NATIONAL JUSTICE COMPANIA NAVIERA SA v PRUDENTIAL ASSURANCE CO LTD (THE "IKARIAN REEFER") COURT OF APPEAL (Civil Division) [1995] 1 Lloyd's Rep 455 HEARING-DATES: 11, 12, 13, 14, 17, 18, 19, 20, 21, 24, 25, 26, 27, 28, 31 October, 1, 2, 3, 4, 8, 9, 10, 11, 14, 15, 17, 18, 22, 23, 24, 25 November, 8 December 1994 8 December 1994 CATCHWORDS: Insurance (Marine) -- Constructive total loss -- Owners connivance -- Fire broke out on vessel -- Shipowner claimed under policy -- Underwriters alleged vessel deliberately set on fire with shipowners' connivance -- Whether allegation proved to relevant standard -- Responsibilities of expert witnesses -- Whether underwriters liable. HEADNOTE: The plaintiff company were the owners of Ikarian Reefer which was insured inter alia with the defendants. The vessel was insured against inter alia perils of the sea, fire and barratry. Under the policy the vessel was valued at US$3m of which 87.5 per cent was subscribed by the defendants. The plaintiffs formed part of the extensive shipping interests of the Comninos brothers. On Apr 12, 1985 at about 23 00 hours Ikarian Reefer ran aground on the Shoals off Sherbro Island, Sierra Leone in the course of a voyage from Kiel to Abidjan in ballast. At about Ol 00 on Apr 13 fire broke out in the engineroom and spread to the accommodation. At Ol 15 the vessel was abandoned and the crew were picked up at about 03 30 by the Yugoslavian vessel Ljubjana. The plaintiffs claimed under the policy contending that Ikarian Reefer became an actual or constructive total loss in consequence of a peril insured against ie fire (and or perils of the sea). The plaintiffs submitted that loss by fire included deliberate fire. If however the fire was accidental and if contrary to the plaintiffs' primary contention, the Court found that the fire on Ikarian Reefrr was the deliberate act of the master and crew the plaintiffs claimed a loss by barratry. The defendants argued that the vessel was wilfully cast away in that it was deliberately run aground and deliberately set on fire by or with the connivance of those beneficially interested in the plaintiffs. The defendants contended that it was to be inferred that the master, officers and crew would only have cast the vessel away on the instructions or with the connivance of her beneficial owners. The question for decision was had the defendants proved to the relevant standard that Ikarian Reefer was deliberately set on fire with the connivance of the owners. -- Held, by QB (Com Ct) (CRESSWELL, J),that (1) before the grounding there were two impacts and the master's perception at that time was that he was in deep water and in a safe position; however the actions taken by the master were not consistent with the conduct of a master intent on deliberately grounding his ship; the master did not consider he had struck the ground; and the master's errors were a failure to use the echo sounder, a failure to alter course to starboard for a longer period of time, a failure to slow down or stop and a failure to look at the satnav; and the grounding of Ikarian Reefer was not deliberate but due to the negligent navigation by the master; (2) the inspection of Ikarian Reefer and in particular of the quick closing stop valve and the area immediately surrounding the valve was materially inadequate; from the diesel oil service tank (DOST) diesel passed first through the valve; the source of fuel for the fire was diesel from the DOST and the valve was found to be fractured; it was not sufficient to assume that the valve had fractured as a result of the fire and a full and careful examination of other possible causes had not been carried out; (3) the main area of fire damage was in the generator flat in way of number 2 generator; the remaining fire damage in the engineroom spread from that area; the seat of the fire was in the vicinity of the number 2 generator although the evidence was inconclusive as to precisely where ignition first took place; there was only one seat of fire; (4) the tap, which was connected to the drop line through which diesel was fed, was found to be 80 per cent open after the fire; it was likely that prior to the casualty the tap was subject to wear particularly if it had been in use since 1968 and it was possible that it vibrated open (the vibration/fatigue theory); (5) if a deliberate fire was started in the manner alleged by the defendants it was difficult to account for the limited nature and extent of the damage to the tap and the generator flat; (6) on the evidence the defendants had not proved to the relevant standard that Ikarian Reefer was deliberately set on fire; (7) if Ikarian Reefer had been deliberately set on fire by a member of the crew the defendants had not proved that the owners in any way consented or were privy to that action; and if the burden of disproving privity lay on the owners they had discharged it; (8) the submission by the defendants that "it may be no more than standing instructions" that most Greek ship masters would know that if they had an accident their owners would far rather they had a total loss than a partial loss, would be rejected; there were no such standing instructions; and there was no evidence of any meeting between the owners and the master face to face at any time between November, 1984 and the casualty when instructions to cast the vessel away could have been given; there would be judgment for the owners. The defendant underwriters appealed claiming that (1) the grounding was deliberate and effectively the cause of the loss; (2) the fire was started deliberately; (3) both the grounding and the fire were caused with the connivance of the owners. -- Held, by CA (STUART-SMITH FARQUHARS0N and EVANS, LJJ), that (1) the principal findings by the learned Judge that the master was an honest witness could not be accepted; the master's evidence showed that he was an unsatisfactory witness and the Court had to make its own findings as to the events leading to the grounding and its own conclusions as to whether the grounding was deliberate (see p 461, col 1; p 474, col 1; p 476, cols 1 and 2; p 477, col 2; p 478, col 2; p 479, col 1; p 480, col 1; p 506, cols 1 and 2); (2) it was difficult if not impossible to conceive any reason why the master should have navigated the vessel as he did; on the facts and the evidence Ikarian Reefer was deliberately run aground (see p 465, col 2; p 471, cols 1 and 2; p 472, col 2; p 473, col 1; p 474, col 1; p 478, cols 1 and 2; p 506, cols 1 and 2); (3) the owners had failed to put forward a plausible explanation of how the tap might have opened accidentally and in the course of the fire and were unable to resist the clear inference that it was opened deliberately for the purpose of fuelling the fire (see p 485, col 1; p 486, col 2; p 506, col 2); (4) the vibration/fatigue theory required eight sequential steps; each of the required steps was either improbable or highly improbable and the theory was wholly implausible (see p 486, col 2; p 487, col 2; p 488, col 2; p 489, cols 1 and 2; p 490, cols 1 and 2; p 491, col 2; p 492, col 1; p 496, cols 1 and 2; p 506, col 2; p 507, col 2); (5) it was on the facts very unlikely that the fire was accidentally caused by a cigarette or burning rag (see p 492, col 2; p 493, col 1; p 506, col 2; p 507, col 1); (6) there were possible explanations for the fractured valve which were consistent with arson; the evidence was simply not sufficient to exclude all other explanations other than the moving of the plates and the vibration/fatigue theory; and on the evidence "moving of the plates" was a possible mechanism which explained the broken valve (see p 493, cols 1 and 2; p 495, col 2; p 496, cols 1 and 2; p 507, cols 1 and 2); (7) there was clear evidence of motive and a good reason to dispose of Ikarian Reefer by scuttling; the overwhelming inference was that the owners authorized the scuttling; the owners in January/March, 1985 were committed to disposing of Ikarian Reefer and were under considerable pressure from the market situation if not from the bank to do so sooner rather than later during the year; there was no evidence of any discussions as to what steps should be taken to dispose of Ikarian Reefer; and there was no evidence that the crew had any reason to act on their own account (see p 500, cols 1 and 2; p 501, cols 1 and 2; p 502, cols 1 and 2; 503, col 1; p 504, col 2; p 507, col 2; p 508, col 1). The appeal would be allowed. CASES-REF-TO: Akerhielm v De Mare, (HL) [1959] AC 789; Colonial Securities Trust Co v Massey, [1896] 1 QB 38; Elias Issias, The (1923) 15 Ll L Rep 186; Filiatra Legacy, The (CA) [1991] 2 Lloyd's Rep 337; Glasier v Rolb (1889) 42 Ch D 436; Ionna, The (1922)12 Ll L Rep 54; Mersey Docks and Harbour Board v Proctor, (HL) [1923] AC 253; Popi M, The (HL) [1985] 2 Lloyd's Rep 1; Powell v Streatham Manor Nursing Home, (HL) [1935] AC 243; Savage v Adam, (1895) WN 109; SS Hontestroom v SS Sagaporak, (HL) (1926) 25 Ll L Rep 377; [1927] AC 37. INTRODUCTION: This was an appeal by the defendants, Prudential Assurance Co Ltd from the decision of Mr Justice Cresswell ([1993] 2 Lloyd's Rep 68), given in favour of the plaintiffs, National Justice Compania Naviera SA and holding in effect that the plaintiffs' vessel Ikarian Reefer had not been scuttled with the privity of the owners and that the plaintiffs were entitled to claim, under the insurance policy issued by the defendants, for the loss of their vessel. The further facts are stated in the judgment of the Court which was delivered by Lord Justice Stuart-Smith. COUNSEL: G Pollock QC and S Kenny for the defendants; D Steel, QC and N Jacobs for the plaintiffs. JUDGMENT-READ: Judgment was reserved. Thursday Dec 8, 1994 PANEL: STUART-SMITH, FARQUHARSON, EVANS LJJ JUDGMENTBY-1: STUART-SMITH LJ JUDGMENT-1: STUART-SMITH LJ, (Reading the judgment of the court): For the reasons set out in the judgment of the Court, to which all three members of the Court have contributed, this appeal will be allowed and judgment will be entered for the defendants. 1. INTRODUCTION At about 23 00 hours on Apr 12, 1985 Ikarian Reefer bound for Abidjan ran aground on the Shoals of St Ann off the coast of Sierra Leone on the west coast of Africa. Efforts to refloat the vessel were unsuccessful. At about 01 00 hours on Apr 13, 1985 a fire broke out in the engineroom and swiftly spread to the accommodation area. At about 01 15 hours the master of the vessel gave the order to abandon ship. All the ship's company were accommodated in the life boats and some two hours later they were rescued by a Yugoslav vessel Ljubljana with which they had earlier been in radio contact. The Shoals of St Ann lie off an inhospitable coast and extend for some 46 miles. The seabed is made up of fine light brown sand formed by silt brought down by the nearby rivers. The Shoals are shallow and constitute a well-known hazard to shipping. The chart being used by the master at the time of the grounding included a specific warning about the unreliability of the depths recorded having regard to their shifting nature. There has never been a clear explanation from the master, who was navigating the vessel at the time, why it went aground or why it was in the vicinity of the Shoals. It was well equipped for navigation and its instruments included a Decca DS5 satellite navigator and an echo sounder as well as radar. Ljubuana carried the master and crew of Ikarian Reefer to Lome where they were repatriated to Greece a few days later. Ikarian Reefer refioated of her own accord and in the days following the casualty was boarded by local fishermen in the employ of a Greek company who claimed salvage. The difficulties in dealing with the salvors later caused a number of problems but on Apr 22, 1985 two fire experts Mr Cook and Dr Bound acting for the owners and the underwriters respectively were able to go aboard and make an inspection lasting two days. They discovered an open tap on the line of the diesel oil service tank (DOST) close to the seat of the fire and concluded that the fire had been started deliberately. For the next year the owners of the vessel who are the claimants in these proceedings were engaged in negotiations with the salvors and it was not until July, 1986 that an agreement was reached and the owners took steps to tow the vessel to Spain for scrap. Both owners and underwriters were anxious to make a further inspection because photographs taken by Mr Cook showed a fractured valve controlling the diesel line running from the DOST. However, before the vessel was taken in tow, she sank near her anchorage in circumstances which remain obscure and the broken valve was never recovered. The claimants are a Greek company in which two brothers, Costas and Anthony Comninos, are the controlling shareholders. Effectively they were the owners of Ikarian Reefer. They are both Greeks who operate their business out of Piraeus. At the relevant time they and the companies they controlled owned some 25 ships. Ikarian Reefer was the subject of a policy of marine insurance No 132875HD which covered it inter alia against perils of the sea, fire and barratry. The present appellants are the lead underwriters. The insured value of the vessel was US$3,000,000 with an agreed market value of US$450,000. When the present claim was made under the policy the underwriters rejected it and alleged that the vessel had been intentionally grounded and the fire started deliberately. The case came on for hearing before Mr Justice Cresswell on May 13, 1992 and the trial lasted 82 days. The Judge found that the owners had proved on a balance of probabilities that the grounding-of the vessel was accidental. He identified the remaining issue to be decided in these terms: Have the underwriters proved to the relevant standard that the 1karian Reefer was set on fire with the connivance of the owners? In the result he decided that the underwriters had failed to discharge that burden and found for the owners. The underwriters now appeal claiming: that the grounding was deliberate and effectively the cause of the loss; (2) that the fire was started deliberately; (3) that both the grounding and the fire were caused with the connivance of the owners. (A) The law The Judge accurately summarized the legal principles applicable to the case at pp 71-73 (references are to [1993] 2 Lloyd's Rep 68). There is no criticism of this statement of the law. At pars 8-10 he set out the approach of the Court in a case of this sort. While accepting that this is a correct statement, Mr Pollock, QC submitted that the Judge failed to follow his own precepts. His approach was, submitted Mr Pollock, fundamentally flawed in three important respects: (1) He was unduly impressed by the demeanour of the witnesses, particularly the master and Mr Tsakiridis, and failed to test their evidence against established facts and the probabilities of the case. (2) He appears to have decided the case piecemeal. Thus he concluded that the grounding was accidental, having at that stage only heard the evidence in relation to navigation, motive and connivance, whereas all three aspects of the case are closely interrelated. This finding had an important bearing on the Judge's conclusion as to the fire. He should not have made up his mind until he had heard all the evidence. (3) He appears to have approached the case on the basis that he should not reject the sworn evidence of the alleged conspirators unless the underwriters could prove that their accounts were impossible. (B) The approach of the Court of Appeal Mr Steel, QC, on behalf of the owners has made a number of submissions concerning the proper approach of this Court when hearing an appeal from a decision of a Judge sitting alone. (1) The burden of showing that the trial Judge was wrong lies on the appellant. This proposition does not admit of any doubt. As long ago as 1896, Lord Esher in Colonial Securities Trust Co Massey, [1896] 1 QB 38 at p 39 citing the words of Lord Justice Lopes in Savage v Adam, (1895) WN 109 said: Where a case tried by a Judge without a jury comes to the Court of Appeal the presumption is that the decision of the Court below on the facts was right, and that presumption must be displaced by the appellant. If he satisfactorily makes out that the Judge below was wrong then inasmuch as the appeal is in the nature of a rehearing, the decision should be reversed; if the case is left in doubt it is clearly the duty of the Court of Appeal not to disturb the decision of the Court below. (2) When questions of the credibility of witnesses who have given oral evidence arise the appellant must establish that the trial Judge was plainly wrong. Once again there is a long line of authority emphasizing the restricted nature of the Court of Appeal's power to interfere with a Judge's decision in these circumstances though in describing that power different expressions have been used. In SS Hontestroom v SS Sagaporak, (1926) 25 Ll L Rep 377 at p 381, col 2; [1927] AC 37 at 47 Lord Sumner said: None the less not to have seen the witnesses puts appellate Judges in a permanent position of disadvantage as against the