 remain unanswered in paragraphs 18 and 20, whereas a question in paragraph 22 is deleted and a note relating to this paragraph appears in the work-book. Notes based on what Mr Sage must have told the solicitors on that occasion continue to the very end of the statement. There are examples of these against paragraphs 22, 23, 26, 28, 31, 32, 35 and 37, occasionally supplemented by notes in the work-book. Against one of Mr Henshaw's questions, in paragraph 29, appears the note "Can't recall".

    This draft contained the story, which Mr Sage had told Linklaters on 26th May, that the April 1975 meeting took place in the morning, that they had lunch after the meeting at which they were joined by Mr Grob, and that there was a further discussion of "extreme degree" at the lunch. Mr Legg, however, noted at the top of the relevant page: "NB: Flint note times meeting at 3.00pm".

    The notes on the statement clearly show that on 11th July Mr Sage was making comments on the final seven paragraphs of his statement which covered the April 1975 meeting and its aftermath. Relevant notes include:

    [Q - do you remember why Grob was not there?] "Had another engagement". (Para 32).

    "Cannot recall agreement reached on 10% - certainly said it". (Para 33, relating to the lunch discussion).

    "Nor was it mentioned or raised by either party. The point was simply not in issue - it was not what the meeting was about" (para 35, relating to the fact that the Orion representatives did not agree to abrogate the Pooling Agreements).

    "If legal binding document intended would have involved solicitors" (para 36, above the list of questions which arose from the Flint Note).

    "Cannot recall any discussion on note. Seemed to reflect what was agreed on goodwill system - not a legal agreement" (para 37, relating to the circulation of the Russell Record, in answer to some of Mr Henshaw's questions in the text).

    "A note was not unexpected - unusual for it to be initialled [May acquire a quasi-legal status]. Was intended to be a lasting ag[reement]" (against two more of Mr Henshaw's questions in the text of para 37).

    Mr Legg's work-book note supplements the note he had made against paragraph 36. It reads:

    "para. 36 goodwill = good spirit between parties - beginning with Philip Gilbert (10/66) - no friction between parties.

    very important: 'not a legal contract' - wished to deal with outstanding matters in a way that was sensible - if problems then would expect contribution on basis of original pooling agreements as amended in accordance with note of 10/66.

    Third attempt at dealing with this problem - but got it wrong; first attempt was Sphere share sale".

    The next draft of Mr Sage's witness statement was prepared by Linklaters on 13th July 1989. The paragraphs relating to the April 1975 meeting and its aftermath were now numbered 37-45. Mr Henshaw's questions about the Flint Note, now in paragraph 43 of the renumbered text, remained unanswered.

    This draft now contained a new paragraph which read:

    "I am uncertain whether the meeting took place in the morning or afternoon but I note from what appears to be a note of the meeting, taken I believe by Mr Flint, that the meeting is timed at 3.00pm."

    In the next paragraph of the text there is a sentence which had appeared in square brackets in the previous draft. Although Mr Legg made no note on the text of that draft or in his work-book, the square brackets have now been removed.

    The passage relating to the discussion about the meaning of "extreme degree" has now been substantially rewritten. Paragraph 40 contains, in essence, an explanation of the text of the relevant part of the Russell Record. Paragraph 41 reads:

    "Although Mr Grob was not present at the meeting I do recall that I had some discussion with him as to the meaning of 'extreme degree'. Mr Grob may have joined me for lunch, possibly on 23rd April 1975. I suggested to him that if any parties' overall commitment diverged by more that 10% from the figures approved, that would constitute 'extreme degree'. I cannot recall Mr Grob's reaction to this suggestion, but so far as I can remember the meaning of 'extreme degree' was never precisely formulated."