trial Judge and unless it can be shown that he has failed to use or has palpably misused his advantage, the higher Court ought not to take the responsibility of reversing conclusions so arrived at merely on the results of their own comparisons and criticisms of the witnesses and of their own view of the probabilities of the case. Viscount Sankey in Powell v Sireatham Manor Nursing Home, [1935] AC 243 at p 249 said: On an appeal against a judgment of a Judge sitting alone the Court of Appeal will not set aside the judgment unless the appellant satisfies the Court that the Judge was wrong and that his decision ought to have been the other way. Where there has been a conflict of evidence the Court of Appeal will have special regard to the fact that the Judge saw the witnesses. Finally in Mersey Docks and Harbour Board v Proctor, [1923] AC 253 at p 258 Viscount Cave, LC, said: In such a case (where the Court is hearing an appeal from a Judge sitting alone) it is the duty of the Court of Appeal to make up its own mind not disregarding the judgment appealed from and giving special weight to that judgment in cases where the credibility of witnesses comes into question, but with full liability to draw its own inference from the facts proved or admitted and to decide accordingly. (3) When a party has been acquitted of fraud the decision in his favour should not be displaced except on the clearest grounds. This proposition is not in contest and is supported by the House of Lords in Akerhielm v De Mare, [1959] AC 789 at p 806 where the earlier authority of Glasier Rolb, (1889) 42 Ch D 436 is cited. Mr Steel submits that the present case is not one of these exceptional appeals where the Court on the basis of those authorities is entitled to substitute its own judgment for that of the trial Judge. He supports that submission by reference to some of the important issues of the case. Leaving aside for the moment questions of the burden of proof, these may be identified as: (1) Whether the master ran Ikarian Reefer aground, deliberately or through his gross negligence. (2) If the vessel was deliberately run aground were any other members of the crew privy to the plan to do so, and if so which? (3) Did somebody acting on the master's authority set fire to the vessel? (4) Were the two owners, the Comninos brothers, participants in the plan to scuttle and/or to fire the vessel? While it is the Judge's duty to take into account all the relevant circumstances Mr Steel contends that the resolution of these issues to a large extent depends on the Judge's assessment of the witnesses. The master was in the witness box over seven days, giving the Judge ample opportunity to assess his credibility, which was central to many of the issues raised. As has already been pointed out, the master's record was sound and perhaps distinguished and the question arises whether such a man would sacrifice his reputation by acting in the manner alleged by the underwriters. Such an issue is peculiarly suited to the Judge's assessment of the character and credibility of the witness. Both the Comninos brothers gave evidence. It was suggested by the underwriters that the master, if he scuttled the ship, could only be doing so for the benefit of the owners. The Judge again had to decide on their integrity. They too were in the witness box for some time and questioned about their business experience and financial resources. The evidence of the members of the crew was relevant to the above issues, particularly that of the chief engineer and Mr Tsakiridis. It was suggested on behalf of the underwriters at the trial that Mr Tsakiridis had set fire to the vessel and that the chief engineer was party to the conspiracy. The Judge, therefore, had to consider the integrity of these witnesses when making his findings on the truth of this allegation. 2. THE GROUNDING (A) The issue and burden of proof The issue here is whether the grounding was accidental, and therefore fortuitous, or whether it was brought about deliberately by the master's navigation during the period between 20 00, when he took over the watch, and about 23 00 when the ship went aground. So expressed, this is a straightforward factual issue. But the Court's approach has to take account of the legal requirements as to the burden of proof and of the substantive law governing the shipowners' claim and the insurers' defence. For the shipowners to succeed, the evidence has to establish that the grounding probably was fortuitous; this conclusion can co-exist with a residual possibility that it was deliberate (or in scientific terms, a low order of probability) because the plaintiffs are required to prove their case on "balance of probabilities" only. If the plaintiffs fail to discharge this burden, however, their claim under this head must fail, even if the insurers have alleged but failed to prove that the grounding was deliberate and the cause therefore remains uncertain: The Popi M, [1985] 2 Lloyd's Rep 1. In those circumstances, therefore, if perils of the seas was the only insured peril relied upon, it would be strictly unnecessary to determine whether the insurers had proved that the grounding was deliberate: the claim would fail in any event. But the present case is more complicated, by reason of the subsequent operation of another insured peril, fire, upon which the shipowners also rely, and so it becomes necessary, or at least desirable, to consider whether, if not accidental, the grounding is proved to have been caused deliberately, as the insurers allege. On this issue, the burden of proof rests unequivocally on the insurers, and the degree or standard of proof which the law requires makes the burden heavier than that which rests upon the shipowners. Although the same "balance of probabilities" test applies, the standard of proof required is commensurate with the gravity of the allegation made; and no more serious allegation can be made against the master of a ship, a trained and experienced professional who was responsible for its safety and for the lives and welfare of its crew. The Court therefore must take account of the likelihood or otherwise of the master of this vessel intending deliberately to run his vessel aground (per Lord Justice Mustill in The Filiatra Legacy, [1991] 2 Lloyd's Rep 337 at pp 365-366). We do not find it necessary to pursue the question, which may be no more than semantic, whether the burden of proof so described by reference to the balance of probabilities is different in practice from the criminal standard of "beyond reasonable doubt" and if so by how much. The burden of proof is not discharged, in our judgment, if the evidence fails to exclude a substantial, as opposed to a fanciful or remote possibility that the loss was accidental. But we bear in mind that, on the authorities, the burden which rests upon the insurers is derived from the civil, not the criminal standard, and that its nature is as described above. Mr Pollock for the appellant underwriters drew our attention to many passages in the judgment where the learned Judge referred to various "possibilities" and then proceeded to hold, as Mr Pollock submits, that because such a possibility existed therefore the burden of proof had not been discharged. In his submission, the correct approach is to consider the balance of probabilities overall, and the mere existence of an opposing possibility does not prevent the balance from tilting heavily and sufficiently far in favour of the insurers. We are prepared to accept this submission, and we think that there may be some instances in the judgment where the Judge appears to allow a mere possibility greater weight than perhaps it ought to have. But we do not consider that, in relation to the grounding issue, there was a fundamental misdirection of this sort. What can be said is that the judgment contains no discussion of the distinction which may have to be drawn between remote or fanciful possibilities, unsupported by any evidence, and what we have called above the kind of substantial, or substantiated, possibility which when it exists on the evidence has the effect of preventing the Court from holding that the burden of proof has been discharged. (B) The factual evidence directed towards this issue consisted of first-hand accounts from the master and various of the crew of the events of the day of the grounding, Apr 12, 1985, together with a substantially agreed history of the voyage which had begun at Hamburg on Apr 4. The destination was Abidjan on the Ivory Coast where it was said, and not disputed by the underwriters, that the master's eta was the morning of Apr 14. There was also a large body of expert evidence, concerned both with the navigation and with technical aspects of the vessel's navigation equipment, particularly the autopilot and the satnav (satellite navigation) receiver. (C) It is unnecessary to repeat the learned Judge's careful description of the satellite navigation system. Such equipment was relatively new on board merchant ships at the time and it had been fitted on board this vessel during the previous year. The master had had some experience of the system on other vessels and he had a good working knowledge of the particular type of receiver, but the Judge found, as he was entitled to do, that the master's attitude towards it was what might be called traditional; he relied upon it for periodic fixes of the vessel's position, which were accurate to within a few hundred yards if the receiver accepted and registered a "good fix", but he preferred to rely upon his own chartwork for his assessment of the DR (dead reckoning) position. This finding may have led the Judge, however, to pay less attention than he might have done to one of the receiver's most important functions. This is the continuous display of the latitude and longitude coordinates of the ship's current geographical position which, although a dead reckoning position derived from the most recent satnav fix accepted by the receiver, was marginally less subject to error than any chart calculation which the master might make. This was because the receiver's DR calculation of the course steered and the distance covered since the last fix was derived from the course actually steered during the period in question. This course information was fed into the receiver from the gyro compass automatically and continuously. So long as the vessel was being steered by the autopilot, the course information was the same as the navigator would use for the purposes of his chartwork, but if there was any period in manual steering then the accuracy of his chart calculations would depend upon his assessment of the course in fact steered. The receiver would receive precise information, and so in this respect its DR calculation would predictably be more accurate. The same did not apply in regard to the speed component, because at the time in question the vessel's log was not working and so there could be no automatic input of the speed through the water. The input therefore was manual, and if the navigator used the same estimate for his DR calculation the receiver and the chart in this respect would coincide. Subject to these potential but relatively minor sources of discrepancy, the satnav receiver provides for the navigator a continuous update of the vessel's current position. This is always a DR position but the margin of error, if the course and speed information are correct, between that and the vessel's actual position is limited to (a) any inaccuracy in the last "good" satnav fix -- this, as said above, will be negligible and (b) the effect of current, tide and wind during the period since the last fix. If that period has been short and the current etc is not exceptional, then this effect will have been correspondingly small. In the result, therefore, the satnav receiver provides a continuous reading of the vessel's existing DR position which, although making no allowance for current, wind and tide, is unlikely to be much different from the actual position even during the interlude between satnav fixes, unless conditions are exceptional or the interval is unusually long. The machine can be interrogated for details of the last satnav fix received and accepted, and for the expected time of the next fix, but, unless this is done, the receiver gives the navigator the vessel's current DR position and it displays the course and speed inputs which are entered at the time. The judgment contains several references to the receiver's ability to provide the master with a satnav fix at intervals when a successful satellite pass occurred, but none, we think, to the continuous display of the current DR position which, for the reasons given above, was likely to be subject only to minor errors and was available to the master and ship's officers throughout. (D) Satellite pass No 18 -- 19 14 or 1935? The satnav fix resulting from satellite pass No 18 was calculated by the receiver from signals received between 19 14 and 19 33 and was made available to the navigator, who at that time was the chief officer, at 19 35. It is important to bear in mind that the receiver at that time was displaying not the 19 14 fix position but the current (19 35) DR position derived from that fix. The chief officer's uncontroverted evidence is that the autopilot was engaged and that the vessel steered 137 deg (T) during the 19 minute period between 19 14 and 19 35. As regards speed, it can be inferred that the speed input was 17 or 18 knots from his evidence that he reprogrammed the receiver at 16 knots after receiving the fix. There is no reason to suppose that the vessel's speed through the water then was much different from 17-18 knots, and, if so, the 19 35 DR position differed from the actual position at that time only as a result of the effect of current and tidal streams during the 19 minute period after 19 14. It is not suggested that this can have exceeded, or even approached, one mile, and even if there was also some minor discrepancy between the speed input and the speed through the water, this would have no significant effect. So for practical purposes the chief officer, and the master, were likely to have taken the 19 35 DR reading as an accurate fix of the vessel's position at that time, and they were fully justified in doing so. Although various calculations by the expert witnesses (see below) focus on the likely position at 19 14 as the time of the fix, it is more realistic, in our view, to consider what DR position the satnav receiver is likely to have displayed at 19 35, based on the signals received between 19 14 and 19 33. The Judge held that the master was unaware of the 19 14/19 35 satfix, but for reasons which we give hereafter we cannot accept this finding. (E) The shipowners both at the trial and during the appeal relied upon an alternative case which depended upon some malfunction or misuse of the autopilot during the master's watch commencing at 20 00 on the night of the grounding. These suggested explanations, in our judgment, must be rejected and we shall give our reasons below. For present purposes we find and assume that the autopilot performed reliably throughout and that it was not misused either by the master or by the cadet who was on watch with him. (F) Factual evidence -- up to Apr 12 This evidence, as stated above, was not in dispute. It is derived mainly from the fair log book and is summarized in the first report of the shipowners' expert witness, Captain Cockcroft. It is relevant only to the vessel's speed and even for this purpose its value is limited by the facts that the vessel encountered adverse weather conditions in the English Channel (Apr 7-8) and that on Apr 10 and Il she is likely to have experienced a favourable current running in a south to south easterly direction at a rate of about 1/2 knot (see Captain Cockcroft's report at par. 4)2 Moreover, there is no evidence which we have seen of the precise shaft revolutions achieved during the previous days. The recorded distances and speeds for noon each day are as follows: Date Steaming time (hrs) Av speed (knots) Apr 5 15.5 18.26 Apr 6 25 17.88 Apr 7 25 12.00 Apr 8 24 15.12 Apr 9 24 17.92 Apr 10 24 18.41 Apr 11 24 18.67 From noon on April 11 to 04 00 on April 12 (the last recorded position -- see below) the total distance run was 282.3 miles av speed 17.64 knots. During this part of the voyage the predicted current remained southerly but its strength decreased to an estimated 1/4 knot. (G) Factual evidence Apr 12 (1) Sources The master and crew were rescued from lifeboats by the Yugoslav vessel Ljublkana shortly after they abandoned Ikarian Reefer. They disembarked at Lome in Togoland on Apr 16 and were then flown to Athens where they arrived the next day. Holman Fenwick & Willan were instructed to act on behalf of the owners, and Mr Charles Lowe began interviewing the master and crew-members at Piraeus on Apr 22. His notes of interviews are contained in Bundle 35. Ince & Co were instructed by the underwriters and from May 1-3 Mr Arditti conducted a second series of interviews in company with Mr Lowe. These were described as joint interviews, but