    The passage dealing with the non-abrogation of the pooling agreements, which now appears in paragraph 42, has been redrafted to include the comments Mr Sage had made on 11th July. Paragraph 43 (the former paragraph 42) is, as I have said, unaltered, but there are significant changes to what is now paragraph 44, to reflect the answers which Mr Sage had given to the questions raised by Mr Henshaw in the text of the former paragraph 37. In his comments on the Russell Record, Mr Sage now says

    "Although it was not unusual for a note of a meeting to be prepared and circulated in this fashion, it was unusual for it to be initialled. I certainly did not consider that we were entering into a legally binding arrangement; that certainly was not my intention either at the meeting on 23 April 1975 or when I subsequently initialled the note on behalf of Orion. To my mind it was no more than a goodwill arrangement and as such the note seemed to me to reflect what was agreed at the meeting on 23 April."

    Although he had answered all Mr Henshaw's questions on the former paragraph 37 on 11th July, they are all reproduced again in this draft following the words ".. on 23 April".

    None of the alterations incorporated in this draft of the statement were enclosed in square brackets or otherwise marked so as to make it easier for Mr Sage to be invited to confirm their accuracy on the next occasion he was asked to consider the correctness of the draft.

    On 14th July Mrs Good sent Mr Sage a copy of this revised draft, and invited him to read through it very carefully to ensure that he was happy with it. On the same day she prepared a further set of instructions for counsel. She was mainly concerned to seek their advice on evidence.

    Page 4 of these instructions drew the attention of counsel to "the (seemingly) contemporaneous" Flint Note and to the understanding that the first line read "Goodwill agreement and not a legal contract". It was added that "this interpretation of the first line of the note, if correct, supports Orion's contention that the arrangement reached on 23rd April 1975 was not a binding legal agreement".

    Mrs Good had included with the instructions a copy of Mr Sage's revised statement. She made the following comment on this statement:

    "[Mr Sage] ... can speak to the events of the Sphere Share Sale Agreement, the Manor reinsurance arrangements and the events of 1974 and 1975, although it must be said that his recollection of all these events is not very clear. That said, his conviction that Sphere Drake would be liable in the event of extreme degree is strong."

    Five days later Mrs Good, Mr Henshaw and Mr Legg travelled to Orion's offices at Folkestone to continue the work of preparing witness statements for Mr Rousell and Mr Heritage. When asked about the Flint Note Mr Heritage said that the reference to a goodwill agreement did not ring much of a bell with him. Mr Rousell for his part could not recall any discussion about the matters which became paragraphs 3 and 4 of the Russell Record when he was asked about it.

    On 24th July Mrs Good, Mr Henshaw and Mr McDonnell attended a consultation with leading and junior counsel. Stewart Boyd QC, who had only recently been instructed in this matter for the first time, expressed the view that the Flint Note was arguably evidence of an understanding that the arrangement made at the April 1975 meeting was to be a goodwill agreement only. He therefore proposed that the Points of Reply should be appropriately amended.

    Two days later Mrs Good stated in a brief for the insurance experts whom she proposed to instruct for the trial that Orion's case was that the April 1975 agreement "was no more than a goodwill agreement", but that if it was legally binding it was subject to a reservation that in the event of extreme degree the parties would continue to be liable under the pooling agreements.

    Mr Sumption laid heavy stress on the fact that up to this time, over 14 years after the critical meeting, and throughout the five different stages of the history which he had showed to us, Mr Sage had never at any time told anybody that he remembered saying at the meeting words to the effect that any agreement that was made would be a goodwill agreement and not a legal agreement. He did not even say it at the meeting on 11th July 1989 when Mr Sumption said there must have been a general discussion of the issues raised in what was then paragraph 36 of his draft witness statement. Mr Sumption submitted that it was completely clear that the very first time Mr Sage professed to such a recollection was after he had received a letter which Mrs Good sent in identical terms to him and to Mr Rousell (with a copy to Mr Heritage) on 27th July. This letter included the following passages:

    " We will be serving our amended pleading (tomorrow) in response to the amended defence and counterclaim. I will let you have a copy of this in due course. The contents do not affect your evidence but rather reflect the evidence you have given us....