it is clear that the master and crew were made available for interview by Mr Arditti on behalf of underwriters and that some of his questions were in the nature of a cross-examination, which was entirely proper, rather than a simple factfinding exercise. Mr Lowe was in attendance to make his own note of the answers and to intervene whenever necessary for the purpose of clarification or to avoid misunderstanding. His and Mr Arditti's respective notes of these interviews are contained in Bundle 25. Witness statements were prepared by the shipowners' solicitors (now Clifford Chance) for the purposes of the hearing, and in accordance with the current practice of the Commercial Court the statements of factual as well as expert witnesses were exchanged before the hearing. They were formally admitted in evidence under the Civil Evidence Act, 1968 whether or not the maker gave evidence as a witness at the trial, and in addition, parts but not all of the interview notes were similarly admitted. This last seems to have involved a process of editing which we find as surprising as it is clearly unsatisfactory; in the result, for the purposes of the appeal, the whole of the interview notes, the statements (Bundle 2) and the daily transcripts of oral evidence were all relied upon indiscriminately as admissible evidence, without regard to the formal route, if there was one, by which they achieved that status. (2) Navigation until 20 00 The log book records a position (Lat 11 deg N Long 17 deg 04 min W) at 0400 which is accepted as approximately correct. It is about 220 miles south of Cap Vert (Senegal) and off the mouth of the river Geba (shown on the chart as Portuguese Guinea). The Archipelago dos Bijagos lies off the estuary and the recorded position is about 15 miles seaward of the 10 fathom line surrounding these islands. The Shoals of St Ann where the vessel ran aground are between 250 and 300 miles south east of the 0400 position and the course from that position to about 20 miles off the Shoals is 137 deg (T). The whole of this distance is covered by Admiralty Chart No 1147 which the master was using, but Cape St Ann, from which the Shoals run in a WNW direction for more than 40 miles to what are described on the chart as the North West Patches, is at the right hand edge of the chart. The next chart eastwards is No 3139 and this has to be consulted before laying off the next course from an alter-course position off Cape St Ann. The watch-keeping officer between midnight and 04 00 was the second officer but the master's evidence was that he was on the bridge at 03 30 and supervised the alteration of course to 138 deg (G) which allowing for a 1 deg high gyro error was equivalent to 137 deg (T). These courses were entered in the log, together with the magnetic equivalent 150 deg (C) (variation 12 deg E). The chief officer (watches 04 00-08 00 and 16 00-20 00) and the second officer (12 00-16 00) both said that this course was maintained throughout the day, and it is not suggested that their evidence is incorrect. The master's evidence was that the same course was maintained throughout his forenoon watch (08 00-12 00) also. There is no evidence to suggest that this is incorrect but, as the underwriters submit there is no independent corroboration of the master's evidence in this respect and there is no record of the noon position. In our judgment, however, the suggestion that the course was altered even for a short period during this watch is wholly speculative, and no assumption should be made in the underwriter's favour that this may have occurred. The master also said that the course of 137 (T) was marked on the chart and that he estimated an alter-course position off Cape St Ann which would be reached during his evening (20 00-midnight) watch. Whether that position was marked on the chart, with or without an estimate of the time when the vessel would reach it, is an important issue which we will consider below, but as regards a track or course line of 137 deg (T) drawn from the 04 00 position the master's evidence is supported in the clearest possible terms by the chief officer. It is normal prudent navigation to draw such a course line, and there is no reason to doubt that in this respect the master's evidence is correct. More problematical is whether any and if so what speed was estimated for the purposes of the day's run, and what speed was in fact achieved until 20 00. The chart shows, however, that if the speed was estimated as 18 knots then the DR position (without allowance for current, tidal streams, stoppages, etc) for 20 00, a total distance of 288 miles from the 04 00 position, was almost exactly 20 miles off the Shoals. Moreover it could be an appropriate waypoint for an alteration of course, though not such as to clear the next headland, Cap Palmas, by about 10 miles, which the master said was what he intended to do. During the day there were various satellite passes which would, or should, have produced accurate satnav fixes at least once during each watch (see the judgment at pp 47-48). The evidence of the watchkeeping officers is unclear as to whether or not they received individual fixes, and if so when, and as to whether any fixes were plotted on the chart. It is unsurprising that they should have no clear recollection of what were entirely routine matters at the time, and likewise there is no suggestion from any of them that the course of 137 deg (T) was altered as the result of any such fix. We can see no reason for doubting their evidence that such fixes as were received and plotted showed that the vessel was, if anything, tending to east of the track but that no alteration of course was called for at any time during the day. The sole exception to this general dearth of direct evidence is the chief officer's recollection of the satnav fix received at 19 35. His evidence in his statement and at the trial was that he plotted the fix and that it showed a position some five-six miles east of the track marked on the chart; moreover, and importantly, that he calculated an average of 16 knots as the speed made good "since the last [position] fix." Assuming that this evidence is accurate, then the actual position of the vessel at 19 35 (strictly, as a 19 35 DR position derived from signals received between 19 14 and 19 33) can be plotted, assuming only (1) that the track on the chart referred to by the chief officer was 137 deg (T) from the 0400 position stated in the log, and (2) that the average speed calculation was based on the distance covered, not from some earlier fix, but from the 0400 position which, although strictly a DR position, understandably was treated as equivalent to a fix. In our judgment, both of these assumptions must be made, for the following reasons and subject to the following qualifications. First, the chief officer gives clear evidence that initially the track shown on the chart was 137 deg from the 04 00 position, and he does not suggest that this had been altered at any time before 19 35. Secondly, if the track or course line had been replotted on the basis of an earlier fix, and the fix had shown the vessel inside the original 137 deg (T) track, then either it showed a course different from 137 deg (T), in order to preserve the same distance off Cape St Ann, or if it remained as 137 deg (T) it showed a smaller distance off the Shoals than was originally allowed for. There is no evidence which suggests that either of these was done, and as regards the former, an alteration from 137 deg (T) is contrary to the evidence of all three witnesses, referred to above, and to the evidence of the cadet, referred to below, that the course steered at 20 00 was still 137 deg (T). The second assumption must also be made, in our judgment, for the following reason. If the chief officer calculated an average speed of 16 knots back to 04 00, then the total distance covered had been 249 1/3 miles (15 hrs 35 mins at 16 knots). If the calculation went back only to some time later than 04 00, say, a noon position, then if the average speed until that time, after 04 00, had been higher than 16 knots, then it follows that the 19 35 position was further south than the total distance of 249 miles from the 04 00 position calculated on the basis of an average speed of 16 knots overall. As will appear below, both parties' expert witnesses have attempted reconstructions of the track followed by the vessel before her grounding and there is no significant difference between them as to the distance covered during the period from 04 00 to 19 35, except possibly the owners' alternative case which does not depend on the 19 35 position. This distance was about 250 miles. It follows, in our judgment, that on the evidence of the chief officer, which we accept on the basis that it is approximately correct, the satnav fix at 19 35 showed the vessel five-six miles east of a 137 deg (T) course line from the 0400 position and not less than (say) 250 miles from that position. We emphasize "approximately correct" for a number of reasons, the most important of which is the risk of allowing these and similar reconstructions of the navigation to become too precise. Every position and every course line is subject to a margin of error, even a supposedly accurate satnav fix and a fortiori any DR positions derived from it. On a small-scale chart such as No 1147 even the width of a pencil line can be significant if precise accuracy is sought. Conversely, therefore, any findings as to position, course and speed must necessarily be subject to an appreciable margin of error. (3) Navigation after 20 00 Three significant events took place during this watch. There is also an issue as to whether or not a "good" satnav fix was received and accepted by the satnav receiver during satellite pass No 21 between 21 20 and 21 39 and available to the watch-keeper at 21 41. This will be dealt with separately below. As to the three events referred to, issues arise only as to their timing and in two cases their precise details. These events were: first, an alteration of course to port; secondly, two impacts felt on board the vessel during a short period, not more than 30 minutes, before the grounding; and thirdly, the grounding itself at a position which the master, after some discussion with the chief officer, reported by radio as Lat 07 deg 38 min N Long 13 deg 08 min W. This places the grounding at some miles west of Turtle Island, which is situated a short distance slightly north of west from Sherbro Island (marked on Chart No 1147 as Cape St Ann). These can conveniently be taken in reverse order, for reasons which will become apparent when the expert evidence is considered. (i) Grounding position The grounding position was fixed, according to the evidence of the master and chief officer, by radar bearings and distance from Turtle Island and some other nearby islets. These were distinguishable, according to the chart, not because the land itself is particularly prominent, but because of tall trees which, it is said, were likely to give firm radar echoes. Neither the master nor chief officer accept that the satnav receiver DR position was noted, and this is understandable, perhaps, if the master believed, or told the chief officer, that it was unreliable and had let him down. While it is questionable, therefore, whether radar echoes could have given a reliable position, there is no basis in our judgment for rejecting their evidence that they did. There is also evidence that the satnav did later confirm that the position was approximately correct. The chief officer said that after the vessel went aground he reprogrammed the receiver at zero speed, which clearly was the correct thing to do, expecting that it would receive and accept signals during satellite pass No 25 between 23 51 and 00 10 (available 00 12). This duly occurred and their evidence is that the resulting display "confirmed" the position that had been given by radio before midnight. The inference seems clear that there was not such a difference between the reported and the satnav positions as to make it necessary to send a correcting signal, and in these circumstances the reported position apparently was regarded as approximately correct. The shipowners' expert witness, Captain Cockcroft, however, relies upon evidence from the rescuing vessel Ljubljana to support his opinion that the grounding was not at the reported position but 2.25 miles south west of it. The learned Judge found that that was indeed where the grounding occurred. With respect to him, we cannot regard even the Ljubljana evidence as sufficient to support this conclusion, and there is no other evidence of it apart from Captain Cockcroft's suggested reconstruction of how the grounding occurred; but his conclusion cannot be regarded as evidence supporting the assumption which he made. Because of the wide margins of error involved, it is easy to exaggerate the importance of this point, but we should set out the Ljubljana evidence shortly and our view as to the strength of the assumption which Captain Cockcroft has made. The Ljubljana logbook contains a narrative from 23 50 when a distress signal was received until 02 15 when she anchored in a depth of approximately three shackles (45 fathoms) at what is recorded as three nautical miles from Ikarian Reefer. It is clear, and accepted by the shipowners, that three shackles must refer to the length of cable used rather than the depth of water at the point of anchoring. This would be consistent with a water depth of 15 fathoms or less, and the logbook also records: The shallow water does not allow us to get any closer. Ljubliana drew about six metres, and so this suggests a depth which was less rather than more than 10 fathoms. The position recorded in the log is just outside the 10 fathom line shown on the chart (07 deg 32 min N 13 deg 07 min W). This is, however, more than six miles from the reported position where Ikarian Reefer was aground. The shipowners submit, therefore, that the "3 miles distance" estimate recorded in the log should be given weight and that the grounding position should be located, not three miles away, but in the position suggested by Captain Cockcroft, which is five miles north west of Ljubljana's anchorage position as recorded in its log. Against this, a distance of three miles from the grounding position reported by the master puts Ljubljana anchorage on or just outside the six fathom line, which is consistent with the reference to shallow water preventing her from approaching closer. At this stage, we should record that all the charts of this area, including Nos 601 and 685 on a large scale as well as Nos 1147 and 3139, give soundings based on surveys carried out in the first half of the 19th C (c 1839). Neither party suggests that precise reliance can be placed on the charted depths, particularly in an area of sandy bottom off the mouth of the Sherbro river and subject to significant tidal streams. Therefore, this evidence as to the depth of the water is of limited assistance in fixing the exact position where Ljubljana anchored. It should also be noted that she was not fitted with satnav and her recorded anchoring position must have been derived either from dead reckoning coupled with use of the echo sounder, or from the use of radar, or (most likely) a combination of both. In either event, the position cannot be regarded as precisely accurate, and there is no evidence as to how it was arrived at. In our view, there is insufficient evidence to displace the inference which arises from the ship's evidence that the grounding position is unlikely to have been significantly different from the position reported at the time; but, because of the uncertainties involved, it is important to bear in mind that this position is subject to some margin of error. No one suggests, however, that the grounding position was further north than was recorded at the time. (ii) The two pre-grounding impacts These are of little, if any, significance in relation to the navigation of the vessel during the watch, and they will be considered separately below. Their only possible relevance in the present context arises from the facts that the southern edge of the Shoals runs along a line not much different from 115 deg/295 deg and, according to the chart (subject to the reservations noted above), the bottom shelves fairly rapidly if the area of the grounding is approached directly from seaward. Therefore, if there were earlier groundings during a period of up to 30 minutes earlier, during which the vessel covered at least eight miles and perhaps more, these were more consistent with steering a course of 115 deg rather than, say 095 deg during the period before. Some such course was predicated as part of the shipowners' alternative case. Their primary case, however, relies upon the master's evidence that the course steered after the alteration from 137 deg (T) was 115 deg (G). This was corroborated by the cadet who was on watch with him, though absent on occasions from the bridge (see below), and it is further supported, to the limited extent described above, by the fact