    The main purpose of this letter is to assist you in preparing your draft statement.... I have been going back through the discovery documents to see whether there are any further documents which may assist in reminding you of events at the time. I enclose the following:

    (1) The manuscript note seemingly prepared by Richard Flint during the meeting on 23rd April 1975. (You have, I believe, seen this before). It contains certain notes in shorthand. We are obtaining the opinion of a short-hand expert as to what these notes mean and we understand that the first line refers to LS having said:

    "Goodwill agreement and not a legal agreement."

    There are also various references to the "spirit" of the agreement. All of this is consistent with what you have told us about the intention of the meeting. It is vital to our case that we establish that what was agreed on 23rd April 1975 was a goodwill agreement only and not a legally binding full and final settlement. I trust this note may help to confirm what you have already told us....

    (3) A number of documents which make it clear that the meeting was first fixed to take place at 3pm on 23rd April in Ken Grob's office. At the last minute the arrangements were changed to Orion's offices but still at 3pm..."

    On 28th July Linklaters sent to Sphere Drake's solicitors their proposed Amended Points of Reply. Paragraphs 5 and 5A (a new paragraph added by amendment) of this pleading were in these terms:

    "5. It is admitted that the meeting referred to at paragraph 20 of the Amended Points of Defence occurred. Any such agreement as was reached at the meeting (as to which no admission is made) was a goodwill agreement only and not intended to be a legally binding agreement. Without prejudice to the foregoing denial, insofar as any legally binding agreement was reached at the said meeting it was agreed, inter alia, as follows:

    (a) That, subject to the general reservations set out at sub-para (b) below, the figures set out in sub-paragraph 20(a) of the Amended Points of Defence were deemed to represent the final liabilities of the parties to the Pooling agreements;

    (b) in the event that the actual settlements falling due to be made in respect of the run-off under the Pooling agreements thereafter increased to an extreme degree the commitment of the parties to the Pooling agreements or any of them, any such agreement as was made at the said meeting or otherwise would be determined and, in such circumstances and/or in default of any subsequent agreement otherwise, the arrangements prevailing prior thereto, namely the Pooling agreements as referred to in the Points of Claim, would continue to apply to the rights and obligations of the parties.

    5A Save insofar as hereinafter set out, the last two sentences of paragraph 20 of the Amended Defence are denied. A draft minute relating to the discussions of 23 April 1975 was drawn up by C. F. Russell acting on behalf of the Defendants. Some 5 to 7 weeks after the meeting, it was circulated and initialled. No admission is made as to the accuracy of the said minute. At no material time was it the intention of the parties that the said minute would constitute a written legally binding agreement. Alternatively, if and insofar as the effect of initialling and exchanging copies of the said minute was to constitute or evidence a written legally binding agreement (which is denied), the Defendants are estopped from alleging or relying upon the same. At the meeting of 23rd April 1975 it was expressly agreed and understood by all parties that the discussions which then occurred could only lead to a goodwill agreement which was not legally binding, and it was upon this basis that the minute thereof was initialled."

    Words added by way of amendment are in italics.

    On 1st August Mrs Good sent Mr Sage a copy of this pleading. Among other matters she pointed out to him that Orion was arguing at paragraph 5 that the April 1975 arrangement was a goodwill arrangement only, but that if it was binding it was subject to the reservation as to reopening.

    (It may be appropriate at this stage to mention that although unsuccessful at first instance Orion obtained leave from the Court of Appeal to make these amendments shortly before the trial before Hirst J began.)

    Mr Sage came to see Mrs Good on 7th August. He brought with him the latest draft of his witness statement, heavily annotated with deletions and suggested additions. In a covering letter he commented that the Flint Note supported goodwill and not legal agreement and that he had enlarged in his evidence. He added that Mr Comery's comment also referred to the spirit of the agreement.

    In notes written in the margin of the text of his statement, he answered the questions which Linklaters had posed to him three months earlier in these terms:

    "Is the Flint Note consistent with your recollection?" "Yes."

    "If so, (a) were you referring to the termination agreement as a whole, or only the let-out clause?" "Agreement as a whole."

    "(b) what was the reaction of the other persons present?" "They agreed."