that two earlier impacts, if they were groundings, occurred. (iii) The alteration of course The time when this occurred is a factual issue of prime importance, for the following reasons. First, it bears directly on the reliability of the master's account of events during his watch, and secondly, the time gives some indication of the likely position of the vessel when it turned from the course of 137 deg (T) to steer the course which led to the grounding position. The master made entries in the fair log book after he returned to Athens. These were not copied from any chart or record which had been brought from the vessel, and they were the master's first written account of what had occurred during his watch. The relevant entry was this: (Time) Course (True) Course (Gyro) (Remarks) 22.00 115 At position 7o 34' N and 13o 42' W, changed course This became known as the master's 22 00 DR position. It is clear that the vessel never was there in fact. During his first interview by Mr Lowe the master said that this was a satnav DR position: . . . which I checked on DR against previous fix . . . using the same speed made good of 18 knots [35/1-18.] He gave this as the time and DR position of the alteration of course to 115 deg. He had also said that the last fix was at 17 00 (35/1-2). It was pointed out to him at the joint interview, however, that the 2200 DR position was 34 miles from the grounding position; the vessel therefore could not have been there at that time. My opinion in hindsight about 22 00 pos is that pos not good. [25/1-22.] The master's witness statement repeats that to the best of his recollection the last position fix obtained before this was between 17 00 and 18 00 (par 61). He had no recollection of a 19 30 fix (par 64). He then described how after his return to Piraeus he had tried to plot the navigation of his ship up to the estimated time of grounding, which he put at 23 00, and: I calculated as best I could that I should have been in a position about 07o 34' N 13o 42' W at a time I estimated to have been about 22 00 and plotted it on chart 1147 [par 74]. He continued: The 22 00 position and the timing of it are unreliable. I calculated a time and position to the best of my recollection, but I cannot recall the basis for it . . . [par 76]. He no longer claimed, therefore, that the ship was in fact in that position at the time of the alteration of course, or indeed that the time of the alteration was 22 00. Both the position and the time were unreliable. His evidence at the trial was that he had believed that the vessel was in deep water and in about the position shown on the 22 00 DR, but he accepted that he must have turned to 115 deg some time earlier than that. By the time of the trial, the shipowners' expert witness Captain Cockcroft had made his first attempt at reconstructing the vessel's track during the day and up to the time of the grounding. He had estimated that the course had been altered to 115 deg at 21 00. The master said in his evidence in chief that he could not remember at precisely what time he altered course (Day 16, p 36) and he was not asked whether he could remember approximately when it was, although he said that the log book entry of 22 00 was not accurate (Day 17, p 5). Later, however, he said that: It had emerged during the course of subsequent investigations . . . that it must have been at about 20 30 that something happened. The cadet who was on watch with the master confirmed the estimate of 20 30 which the master was prepared to accept. He did so with some corroborative detail (which the master agreed: statement par 77), to the effect that the master himself had made the alteration just when the cadet was leaving the bridge to obtain tea or coffee for them. This was either shortly after the beginning of the watch, at a time which the cadet put as early as 20 20/20 30, or an hour or so after that. But a time as late as 21 20/21 30 was inconsistent with the master's evidence, whereas the earlier time of 20 20/20 30 coincided almost exactly with it. It is confirmed also by the fact that an alteration 20 minutes or so after the beginning of the watch is entirely consistent with the time the master might take to do the necessary chartwork and reach his decision to alter course. It is difficult to see why, after this evidence was given by the only two eye-witnesses of what occurred, the trial proceeded on the basis that the alteration took place, or might have taken place, as late as 21 00. The appeal was opened on the basis that the time was between 20 30 and 21 00 and that the precise time probably is immaterial. But the ship travelling at 18 knots, if it was at full speed, covered not less than 12 miles through the water between 20 20 and 21 00, and if any reconstruction is attempted, the difference cannot be insignificant. Nevertheless, after the factual evidence was given, Captain Cockcroft continued to assume that the course was altered at 21 00 (drawing AC1 refers). Strangely, perhaps, in connection with an alternative contention that the autopilot was misused, he assumed a time of 20 45. The insurer's expert, Captain Third, preferred the time of 20 30 and his suggested reconstruction allowed for a time of 20 20/20 30. These inconsistencies seem never to have been resolved in the course of their evidence, and as stated above the appeal was opened on the basis that the precise time between 20 30 and 21 00 is immaterial. The Judge's finding that the vessel probably followed the courses described in Captain Cockcroft's plot AC1 (judgment p 69) adopted by inference the time of 21 00. We are satisfied, for reasons that will be given below, that we are entitled, indeed bound, to form our own view on these issues of fact; and we are also satisfied, for the reasons given above, that the time when the master altered course to port was 20 20/20 30. What was the course? The master's evidence is that the autopilot was set to steer 115 deg. This is confirmed by the cadet, and the expert witnesses for both parties prepared their reconstruction of the ship's track on the basis that it was correct. There is no reason, therefore, for doubting this evidence and it should also be noted that the gyro was subject to a fixed error (known as the delta error) which meant that the course in fact steered, subject to variable errors, was about .5 deg higher -- say 115.5 deg (T). But the log entry at 04 00 suggests that there was an overall gyro error of 1 deg high, which would mean that the course in fact steered was 114 deg (T). (H) Where did the vessel alter course? (a) Direct evidence The only direct evidence, apart from the fact that the course steered from this position led to the vessel grounding where she did at about 23 00, is the chief officer's recollection that the position of the 17 14/17 35 fix was five-six miles east of the track or courseline marked on the chart and showed an overall average speed of (or slightly in excess of) 16 knots. Some uncertainties are inherent in this evidence (see above) but if the figure of five-six miles is accepted and if 16 knots was the average from the 04 00 position, then a precise position can be plotted for 19 35: by our reckoning it is near 08 deg 02 min N 14 deg 08 min W. If the course of 137 deg (T) is projected from this position, DR positions can be plotted for 20 20/20 30 (the time of the alteration of course) and 22 00, depending on what estimate is made of the vessel's speed. If the speed is taken as 16 knots then the resulting 22 00 DR position is such that a course of 118 deg, which is what the master said that he had planned in order to clear the next headland, Cap Palmas, or even 115 deg which is what he steered in fact, is not immediately unsafe. The position is 288 miles from the 04 00 position (18 hours at 16 knots) and although it is about six miles to eastward of the original (137 deg (T)) track, nevertheless it is some 15 miles clear of the Shoals of St Ann. If the vessel had been there at the time of the alteration of course, whenever precisely the alteration took place, then she would not have run aground where she did. This was in fact the 22 00 DR position which the master recorded afterwards in the log, and it seems clear to us that he had calculated the log entry in this way. No other basis on which it could have been calculated before the log entry was made can be identified, and the only alternative explanation, unless the master carried the precise latitude/longitude coordinates in his head, which has not been suggested, is that the master chose it at random as a position in deep water where he could safely allege that he had believed the vessel to be. If so, then the fact that his random choice alighted upon the exact position which can be calculated from the chief officer's evidence is a remarkable coincidence. This coincidence, however, was not explored at the trial; the only evidence of the vessel's likely position at the time of the 19 14/19 35 fix were the expert witnesses' reconstructions of the courses and speeds which led to her going aground. These were projected back to 19 14/19 35, but unfortunately neither of the final versions (plot AC1 for Captain Cockcroft (shipowners), plot 1-18 for Captain Third (underwriters)) resulted in a 19 14/19 35 "fix" position which could be reconciled with the chief officer's evidence that the fix showed the vessel five-six miles east of the track. It next becomes necessary, therefore, to consider the experts' evidence and the Judge's conclusion that Captain Cockcroft's reconstruction was to be preferred. Before doing so, however, we should note that the master's 22 00 DR position is five-six miles to the east of a 20 00 DR on the original (137 deg (T)) courseline calculated at a speed of 18 knots from the 04 00 position. This is because 16 hours to 22 00 at 18 knots also produces 288 miles. The only differences, therefore, are the revised time (22 00) and the revised courseline (five-six miles to eastward), both of which can be derived from the chief officer's evidence, including his estimate of the speed of 16 knots which he used to reprogramme the satnav receiver shortly after 19 35. This makes the master's account consistent overall, even if the fact that the vessel cannot have been at the 22 00 DR position at that time makes it "unreliable" (his word) in that respect. At 04 00 he had laid off a course of 137 deg (T) which at a speed of 18 knots showed a DR estimate for 20 00 at a possible alter-course position off Cape St Ann. When he took over the watch, the revised eta at that position at 16 knots was 22 00 and the course was five-six miles inside the original track. Nevertheless, it remained a safe course and speed and it would have continued to do so if maintained until 22 00, as the master originally contended that it was. In the result, however, the master was unable to give any direct evidence of where the vessel in fact was when he altered course to 115 deg. (b) Expert evidence Both of the expert witnesses produced chart extracts which sought to reconstruct the likely track of the vessel from the time when the satnav fix which was noted by the chief officer at 19 35 was received. They did this on the basis that the courses steered were 137 deg (T) followed by 115 deg (G) and that the "full ahead" speed of the vessel was about 17.5 knots. The accuracy, however, both of the reconstruction and of the estimated position at 19 14/19 35 depended on the validity of their other assumptions, and there were differences between them in almost every other respect. Apart from the time of the alteration of course (Captain Cockcroft 21 00, Captain Third 20 20/20 30), the differences were these: Grounding time (1) 23 00 (Captain Cockcroft) (2) 23 00 or 23 10 (Captain Third) Grounding position Estimated as 2.25 SW of the reported position, by reference to (but not adopting) the Ljubljana evidence (Captain Cockcroft) (2) As reported (Captain Third) Current and Tidal Stream (1) (Captain Cockcroft) 19 14/21 00 Tide setting 070 deg 0.4 knots Current 040 deg 0.7 knots 21 00/23 00 aide setting 033 deg 1 knot Current 090 deg 0.5 knot (2) (Captain Third) No allowance alternatively 19 30/21 30 Combined effect 112 deg 1 knot 21 30/23 10 Tidal Stream 033 deg 1 knot Current nil Speed (1) (Captain Cockcroft) (subject to a reduction see below) 19 14/21 00 27 miles = 17.5 knots 21 00/23 00 35 miles = 17.5 knots (2) (Captain Third) 16 (or 18 knots) throughout alternatively 20 20/23 10 1 hour at 12.4 knots plus 1 hour 50 mins at 17.5 knots. (In addition, Captain Cockcroft worked on largescale Chart No 601, Captain Third on Chart No 1147.) The resulting positions for 19 14 (Captain Cockcroft) and 19 30 (Captain Third) were all to the east of the courseline or track drawn at 137 deg (T) from the 04 00 position (which was referred to by the master and the chief officer, and which we will call "the track") by considerably more than the five-six miles spoken of by the chief officer when he plotted the 19 14/19 35 fix. Inevitably, the positions calculated by Captain Cockcroft were to the west of those suggested by Captain Third, because he allowed for north-easterly and easterly currents and tidal streams, which had the effect of increasing the distance which he estimated the vessel had covered eastwards and south-eastwards after 19 14. Therefore, his 19 14 starting point was further west. The assumed grounding position being to the south (and west) of the reported position, this assumption similarly moved the calculated 19 14 position a little to the south and to the west. As for the vessel's speed through the water, if this was no more than 16 knots (used by Captain Third as one basis of calculation) then the distance covered between 19 14 (or 19 35) and the grounding was reduced, and the calculated 19 14 position was correspondingly further east than if a full speed of 17.5 knots (Captain Cockcroft) or 18 knots (Captain Third) was assumed. There was, in addition, a factual issue as to the extent of a reduction in speed which, according to the ship's evidence, became necessary for some period between 20 00 and 21 00 because of the need to attend to an overheated stern bearing. At one extreme, this could have caused a reduction to 70 rpm or about 12 knots for the full hour, say 5.5 or six miles; at the other extreme, the underwriters doubted whether any appreciable reduction took place at any time. In these circumstances, it is not clear why so much time and energy was directed towards the minutiae of these different calculations. Many approximations and uncertainties are involved, and any conclusion as to the likely position of the vessel at 19 14/19 35 must necessarily be subject to a considerable margin of error. Nevertheless, the shipowners were concerned to establish that the position was sufficiently far west, and south, for them to be able to demonstrate that the master could have believed that it was safe to alter course to 115 deg when he in fact did so. The underwriters, for their part, were equally concerned to show that the position was further east and further north than the shipowners suggested, so that the course of 115 deg led directly to the position where the vessel in fact went aground. The easternmost of Captain Third's 19 30 positions, however, was open to a specific objection; if the satnav fix had shown the vessel already that far east, up to 20 miles east of the track and dangerously close to the Shoals, known there as the Orontes bank, then the chief officer would have to be criticized for not taking immediate steps to remedy the situation whilst he was on watch. The underwriters did not suggest that he was a party to any plot to run the vessel aground, and there are no grounds for doubting his competence as a watchkeeping officer and navigator. Indeed, the underwriters rely upon his qualities in this respect in support of their allegation that the master was made fully aware of the 19 14/19 35 fix. So this part of Captain Third's reconstruction, which was based on a 16 knot speed estimate and no current or tidal stream effects (drawing 2-18), can be put to one side. This conclusion may carry with it an inference that the speed was greater than 16 knots and/or that current and/or tidal streams did have some easterly and perhaps north-easterly effects. It thus becomes possible to assert, as Mr Pollock did, that Captain Third's alternative and more westerly positions do not differ significantly from Captain Cockcroft's calculations, and that such difference as remains is immaterial to the issues which arise for decision in the case. On the other hand, it is the underwriters who allege that the master altered course deliberately so that the vessel would run aground, as in due course it did, and we do not understand how this allegation can be maintained unless the evidence shows approximately where the vessel was when the alteration of course was made. Moreover, the shipowners, through Captain Cockcroft, put forward an explanation of how the master came to make that decision when he did, and the validity of the explanation must be considered carefully before any conclusion can be reached. The explanation is depicted graphically on drawing AC1 which therefore has two separate functions: first, to reconstruct the position of the 19 14/19 35 fix, and, secondly, to indicate what the master's chartwork and DR calculations, derived from that fix, might have been. In essence, the explanation is this. If the master drew in the 137 deg (T) courseline from the (calculated) position of that fix, took the vessel's speed as 18 knots, and made no allowance for current and tidal streams, then the resulting DR position at 21 00 was one where it appeared sale to alter course to 115 deg, assuming again that no allowance was made in the choice of course for any current or tidal stream setting to the north-east. This suggested explanation does not account for the fact that any competent and experienced navigator, as the master was, could hardly fail to consider the likely effects of current and tidal stream in a situation where on his own evidence the master was taking the vessel closer to the shore than the 15-20 miles which would customarily be allowed. This objection does not seem to have been raised by the underwriters, either in cross-examination or in Captain Third's evidence. All attention was focused on the differences between Captain Cockcroft and Captain Third as regards their estimates of current and tidal stream, and as regards the vessel's speed. We do not think that it is an exaggeration to say that the parties seem to have become more intent on winning the battle of the experts than on establishing the facts upon which their respective cases were based. During the hearing of the appeal, Mr Steel produced at our request a modified version of drawing AC1, labelled DS1 for ease of reference. This shows as an alternative assumption that the alteration of course to 115 deg was made at 20 30, not 21 00. This results in a calculated 19 14 position rather more than eight miles from the original 137 deg track; but if the time is assumed to have been 20 20, the earliest spoken of by the cadet, the distance is reduced (by our calculation) to 6.5 miles. This is little more than the chief officer's evidence of the actual position five-six miles off the track. It is clear from this plethora of detailed and disputed evidence that no a priori conclusions can be drawn as to the actual effects of current and tidal stream; nor any, within rather narrower limits, in regard to the vessel's speed. We find it surprising that the underwriters have sought to maintain the contention that there was no significant tidal stream setting in north-easterly direction towards the river even when the vessel was only a few miles or less off the sands and the tide was flooding up to a possible maximum of 1.5 knots (taken from Chart No 685: low water was between 20 00 and 20 30). If there was none, then it follows that the vessel could only have run aground as the result of steering towards the grounding position; if there was some north-easterly set during the period from 20 00 until 23 00 (drawings ACl and DSl suggest that it was as much as six miles), then the question remains, why did the master fail to take any account of current and tidal streams which had the effect of putting the ship that much north of her intended track? (c) Conclusions Reverting to the respective calculations of the position of the 19 14/19 35 satnav fix, we find ourselves in broad agreement with the Judge's conclusion that drawing AC1 (now DS1) should be preferred as a reconstruction of the vessel's track during the period from the 19 14/19 35 fix until the grounding at 23 00 or 23 10, but assuming that the alteration of course was at 20 20/20 30, not 21 00. Captain Third's revised drawings (see below) are not put forward, as we understood Mr Pollock's submission, as a suggested reconstruction; insofar as his evidence is relied upon to support the contention that current and tidal streams had only a negligible effect on the vessel's track, we do not find his evidence persuasive. In our judgment, any conclusion as to the vessel's position at the time of the fix must take account not only of the expert evidence but also of the only direct evidence, that of the chief officer to which we have referred. Before stating our conclusions, we should state one reason why we consider that the chief officer is unlikely to have exaggerated the distance which the fix showed the vessel off-track. The greater the distance, the more likely it would be that the chief officer might be criticised for failing either to alter course or to recommend an alteration to the master before 20 00 when his watch ended. He therefore had good reason to minimize the distance, and his original answers were to the effect that it was not very great. His evidence that it was five-six miles is not, on this view, inconsistent with those earlier statements, and it seems more likely to us that the figure which he gave is an under- rather than an over-estimate. This, effectively, is the result shown by Captain Cockcroft's revised drawing DSl if further allowance is made for the course being changed at 20 20 rather than 20 30. The drawing also shows a 19 14 rather than a 19 35 position, but if the reduction of speed was less than the 5.5 knots, or shorter than the 30 minutes which Captain Cockcroft assumed, then the calculated position at 19 35 becomes closer to the DR position which can be derived from the chief officer's evidence. On the other hand, if the grounding position was closer to the reported position than Captain Cockcroft has allowed, then the distance off-track is increased. The result, in our judgment, is within the margins of error which must be recognized for this exercise. We conclude that at 19 35 the vessel was close to, and to the east rather than the west, of the DR position derived from the chief officer's evidence (Lat 08 deg 02 min Long 14 deg 08 min W). Because this matter was dealt with at the trial without regard to the correlation between the chief officer's evidence and the master's 22 00 DR position which he entered in the log (see above), we have not taken account of that additional factor in reaching our conclusion stated above; but it is apparent, if sheer coincidence is excluded, that the conclusion can be supported on this ground. (d) After the close of argument on this issue, the underwriters submitted further plots based on Charts Nos 1147 and 601 respectively showing Captain Third's reconstruction of the vessel's track back to a 19 14 satnav fix position. These were on the assumption that the alteration of course was at 20 20 and that the currents and tidal streams would have had a negligible effect upon the courses made good, though perhaps slightly accelerating the vessel's speed. They also introduced a new assumption that the vessel grounded as late as 23 20. (We are inclined to the view that the grounding was some little time after 23 00, but not as late as 23 20, because the evidence as to radio messages upon which it is based shows, in our judgment, that the first distress call was sent out, not to Ljubljana, some minutes before 23 20, rather than later.) The reconstructed 19 14/19 35 positions are substantially to the east of a 137 deg courseline from the 04 00 position, by about 11-12 miles, although they show an overall average speed from 04 00 which is little different from 16 knots, thereby supporting the chief officer's evidence in that respect. The owners responded on the last day of the hearing by producing two overlays for the plot on Chart No 601, illustrating their contentions on the assumption that the alteration of course was 20 20. The primary case is on the same assumptions as before; because these included a reduction in speed to as little as 12 knots for 30 (or perhaps 20) minutes, it is not surprising that their reconstructed 19 14/19 35 position is less far north of the grounding position than Captain Third's. They do show, however, a 19 35 position only 6 1/2 miles to the east of the 137 deg courseline from 0400 and to this extent they, too, are supported by the chief officer's evidence. Their alternative case (overlay DS3-2) calculates what might have been the master's DR position if he derived this from an (assumed) noon position which Captain Cockcroft has derived by plotting back from his own reconstruction of the 19 14/19 35 fix. In our judgment, these further plotting exercises all suffer from the same defects as those which went before. Rather than work from the direct evidence, including the chief officer's recollection of where the 19 35 satfix was, they seek to demonstrate on the basis of various assumptions that their alternative explanations of the grounding are correct, namely, the owners' case that the master believed that the vessel was in some other position than where she was, and the underwriters' contention that she must have been in a position where the course of 115 deg was bound to lead the vessel aground. In our judgment, the effect of these various revisions is to confirm our view that this issue should be decided by reference primarily to the direct evidence, supported and explained as necessary but not superseded by expert evidence, and in any event by reference to the whole of the evidence, including that of motive or possible motive for the grounding. (e) Course and speed made good -- 0400-19 35 Much attention has been directed to the reasons, if the 19 35 position given by the chief officer is approximately correct, as we have found it to be, why the course made good overall during the day was 136.5 deg (T) rather than 137 deg (T) particularly when the fixed (delta) gyro error was 0.5 deg or more in the opposite direction, and why the speed made good was only 16 knots. The shipowners say that these differences indicate an adverse current, although the statistical chances of such a current based on published observations are minimal, being less than 23 per cent. This same unpredicted current, they say, explains why the vessel went aground to the north and east of where the master thought that he was. The underwriters, on the other hand, say that the explanation must lie in a surreptitious reduction in speed and possibly an unnoticed alteration in course during the master's forenoon (8 00-12 00) watch; the other watchkeeping officers denied that any such reduction or alteration took place during the rest of the day. There is no evidence to support this suggestion, other than the ship's failure to achieve 17.5 knots or more as might have been expected if the current was favourable or even slight. The master and chief engineer both said that the speed was increased, rather than reduced, from about 13 00, because the master became anxious about the Abidjan eta. We do not consider, however, that these speculations can be relied upon as evidence that the grounding was deliberate. The vessel altered course when and where she did, and if that was done with the intention of putting her aground, then that fact would support an inference that a previous unexplained reduction in speed was deliberate also and part of a grounding plot. But in the absence of any evidence that the reduction if there was one was deliberate, the fact of the reduction cannot be relied upon, in our judgment, in support of the underwriters' case. Insofar as these matters are relevant, we note that the chief engineer thought that the speed was increased from 98 rpm to 102/103 rpm at about 13 00. This suggests that at least during the forenoon the vessel was steaming at something less than full speed. For this reason among others we doubt whether the whole of the shortfall can be accounted for by adverse currents and tidal streams. (1) Where did the master think that he was? (a) The shipowners' primary case is that the master was made aware of the 19 35 satnav fix and relied upon DR navigation thereafter, derived from his own chartwork rather than from the satnav receiver. If he estimated the vessel's speed at 18 knots on its course of 137 deg (T), then at 20 30, and even at 20 20, it was in a position where a course of 115 deg (G) would take the vessel south of Cape St Ann and just clear of the Shoals. It is to be noted that this explanation requires, first, that the master ignored the chief officer's estimate that the satnav receiver should be reprogrammed for 16 knots, which the chief officer states that he did, and secondly that the master was unaware of, or made no allowance for, a reduction in speed due to the overheating stern bearing during the first hour of his watch. (If the alteration was at 20 20, the effect of any reduction would be greatly reduced. More generally, the greater the reduction in speed, the more difficult it becomes to accept that the master was unaware of it and made no allowance for it. If he did make no allowance, then it is likely that the reduction was not significant, and any difference between his DR position and the vessel's actual position would be correspondingly small.) It follows from our finding that the 19 35 satnav fix was somewhat to the west of the position shown in Captain Cockcroft's drawing DSl, while to the east of the original 137 deg courseline, that at 20 20 (or later) the vessel was in a position where the course of 115 deg which was then steered, although dangerously close to the Shoals, would not run her aground provided that current and tidal streams had no effect upon her. Conversely, if account was taken of current and tidal streams which affected her in fact, then it was an appropriate course to steer in order to bring about that result. It is implicit in the shipowners' case, as advanced by Captain Cockcroft, that the master made no allowance whatever for the effects of current and tidal streams. The master was not asked whether he failed to do so, and if so, why he ignored the elementary rules of prudent navigation in this respect. It is difficult, however, to see what explanation he could have given. A paradox begins to emerge at this stage; without the assistance of accurate satnav fixes, a prudent navigator would not normally close within 15-20 miles of the coast (this was established by the expert evidence); with satnav assistance, the prudent navigator might properly decide to do so, but he would not then ignore the satnav readings, as the master said that he did, and make no allowance for current and tidal streams, when the need for some allowance was obvious from the configuration of the coast. For these reasons, in our judgment the question is not only where was the vessel when she altered course, to which so much time and effort was devoted, but why did the master make no allowance for current and tidal streams, if those were indeed the forces which drove the vessel ashore? Before returning to this question, we should refer to one other matter which is also relevant. (b) Why steer the course of 115 deg? The master's evidence was that he had planned, earlier in the day, to alter course to 118 deg in order to clear Cap Palmas by 10 miles. This he derived from the next Chart No 3139 on the same small scale as Chart No 1147. This implies an alteration of course position on the original courseline (137 deg (T) drawn frgm 0400) which would not be reached until after 22 00 even if the average speed made good was 18 knots. If the change of course was made earlier, as it was, then the new course would be greater rather than less than 115 deg and if the average speed made good was less than 18 knots, as was also the fact, then this was a further reason why the alteration from 137 deg would have to be smaller than had been planned. The master sought unconvincingly to explain why he chose the course of 115 deg. He said that his intention was to cut into the shore in order to gain the advantage of an easterly current; but this would not increase, and might reduce, any advantage which an easterly current would give. It would also imply that he ignored the northerly element of the current which, the shipowners assert, had been experienced during the day up to 19 35. He said that he intended to "cut the corner" by steering for a time at 115 deg but as is apparent from these figures this would imply a course above rather than below 118 deg. Finally, Mr Steel submitted that there is no significant difference between 118 deg and 115 deg and that the substantial question is whether the alteration was made with the intention of running the ship aground, which 115 deg on Captain Cockcroft's evidence did not. The underwriters responded by producing a further drawing (GP1) based on Chart No 3139 which shows that both 115 deg and 118 deg led the vessel ashore before Cap Palmas was reached even if she was as far south as the 2200 DR position when the alteration was made, which she cannot have been. Finally, the master suggested that he altered course to 115 deg because he remembered that the next planned course was 118 deg (the difference being unexplained) and that he may have done so without consulting Chart No 3139 again. At this point, in our judgment, the master's evidence becomes so unconvincing that the only possible conclusion is that the choice of 115 deg has not been explained. In short, the