    He added, on a separate sheet of paper, the following rider:

    "I recall clearly that I made the point that this, meaning the agreement as a whole, was a goodwill agreement and could not be a legal contract. I could not override the earlier agreements ratified by the Orion Board except by their authority. (This also, I would suggest, applied to all the parties). The reaction of the other parties was one of agreement."

    This was the origin of what became paragraph 65 of Mr Sage's witness statement at the trial. The words "I recall clearly that" remained in the draft witness statement when the sixth draft was prepared on 27th August. Mr Legg then annotated this draft with the words "how clearly". These four words did not appear in the next draft, prepared on 30th August.

    Mr Henshaw had conducted a very long proofing meeting with Mr Sage that day. He could not remember how the words came to be deleted. The relevant part of his evidence ran along these lines:

    "I mean, speaking from how I would go about this sort of change ... I would have wanted to find out really what I thought, or how clear I thought the recollection was. If, having done that, it seemed that to say 'how clearly' was overstating it, then I think it would have been natural to take that passage out ...

    Q. The result of that process on 30th August must have been that Mr Sage decided, after discussion with you, to delete the words 'I clearly recall'.

    A. Well, I think we are talking about two different issues here. One issue is whether he has an independent, let us say a refreshed, recollection. The other issue is whether to say 'clearly' is an overstatement. This seems to me to go to the second one.

    Q. Let us look at this in stages. First of all, we can assume, can we not, that if the words 'I clearly recall' disappeared, it is because Mr Sage considered, in the light of the discussions that you had on the point on 30th August, that those words were unjustified and had to be deleted?

    A. Well, if you are right and, as it appears if the discussion was of the question how clearly, then the process would have been, 'let us try to ascertain how clear this recollection is'. If the upshot was that it really is overdoing it to say 'clearly', which I think is quite natural given the lapse of time, then the phrase would have come out." (T8/48-50).

    Mr Henshaw went on to say the same thing in different words at T8/51 (lines 4-9) and T8/52 (Lines 6-8). At T8/53-54 he said:

    "All that is happening here, as far as I can tell, is that one is thinking about how clear this recollection is. We have concluded that 'clearly' is going too far, so the phrase has come out."

    At T8/55 (lines 12-16), he said:

    "Again, I do not remember the process, but if we remained satisfied that he had a recollection but we felt that to say 'clearly' was over-egging it, then I think what has happened here is the right end result."

    Even if Mr Henshaw had not given this evidence, it does not need much imagination to conclude that he must have told Mr Sage that he could not say that he recalled clearly making the point mentioned in his statement, with the result that those four words were omitted from the next draft. Given Mr Henshaw's evidence on the point, the conclusion is irresistible.

    In its final form, as signed by Mr Sage, the paragraph read:

    "I made the point at the outset that this (meaning the agreement as a whole) was a goodwill agreement and could not be a legal contract. (Indeed, I could not override the earlier agreements ratified by the Orion Board except with its authority and in the proper form. So far as I was aware, the other persons present were in the same position). I believe this was clearly understood and agreed by the other parties. My comment appears to be reflected in notes made by Mr Flint of Sphere Drake, partly in shorthand, of which I have subsequently seen a copy."

    The final sentence was subsequently deleted from the witness statement by order of Sir Peter Pain.

    The judge was of the view that he did not think the deletion of the words "I recall clearly" was of any significance. He said he could well understand why the words were considered inappropriate after 14 years. He accepted Mr Henshaw's evidence that the effect of their deletion was to put the resulting draft not outside the realm of recollection, but only outside the realm of clear recollection. He also accepted Mr Legg's evidence that this was a drafting point and nothing more.

    Mr Henshaw said, and the judge accepted, that Mr Sage's note "I recall clearly" represented the first occasion on which Mr Sage had said to him, and so far as he was aware to anyone, that he had actually said at the April 1975 meeting that the discussions could not lead to a legally binding agreement.

    It is convenient to mention at this stage that although Mrs Good had told Mr Sage what she understood the first line to mean, the expert shorthand writers whom she consulted on counsel's advice were by no means so positive about the meaning of the hieroglyphics between the words "Goodwill agreement" and the word "legal" in the first line. The first firm she consulted commented, in a letter dated 31st July, that the shorthand was not very good. The best they could do with the first line was "LS Goodwill agreement ? not a legal contract".