master's evidence is that at a position which he cannot identify he altered to a course which he cannot explain or justify. (c) Why make no allowance for current and tidal streams? For the reasons given above, this elementary question lies at the heart of the issue whether the vessel was grounded deliberately, or not. Since it was not explored at the hearing, or in argument before us, we do not feel entitled to consider the question in isolation or to answer it adversely to the owners without taking into account all other matters which were canvassed in the evidence and are relevant to the issue. Our difficulty in this respect is compounded by the fact that the underwriters through Mr Pollock have stated that they do not seek to raise any positive case, notwithstanding Captain Third's evidence, as to the vessel's position at the time when she altered course. We confine ourselves, therefore, at this stage to commenting that we find it difficult if not impossible to conceive any reason why the master, unless he behaved irrationally and wholly out of character during his watch, but not before it or after the grounding and without making any other person aware that the change had taken place, should have navigated the vessel as he did. If the grounding was accidental, then it was the result of an extraordinary series of lapses, beginning with an unexplained alteration to an unsafe course; if it was deliberate, then it was a correct choice of course in the conditions then prevailing to achieve the desired result. (d) The shipowners' alternative case This is based essentially on the master's evidence that he believed that the vessel was far south of where she was, in a position close to the 22 00 DR. This case is inconsistent with the master having had knowledge of the 19 14/19 35 fix because on any possible reconstruction of events the vessel must have been far north of the 22 00 DR position when she altered course, and that would have been shown by any DR calculations based upon the fix. The Judge's findings were that the master did remain orant of the fix, whether due to his own careign lessness or to a misunderstanding or failure of communication between the chief officer and himself. But strictly these findings were inconsistent with the further finding that the appropriate reconstruction was set out in Captain Cockcroft's drawing AC1, which was based on the 19 14/19 35 fix and assumed that the master derived his DR calculations from it. This inconsistency was acknowledged by Mr Steel and it means that we have had to make the relevant findings for ourselves. Mr Pollock mounted a strong challenge to the findings that the master was unaware of the satnav fix and that there had been a misunderstanuing between the chief officer and the master, and Mr Steel sought rather to uphold the Judge's approval of drawing AC1 and therefore, as he submitted, his implied acceptance of the owners' primary case. We therefore find it unnecessary to say more than that we regard Mr Pollock's criticisms of this part of the judgment as unanswerable. Moreover, the primary case requires acceptance only of the submission that the master, who correctly calculated his DR position, negligently failed to allow for the effects of current and tidal streams, both in estimating his position and when deciding what course to steer. The alternative case means that the master was unaware of, or paid no attention to, any recent satnav fixes and was content to alter course on the basis of a DR position derived apparently from chartwork which he had carried out not later than noon and probably much earlier, at 04 00. To this must be added the fact that his choice of 115 deg as the new course to steer was inconsistent with those earlier plans and even now is unexplained (see (b) above). For these reasons, we have no hesitation in rejecting the owners' alternative case. (J) Autopilot and echo-sounder We have referred in par (E) above to the owners alternative case that the grounding was due to some malfunction or misuse of the autopilot. Three possibilities were put forward. Each would have the result that the course steered after 20 20 was to the north of 115 deg which was intended, with the inevitable result that the vessel went aground where she did. The first suggestion was that although the pointer by which the helmsman or officer of the watch sets the course on the autopilot was set correctly at 115 deg the equipment itself failed to work properly. The instrument is described in the judgment at p 81. It is of a kind which is in universal use and the particular model was well established and of reputable manufacture and design. There is no reason to suppose that it may have malfunctioned on this occasion, except for the suggestion which the owners put forward, namely, that as a result of a sandstorm which the vessel is said to have experienced on the previous day some particles of sand found their way into the gyro compass repeater on the bridge and prevented it from operating freely. It is possible that this was intended by the master as his original and preferred explanation for the navigational error which he made; his manuscript additions to the fair log after the casualty included a slightly mysterious entry for 15 00: "Washed away the existing sand with fresh water" and for 20 00: "Likewise". The reason why the master was doing this personally is said to have been the fact that the day was Good Friday and therefore a holiday. But however that may be, the fact is that no other officer of the watch or helmsman noticed anything amiss during the whole of the day. There are two reasons for rejecting this explanation outright. First, it is quite impossible to explain how sand could have got into the compass repeater, which is covered, even if there was a sandstorm; and secondly, it is inconceivable that a malfunction on the scale alleged would not have been noticed by comparison with the magnetic compass or by reference to satnav fixes during the day. The second and third possibilities suggested both involve some misuse of the autopilot itself. It is unnecessary to go into great detail. Assuming that it was intended to alter course from 138 deg (G) to 115 deg (G), the helmsman could well decide to disengage the autopilot, steer manually to the new course and then re-engage the autopilot after resetting the pointer to the new course. One method of doing this could mean moving the pointer well away from the relevant part of the dial so that it would not interfere with the helmsman's sight of the new heading. When the ship was steady on that heading the pointer would be brought back to it. Alternatively, the pointer might be moved direct to 115 deg while the autopilot was disengaged. The suggestions are that if either of these procedures was adopted the pointer might have been moved or returned inadvertently to (say) 095 deg. This particular course was identified for two reasons. The dial shows one-degree gradations with figures only at the 10-degree markers. The intervening five-degree markers are more pronounced but they are not numbered. Therefore, it is suggested, the helmsman could move the pointer to the wrong five-degree marker, failing to notice that the 10-degree figures on either side of it were not the ones he intended. Secondly, an alarm bell sounds if the autopilot is engaged when the pointer is more than 25 deg off course. No alarm was heard, therefore that was the maximum error which could have occurred in this case. If this was a realistic possibility, even for a casual and negligent operator, then the number of shipping casualties might be expected to be much greater than it is. We are satisfied that nothing of the sort occurred. The alteration of course took place at about 20 20. The grounding occurred between 2 1/2 and three hours later. In the normal course of events the gyro compass heading would be checked against the magnetic compass and there would be regular checks of the gyro compass, at least, throughout the watch. However negligent the master was, he cannot have failed to check the gyro course from time to time, and there are no grounds for supposing that the cadet was negligent also. More specifically, the error would have been discovered at the time of the 21 41 satnav fix (see below). Much time and effort was spent in trying to support these suggestions by expert evidence, including detailed information obtained from the manufacturers of the autopilot in Germany. Predictably, the evidence that was given all tended to show not merely that these suggestions were speculative, but that they can be wholly discounted in the present case. Mention should also be made of the echosounder, which the underwriters alleged should have been used by the master during his watch and would have alerted him to the fact that the vessel was in shallow waters. His failure to do so is significant in relation to the two impacts before grounding (see further below), but as regards the period before and the reasons for the casualty we do not consider that this non-use of the echo-sounder, assuming that the master did not use it, supports the allegations made against him. His evidence, although not verified in detail, was that he believed that the vessel was in deep water some 15 miles off the coast. Although this is above the continental shelf, and so the depth could have been measured, any echo-sounder readings would not have been a reliable method of safe navigation, in comparison particularly with the satnav receiver. If the master thought that the vessel was in the DR position shown in Captain Cockcroft's reconstruction, then it does become difficult to explain why he ignored the echo-sounder, if he did so; but not more so than his failure to programme the satnav correctly, if that is what he did, or to ignore the satnav altogether both generally and in relation to the 21 41 satnav fix. (K) The basis of the Judge's finding that the grounding was accidental In reaching his conclusion that the grounding was accidental, the Judge made a number of important findings of fact: (1) That the master was unaware of the 19 35 satfix position, and mistakenly thought that he was further south and west of his actual position. (2) That he was unaware of the 21 41 satfix position. (3) That the master reasonably concluded that the cause of the impact was that the vessel had struck logs and not the ground. (4) That there had been a significant slow-down of speed at the start of the 20 00 watch, but the master was unaware of it. (5) That the master's decision to alter course to 115 deg, though mistaken, was due to a genuine belief that it would save time. (6) That the master was an honest and truthful witness. Mr Pollock submits that the Judge was in error in each of these matters. We must therefore consider them. (L) The 19 35 satfix The chief officer's evidence was that towards the end of his watch he had received a good fix at about 19 30 and he marked it on the chart. This showed that the vessel was some five or six miles to the east of her course of 137 deg from the 04 00 position. He calculated the speed of the vessel on the basis of the distance travelled from a previous fix and that at the hand-over of his watch to the master just before 20 00 he pointed out the position to him and told him of the calculated speed of 16 knots. This had been his evidence consistently from the beginning, save for two matters. At the joint interview he said the vessel was a little to the east of the track; he also said that he reprogrammed the speed on the satnav, a point he could not recall when he gave evidence. Apart from the fact that the master denied that the chief officer had told him that the speed was 16 knots and that he had reprogrammed the satnav, the master did not dispute this evidence. He said he could not remember. He accepted that what the chief officer said he had done was consistent with the usual procedure. It was the owners' primary case before the Judge and in this Court that the master was aware of the 19 35 satfix; but was not told of the speed of 16 knots or that the satfix had been reprogrammed to this speed. Mr Steel, however, submits that even if the master was aware of the ship's position as a result of this fix, error crept in thereafter. It has, of course, always been the underwriters' case that he knew of this position and was either told of the speed or, if not, would certainly have calculated it himself. But the Judge, in a crucial part of his judgment, did not accept that the master was aware of the 19 35 position. He accepted an alternative and inconsistent submission made by the plaintiffs in these terms (pp 90-91): The plaintiffs submit that one possible explanation for the casualty is that the information obtained by the chief officer as to position and speed was not relayed to the master, who proceeded on the basis that the vessel was on her course line and proceeding at 18 knots. The plaintiffs submit that a second possible explanation is as follows. There was no satfix update of the DR until 19 35. For a considerable period of time the chief officer may have projected forward the original DR track at 18 knots and marked the 20 00 hand over position on the chart. When the satfix was obtained at about 19 35, the chief officer may or may not have marked the fix on the chart, although it must have been some 15 miles or so further back up the track and to the east. It is possible that he omitted to erase the 20 00 hand over position based on the previous satfix. When the master came on watch, he may or may not have been told there was a recent satfix. However, the master would have looked at the 20 00 DR estimate and, knowing perhaps that there had been a recent fix, he may have felt quite confident about altering course at a position which would have been down the track and to the west of the vessel's actual position. I find that there probably was a failure of communication between the master and the chief officer at the hand over in one of the ways outlined above. At pp 94-95: It is probable that there was a line on the chart from the 04 00 position of 137 deg true with a projected 20 00 hand over position marked. I accept the plaintiffs' submission that the most likely explanation for the grounding is that as a result of a failure of communication between the master and the chief officer at the hand over (see under the heading 16 00-20 00 above) the master mistakenly thought that the vessel was making a course of about 137 deg true and a speed over the ground of about 18 knots. He failed to appreciate that the current was not a following current and no allowance for the reduction in speed due to the hot bearing was made. He relied on DR from a previous position and not on the satnav. The master was concerned about making Abidjan on time. He had originally intended to alter course to 118 deg but in fact altered course to 115 deg because he believed (in fact mistakenly) that he would gain time by altering to 115 deg (and subsequently reverting to 118 deg). He altered course to 115 deg wrongly thinking that the vessel was to the south-west of her actual position. He also failed to give consideration to the possibility of the tide setting the vessel to the north-east. The grounding probably occurred broadly in the manner illustrated on plot AC1 (or AC2 and AC3). Mr Pollock submits that this conclusion cannot be sustained, not least because it is contrary to the uncontested evidence of the chief officer and all the probabilities. Mr Steel concedes that both passages cannot be right and that the Judge is confused in his reasoning. The Judge's conclusion is only explicable on the basis that he had determined that the master was a truthful witness, and consequently was to be believed when he said in his evidence that he thought he had altered course when he was 15 miles off the Shoals and that when the ship encountered the two bumps, he believed he was in deep water and nowhere near the Shoals. Such evidence may be reconcilable with the Judge's conclusion. It is not reconcilable with the fact that the master knew the vessel's position at 19 35, nor is it consistent with the position in which it was suggested by Captain Cockcroft the master believed himself to be, as shown in the plot AC1, as now updated, to show an altered course at 20 30 as opposed to 21 00 (DS1). In our judgment, the only possible finding on the evidence is that the master was aware of the 19 35 satfix and that this showed the vessel five or six miles to the east of her track (as the Judge found) and that this was pointed out to him. We also consider that it is probable that the chief officer told him that he had calculated the speed as 16 knots and had reprogrammed the satnav with this speed. But it is not really necessary to make this finding because there can be little doubt that if the master was not told of the calculated speed, he would have calculated it for himself. In our judgment, this is a fundamental error in the Judge's reasoning and has important implications on the master's credibility, as we shall show. (M) Satellite passes 21 and 22, the satfix at 21 hrs. The questions that arise under this head are two. Did the satnav on the vessel receive a good fix from pass 21 (or 22)? If so, was the master aware of it? This is an important issue and was so regarded by the Judge. Captain Cockcroft accepted that the master's conduct is difficult to explain in terms of an accidental grounding if this satfix was known to him; and Mr Steel concedes that, if he received it, it would have been apparent to him that he was unduly close to the southern edge of the Shoals. The Judge found that the satnav did not receive a good fix from either of these passes; or if, contrary to this conclusion, a good fix was obtained, the master did not perceive it. He reached this conclusion on the following basis. He said that the master's evidence was that he did not receive it. He discounted for various reasons Mr Blanchard's evidence that he was "quite certain" that one of these two passes had given a good fix, more probably pass 21 rather than 22, and he rejected the cadet's evidence to the effect that the master had received a good fix during his watch on the basis that he had said inconsistent things. In the light of this analysis he was not prepared to reject the master's evidence. Mr Pollock submits that the Judge was wrong to reach this conclusion. First, he points out that the master did not say that he never received this fix. On the contrary, he said that he could not remember whether he had or not. At the Lowe interview he said he could not be sure whether he had or not. In his evidence in chief, when asked if he had received any satellite fix after 8 pm, he said: "This is something I have no recollection of." In cross-examination he said: I do not remember that I did take a fix. I may have taken another fix -- obtained another fix, however the fact is that I don't remember doing so. It is remarkable, not to say astonishing, that the master should not remember whether he got a fix at 21 41, since this was so close to the grounding, and he made it plain that he was both expecting one and anxious to know his position from it. Mr Steel submits that a deliberate scuttler would say that he never received it. But that could be dangerous if the master knew that his cadet had witnessed him obtaining it, as the 'latter said. It is now necessary to consider the expert evidence as to whether or not passes 21 (or 22) would have been received by the ship. Of the two experts who gave evidence on this point, the Judge said: Mr Blanchard's (for the underwriters) qualifications in this field were particularly impressive. With no disrespect to Mr Beatty (for the owners) it is clear that Mr Blanchard was more experienced and knowledgeable on the technical side of satellite navigation. There were two possible problems with passes 21 and 22 which might result in a no-fix. First, because the two satellites rose at about the same time it might be that the machine did not get a good fix on either. Mr Blanchard had considered this point in his report. He said: The NAG data shows both passes occurring at the same time and on that evidence they are so similar that it might be thought that no satfix could have resulted. However, re-computation using the correct position for the ship (Table 4) shows that satellite 2l rose 1 minute 24 seconds ahead of satellite 22 and a close examination of the Doppler-corrected frequencies shows that the two frequencies only crossed over late in the pass and for almost all the important part of the pass were separated by more than the band width of the receiver (l00Hz). Fig 2 is a plot of the figures given in Table 4, showing that they are sufficiently clear of each other for the D85 to have tracked one or the other quite successfully over the important part of the pass -- the middle. By 21 24 they are clear enough to permit correct tracking of 1, and do not approach within a 100Hz until 21 34. This would have provided 10 minutes of clear tracking -- quite enough for the DS5 to have produced a good fix. It can do so after only 6 mins of data. It is quite likely that one of these passes was tracked successfully resulting in a good fix and one is therefore listed. Mr Beatty accepted Mr Blanchard's explanations in relation to the Doppler-corrected frequencies; but he suggested that there might be reasons why the machine might lose lock on satellite pass 21, then, in seeking to track it again, pick up satellite pass 22, with the result that no fix would be obtained. He advanced a number of reasons for this: radio noise, incorrectly balanced antenna height, risk of computing on the wrong side of the orbital subtract, ionospheric refraction, mishandling of the high/low pass facility, satellite uploaded with faulty data and satellite not working for any reason, interference by scintillation. The Judge rejected all these suggestions: (Judgment p 89). In the end there appeared to be only two possible explanations as to why the machine might lose lock on satellite pass 21 and fail to pick it up again for the requisite six minutes to obtain a good fix, but become confused by pass 22. This condition is known as multipath and might arise in two situations. First, interference with reception of the signal by the radar gear box on the mast. Mr Blanchard explained that this was not relevant because of the relative position of the radar gear box and the satellite receiver and the direction in which the vessel was heading in relation to the bearing of the satellite orbit. Secondly, in certain conditions of the sea it may be possible for the signal to be reflected off the surface of the sea. This was not relevant because it requires sea state one or less. The sea state at the time was three or four. In the course of cross-examination Mr Beatty said this at D26/18G: [Q.] But is it not the case on the basis of your investigations and the recalculated Doppler shift and the basis of your own investigations with the machine in your home, that you have no reason whatever to think that satellite no 21 would not have given rise to a good fix? [A.] Based on all of those things you've outlined, I have no reason to believe that it should not have given a fix. The investigations referred to are certain experiments which Mr Beatty had carried out at his home to see if he could identify instances of two satellite passes occurring at the same time where, after the appropriate Doppler correction, he failed to get a good fix. He did not succeed. By the time Mr Blanchard gave evidence he, too, had carried out experiments over a period of several months with a view to establishing the same point. In particular, during the period of a week there had been 14 occasions when satellites had risen within two minutes of each other. Only in the case of two was a good fix not obtained and the failure was due to the fact that there was insufficient difference in frequencies. The results of these experiments were not available when Mr Beatty was giving evidence and he was not cross-examined about them. Accordingly, arrangements were made for Mr Beatty and Mr Blanchard to meet and discuss the experiments on the clear understanding that if Mr Beatty wished to comment upon them he could be recalled to do so, and no doubt further cross-examination of Mr Blanchard would have been permitted. They did meet. Mr Beatty was not recalled and no further cross-examination of Mr Blanchard took place. As a result partly of these experiments, and perhaps also from further consideration, in his evidence Mr Blanchard said that he was now quite certain that the machine on Ikarian Reefer had tracked one or either of passes 21 or 22 and it was more likely to have tracked pass 21 because it rose first. He accepted, of course, that he could not be a hundred per cent certain because he had not been present on the vessel. Why then did the Judge not accept Mr Blanchard's evidence? We must examine some of the reasons that he gave (pp 88-90): (1) Because Mr Blanchard had changed his opinion from "quite likely" to "quite certain". We cannot see the relevance of this, since Mr Blanchard gave a perfectly good reason for doing so. (2) Mr Beatty "had no or no proper opportunity of dealing" with the evidence relating to Mr Blanchard's experiments. If this had been correct, it might have been a powerful objection. Unfortunately, it is clear that the Judge was in error. As we have indicated, there was a full opportunity and it is plain that Mr Beatty had nothing further to add or contradict Mr Blanchard. (3) At one stage in Mr Beatty's evidence there was a suggestion by him that if a facility called "programme tracking" had been incorporated in Ikarian Reefer, it might overcome some of the problems. Mr Tomlinson intervened to say that Mr Blanchard "had ascertained" that because the serial number of this set was over 31,000 it did have this facility. It is not clear whether Mr Blanchard has in Court at this time. However, when he came to give evidence it transpired that this was based on Mr Blanchard's recollection. The manufacturers no longer had their records and could not confirm it. In his evidence Mr Blanchard explained that this facility was in any event irrelevant to the problem of loss of lock. There is no suggestion that Mr Blanchard was trying to mislead the Court. We cannot see any reason why this incident, which may well have been due to a misunderstanding by Counsel, should detract from Mr Blanchard's evidence; yet the Judge referred to this matter, without deriving any conclusion from it, as one of his reasons for having reservations about Mr Blanchard's evidence. (4) That Mr Blanchard's experiments did not exactly parallel passes 21 and 22. This is true; but Mr Blanchard said there were sufficient similarities for him to reach the conclusion he did. The Judge said that there was no proof that the software of the DS5 on Ikarian Reefer was the same as that in Mr Blanchard's house. Again, the evidence was that there could be no relevant difference. (5) The Judge referred to the anecdotal evidence of Mr Beatty and Mr Burnett that fixes expected are sometimes not obtained. Mr Blanchard said that whenever he had investigated a complaint of this sort he was always able to find an explanation for it. (6) The Judge referred to the fact that the second officer and chief officer were uncertain as to the times or number of fixes during their watches. We do not regard this as a convincing reason for supposing that certain fixes were not obtained among their watches, as to which there was no dispute. Mr Pollock submits that none of the Judge's criticisms are valid. In our judgment, there is much force in this submission. The cadet's evidence was that he observed the master obtaining two fixes during the last watch. He had heard the alarm. The Judge dismissed his evidence as hopelessly inconsistent. There certainly was some inconsistency; but the same could be said for almost every witness. To Mr Lowe he is recorded as saying that he heard the alarm for the satnav in the morning watch; but he could not remember whether there was an alarm in the evening watch. At the joint interview he said the same as he did in evidence, save that he added that the master had put his position on the chart. In evidence he said he had not seen that. It is not clear why he should have said there were two occasions in the last watch, unless he included the fix received after the grounding, which seems unlikely. He said he knew the difference between the sound of the alarm and someone pressing the buttons to obtain information. Although the cadet's evidence therefore is not entirely satisfactory, it is corroborated by the expert evidence. In our judgment, the Judge's finding that a fix for passes 21 or 22 was not obtained is contrary to the evidence. The conclusion must be that there is a high degree of probability that one or other of these passes, probably 21, was received. The Judge held that if it was received, the master did not perceive it. This finding can only be based on the Judge's mistaken belief that the master denied receiving it. In our judgment it is inconceivable that if, as we find, it was received, the master was not aware of it. The Judge found that the alarm was on. The master made it plain that he was anxious to receive a fix. On any basis, this must be correct; if he was innocent he knew he was sailing very close to the Shoals and not giving them anything like the 15 miles distance which he accepted that a prudent navigator would. If he was intent on scuttling the ship, it was equally important for him to know exactly where he was. The master accepted that if the fix had been obtained, he would have been aware of it. He would know exactly when to expect it. In our judgment, the only possible conclusion on the evidence is that the vessel received and the master was aware of the 21 41 fix. (N) The two pre-grounding impacts In the 30 minutes or so before the grounding the vessel sustained two impacts. The best timing for the first comes from the evidence of the deckboy Maherides who was asleep in bed, but very severely shaken. He looked at his watch and the time was 22 50. Some 10 or 15 minutes later there was a second bump which again was about 10 minutes before the final grounding. It is now obvious that these impacts were due to the vessel striking the bottom. At the time of the first impact the cadet was not on the bridge but in the messroom. His evidence was that on feeling the impact he shouted "we touched the bottom". He returned to the bridge and asked the master what happened. He replied: "We'll see" and went to investigate; after a few minutes the master said that he thought it was a log. After this he told the cadet to alter course to starboard, which he did for a few minutes before being told to resume the course of 115 deg. It is of some significance that the master had not himself already altered course to starboard; but only did so when the cadet returned to the bridge. The chief officer was lying on his bunk at the time of the first impact, which he described as a slight bump and vibration coupled with a sudden jerk as if the ship had touched ground. He immediately went to the bridge and asked the master what had happened. The master told him that he thought the ship had hit a heavy log or something like that. He said to the master that it felt as though they had touched bottom, but the master said they were in deep water and it was probably a log or debris. The master then instructed the chief officer to check the bilges. This he did with the bosun and, after checking No 1 and No 2 hold bilges, he called to the bridge that they were alright. He then went to the engineroom to check the bilges for holds No 3 and No 4. It was while he was in the engineroom that there was a second impact. According to the the chief officer it was so severe that he was: . . . very badly shaken, physically shaken by the second bump, I nearly killed myself. He realised they had hit bottom, and returned to the bridge. After the second impact the master instructed the cadet to alter course to starboard, which he did, again for a few minutes. However, by the time the chief officer returned to the bridge, the master was ordering the cadet to resume the course of 115 deg, instructions which the chief officer relayed to the cadet. Shortly after this the ship ran aground. The master's explanation of the second impact was that he thought he had hit another log. He said that he did not refer to the satnav to check his position; he checked his position using DR. He did not use the echo-sounder because, as a result of checking his position, he was satisfied that he was in deep water and consequently the impact must have been with logs. The Judge criticized the master for negligence. He said that he failed to use the echo-sounder, to alter course to starboard for a longer period of time, to slow down or stop, or look at the satnav. But he accepted his explanation that he thought he had hit a log and not the bottom. He reached this conclusion for two reasons. First, he said that the master was not alone in considering that the vessel had struck a log. He cited examples of other members of the crew who thought the same. As Mr Pollock points out, this was only a partial selection and omits, among others, the evidence of the cadet and chief officer, to which we have referred. Mr Steel accepts that the preponderance of view among the crew, until they were told that the master had said it was impact with a log, was that the vessel had struck bottom. The crew would have no reason to question the master's opinion, at any rate after the first impact; though they plainly did after the second. The second reason was that the Judge accepted that the master thought he was in fact in deep water and in a safe position. This conclusion derives from the Judge's earlier findings in relation to the satfixes at 19 35 and 21 41. Once it is established, which it is for the reasons we have already set out, that the master was aware of both these satfixes, it is quite plain that the master cannot possibly have thought he was in deep water. Even if he only received the 19 35 position, he must have known perfectly