    A second firm was then instructed, and on 11th August a representative of that firm told Mr McDonnell that the note was essentially a summary of minute of a meeting that had not been written at speed. It was a fairly well written piece of basic shorthand but had not been written by an expert. He interpreted the first line as reading:

    "LS Goodwill agreement all [of] in legal contract."

    He was positive that the line did not say "not a legal agreement" and gave reasons for this view. He was most surprised that an interpretation of "not" had been given. He confirmed this opinion in a letter dated 16th August in which he expressed the view that the three shorthand outlines on the first line were more probably "all in a" than anything else. Again, he supported this opinion with technical reasons.

    By this time, however, Linklaters had contacted Mr Flint himself who interpreted the line as saying "goodwill agreement and not a legal contract" and this was the basis on which the matter went forward to trial.

    The only reason why I have mentioned this history is that when Mrs Good and Mr Henshaw were asked in 1998 about their reaction to Mr Sage's "I recall clearly..." statement, they both told the judge they were not surprised by it. The history shows, however, that on 7th August there was a justifiable doubt whether the Flint Note did indeed record the whole of what Mr Sage was professing he recalled clearly, and this doubt must surely have increased in the days that followed. This point was not, however, put to the witnesses. I only mention it to illustrate the problems that are bound to occur when witnesses are asked to recall their reactions to events so many years after they have happened.

    Mr Sumption also relied on part of the evidence relating to what came to be called "the Tardis meeting" in support of his contention that Mr Sage had no recollection of the April 1975 meeting until he was reminded of it by the Flint Note. This was a meeting which took place during the course of the hearing before the Court of Appeal in July 1991. As a result of some observations made by Stuart-Smith LJ the Orion team thought it would be helpful to convene a meeting at their office near the Law Courts (which was nicknamed "the Tardis") in order to consider whether there might be any merit in Orion waiving privilege so as to show the Court of Appeal that Mr Sage had been saying that the 1975 agreement was not legally binding long before the pleadings were amended to this effect in July 1989. Mrs Good flew over from Brussels to attend the meeting.

    The judge heard a great deal of evidence about this meeting. He set out his findings at pp 126-142 of his judgment. I will be considering aspects of the meeting in greater detail later in this judgment. It is sufficient for present purposes to record that Miss Barttram, who was at that time a trainee solicitor with Linklaters, made a lot of notes in manuscript as the meeting progressed on which she based the five-page minute which she and a junior solicitor then prepared, and that Mrs Good compiled her own six-page minute of the meeting four days later, following her return to Brussels. Mr Boyd attended the first part of the meeting, but once it had been agreed that there should be no waiver of privilege he left in order to continue preparing his submissions for the Court of Appeal the following day.

    Part of Mrs Good's note of the meeting is in these terms:

    " DFG [= Mrs Good] summarised by saying that Sage had always talked about the context of relations between Orion and Sphere Drake, the spirit of good faith and trust which were laid down in the golden rules of 1966 and the fact that he would never have entered into an agreement of this kind without a 'let out'. She was certain that Sage believes he had always told us that the agreement reached in l975 was a goodwill agreement. As he said in the witness box, he had it in his mind and it is possible that this was what he was always trying to explain in a rather unfocused manner. Stewart Boyd went on to say that once we had the Flint Note this was better evidence than Sage trying to remember what had been said. Sage had, in fact, gone too far in the witness box in saying that he recalled saying this at the outset of the meeting and had always said so."

    Mr Boyd denied that he had said that Mr Sage had gone too far in the witness-box, and the judge found that he might well be right in maintaining that this comment was not rightly attributed to him in Mrs Good's note. The reason for this observation is that Miss Barttram's contemporaneous manuscript note ascribed to the part of the meeting which had followed Mr Boyd's departure the words:

    "Prob - Sg carried away in W. Box. 'I recall and have always recalle