well that he was sailing dangerously close to the Shoals. The Judge accepted the plaintiffs' submission that the master's conduct in relation to these impacts was not consistent with that of a man intent upon deliberately grounding his ship. We find it very difficult to follow this reasoning. Only if one believes uncritically everything the master says is it possible to accept the plaintiffs' submissions. (O) Did the vessel slow down at the start of the 20 00 watch and, if so, was the master aware of it? The Judge found that there was a significant slow-down at the beginning of the watch because the engineroom crew had to deal with a hot bearing, but the master was unaware of this. Although this is an important part of the owners' primary case, as shown in AC1 and DS1, it is difficult to see how it plays any real part in determining where the master thought he was, having regard to the Judge's conclusion which we have set out above. But it is a matter which we should deal with briefly. There was a good deal of evidence that a hot bearing was discovered by the chief engineer when he came on watch at 20 00; the engineroom crew from the previous watch were either called back to deal with it or never left. The evidence relating to the slow down in speed is much less satisfactory. In the initial interviews, although the problem of the bearing was mentioned, there is no suggestion of loss of speed, and the chief engineer is recorded as saying that there was no reduction. In evidence the chief engineer said that the engine speed was reduced to as slow as 70 rpm for about an hour and then gradually increased over a period of 20-30 minutes to 102-103 rpm. This is what the Judge accepted. It is impossible to reconcile this evidence with any realistic estimate of the 19 35 position and the grounding position at about 23 00; the ship simply could not have covered the distance between them in the period, even making every allowance for error in the two positions. The Judge also found that this slow-down was not observed by the master. If it happened, this finding is contrary to all the evidence, save possibly that of Captain Cockcroft, who had no experience of reefer vessels. The chief engineer said he would have notified the bridge, and it was clearly his duty to do so. It would be a serious dereliction of duty not to have done so if the reduction was of this order and lasted for anything like this period of time. The master said he would have noticed such a slowdown. Captain Third said it was inconceivable that he would fail to do so. Our conclusion is that there was a problem with an overheating bearing which was dealt with without any significant slow-down of the vessel and that any slow-down did not last for more than about 20-30 minutes. It would have made no appreciable difference to the expected and actual position of the vessel at the time of the alteration of course to 115 deg. (P) The choice of course of 115 deg. The Judge accepted that, although he was mistaken, the master genuinely believed that a course of 115 deg would save time. We have already dealt with this point. The master's explanation, such as it was, made very little sense, even if the alter-course position was what the master said was the 22 00 DR position. If it was where it must have been, and where, within a mile or so, the master must have known it was, it can only be explained on the basis that it was intended to lead to the position where the vessel went aground, otherwise it is incomprehensible. (Q) The Judge's assessment of the master At an early stage of his judgment the Judge set out his impression of the witnesses. Of the master, he said this (p 84): I formed the impression (based on seeing and hearing the master and on his demeanour) that he was a truthful witness and that he did not deliberately run Ikarian Reefer aground. He was of course in difficulty in explaining certain matters particularly as, even on the plaintiffs' case he had been grossly negligent. At times he showed understandable frustration at his long ordeal. The master was at all times adamant in his denial of the allegation that he deliberately cast the vessel away. On several occasions his answers were to this effect: I may have made a mistake -- I have accepted that. I have never accepted and I will never accept under a penalty of my life that I did something intentionally. My impression was of a slightly stubborn man who was proud of his sea-going career. Although he was due to retire shortly I do not think that this master would have been prepared to sacrifice his good record and reputation for a financial inducement. The master would have realised that he would probably face ASNA/disciplinary proceedings and/or criminal proceedings. Further as Captain Cockcroft (the plaintiffs' principal navigation expert) pointed out it would have been dangerous to steer the vessel at 16 or 18 knots into the poorly-charted shoals. If and to the extent that the master did not take part in discussions with other members of the crew after the casualty this was a reflection not of dishonesty but of despair. I accept his evidence that the casualty had a profound effect on him. He referred to "psychological state" and to "keeping . . . going with medication". Mr Pollock submits that the Judge has adopted a fundamentally wrong approach to the case, in as much as he had determined from his observation of the master that he was telling the truth and thereafter had to make his findings of fact fit with that assessment, instead of testing the master's account against what is known, and in the light of the probabilities. If, for example, it can be shown that his explanations require the Court to assume that a large number of improbable things occurred or failed to occur, the master's account should be rejected. Further, he submits that it can be demonstrated that the master's evidence must be dishonest in a number of respects. The Judge cannot be criticised because he had set out his conclusions on the master's credibility at this stage of his judgment, provided of course that he did weigh the account in the way that he should have done against the probabilities and in the light of all the evidence in the case. But even on the basis that the Judge's conclusion that the master was not aware of the 19 35 satfix was correct, we have to say that his conclusion that the master was a truthful witness is surprising. In Piraeus, a few days after the grounding, the master made up the fair log book. He made the following entry: 22 00 115o At position 7o 34 N and 13.42 W changed course. This became known as the 22 00 DR, being the point at which the master said he altered course. This is what he told Mr Lowe (B25/1/18). In the joint interview he accepted that this could not be the correct position for altering course (B35/1/22). It is difficult to see how he could honestly swear on oath in the ASNA proceedings on Oct 8, 1985 that he confirmed what was in the log book (B5/5/l). But once it is accepted, as in our view for the reasons we have already set out it must be, that the master knew of the 19 35 satfix, it becomes clear beyond doubt that he cannot have been telling the truth in his evidence when he said that he turned at a point where he was 15 miles off the Shoals and that when the vessel suffered the impact he believed he was in deep water and nowhere near the Shoals. The owners' reconstructions of the actual course taken, as opposed to the course the master thought he was on (his supposed course) are set out in DS1 (for an alter course at 20 30) and DS2 (for an alter course at 21 00). But the supposed track of the vessel, which, as we have already indicated, makes every assumption in favour of the master, is quite inconsistent with the master's own evidence. And it is difficult, if not impossible, to suppose that his answers on these points can have been honest. Mr Steel at one time conceded this; but he later withdrew the concession, submitting that the position entered in the log might have been put in at a time when the master was suffering from the psychological trauma of the casualty. Even if this could explain the original entry in the log, and we do not think it can, it cannot explain his evidence or his sworn testimony to ASNA. Mr Steel also submits that simply because the master is dishonest, it does not mean that he is a scuttler; it is equally consistent with a desire to conceal his gross negligence. In many situations that may well be so. But it does not seem probable here, where the master was anxious to say that he was negligent in order to rebut the allegation that he deliberately cast his ship away. Furthermore, if, as we conclude, the master was aware of the 21 41 satfix, it is impossible to believe his evidence that he thought he was in a position of safety until the vessel went aground. We find ourselves driven to this result and to reject the Judge's assessment of him as an honest witness. This assessment was a fundamental plank in the Judge's reasoning. There are many other problems with the master's evidence which, 10 say the least, show that he was an unsatisfactory and unreliable witness. Indeed Mr Steel points out that the Judge did not describe him as either of these things; merely that he was honest and would not therefore scuttle his ship. It is impossible to read the transcript of his evidence without getting a clear impression of these problems. Thus: (1) On many occasions the witness prevaricated and was evasive. He declined to answer any hypothetical or difficult questions. (2) Time and again he resorted to the device of saying that he does not remember. Some of the things like the satfixes at 19 35 and whether or not he obtained one at 21 41 would be very firmly imprinted on his memory in any event. (3) Although he must have been fully aware after the casualty of the significance of the two impacts, all he said to Mr Lowe was: Some 10 minutes before the grounding I felt some light impact as if we hit a log but nothing to worry about. (B35/1/33) At the joint interview he said that he had noticed nothing "unusual" before the grounding (B25/1/24). There is no reference to these matters in the log or in any of his statements or evidence in the ASNA proceedings. His explanation of these omissions is, to say the least, unconvincing. At one stage he suggested that the solicitors had put words into his mouth, overlooking the fact that they knew nothing of these impacts prior to interviewing him. (4) When confronted with inconsistent statements that he had made in these interviews, he alleged that words had been put into his mouth by the solicitors at a time when he was suffering psychological stress. (5) The alteration of course to 115 deg. For reasons we have already set out, the master's evidence on this point is wholly unconvincing. (6) The master, after the outbreak of the fire, at some risk to himself, went down to his cabin and collected the crew's seamen books and the fair log, which was only made up to 08 00 hours on the 12th, and accordingly contained none of the vital information relating to the navigation of the vessel thereafter. This information was contained in the scrap log and the charts. Although the master returned to the bridge after fetching the documents from his cabin, he did not take with him the scrap log or charts. The master's explanation for this omission was that he simply did not think of doing so. The Judge's comment on this is that his failure to take these vital documents was explicable, given the fire (p 93). This hardly does justice to the underwriters' case, namely that if he thought about and was able to rescue the fair log and seamen books, although this involved going near the fire, it is very surprising that he forgot about these documents. Mr Pollock has referred the Court to a number of cases where a failure such as this is referred to as one of the common features found in scuttling cases, because an innocent master would be anxious at all costs to show what he had done and why. Mr Steel submits, on the other hand, that a grossly negligent master would be equally anxious to hide his tracks. There is some force in this submission. But it is not easy to reconcile with the facts here, since the master's original case at interview and in the ASNA enquiry was that the grounding was due to equipment failure. If he believed this, he was wholly innocent, and presumably all the more anxious to establish exactly what happened, or what he believed had happened, from the chart. Moreover, even if a desire to conceal gross negligence was the explanation for not taking these documents, it follows that his answers about his reason for not doing so are untruthful. For these reasons, we have come to the clear conclusion that the Judge's principal findings of fact on this aspect of the case and his assessment of the master as a truthful witness cannot be accepted. Accordingly, we have had to make our own findings as to events leading to the grounding, which we have done, and reach our own conclusions as to whether the grounding was deliberate or accidental. 3. (A) EVENTS FOLLOWING THE GROUNDING ON APR 12 AND 13 The fire Following the grounding, attempts were made to refloat the vessel. These consisted of the main engine being run forward three times and in reverse two or three times with the rudder being alternatively turned to port and starboard. These efforts lasted about 20 minutes before being abandoned. The chief engineer was in the engineroom conducting these operations. He was only able to get about 60 rpm from the engine although this increased from time to time, but never exceeded 75-80 for very short periods. The Judge accepted the evidence of the ship's witnesses, and particularly that of the master and chief engineer, that this caused very strong vibrations, especially when the engine was run in reverse. The ship was somewhat down by the head with the result that the stern was out of draft and accordingly the propeller was not fully in the water and was racing. There was also a lot of evidence that the vessel was rolling from side to side, at any rate during the earlier period. Estimates vary as to the degree of roll, from 12 deg-14 deg either side according to the master, to 5 deg either side according to Mr Nginos. It was sufficient to make it difficult to walk without holding on to something. The master said that he was afraid the vessel would capsize; accordingly he ordered that the two lifeboats be lowered and brought to the stern of the vessel. However, this order was disobeyed by the crew, who brought the portside lifeboat, which was only launched with some difficulty, round to the starboard side. As the vessel rolled from side to side her motion was arrested, probably as the bilge keel on either side struck the ground. The vessel was banging, jerking and vibrating. All but five or six of the crew, including the master, chief officer, chief engineer, radio officer, oiler Tsakiridis and possibly also the third engineer, who the master ordered to remain on board, took to the lifeboats. They were able to take a good many of their possessions, including a television set, with them. The Judge also found that there was significant pounding. This is a phenomenon when the vessel is lifted bodily from the grounded position by the swell and then bumps back on the bottom. The underwriters dispute this finding and we shall have to return to the point later. At about 00 30 the chief engineer instructed Mr Tsakiridis to go down to the engineroom to top up the DOST by turning on the purifier since several people would be remaining on board and he did not want the lights to go out. He was also instructed by the chief officer while down there to check the bilges. This took him about 15 minutes. As he came out of the engineroom, he would go past the DOST. About 10 to 15 minutes after he returned, at about 01 00, smoke was detected arising from the engineroom. Fire quickly spread to the accommodation. Virtually no attempt was made at fire fighting; the underwriters make no point on this. At about 01 15 those remaining on board abandoned ship. By this time flames were seen in the accommodation. The clock in the compressor room stopped at 01 25. About 10 minutes after the lifeboats got clear of the casualty the lights on the vessel were seen to go out. This was either because the generators had run out of fuel or the wiring was damaged by the fire. The crew were picked up by Ljubljana at about 03 30. On Apr 22 and 23 Mr Cook and Dr Bound, accompanied by Captain Katakos, one of the owners' port captains, carried out an inspection of the vessel. At about 15 00 on Apr 23 the experts discovered that the tap on the diesel line running from the DOST to the generators was 80 per cent open. Shortly after this they discovered that a valve adjacent to the DOST itself was fractured. They left the vessel at 19 55 hrs, having concluded, subject in Mr Cook's case to his obtaining any innocent explanation from the crew as to the open tap, that the fire had been deliberately started in the bilge or saveall below the generators, fuelled by diesel from the open tap. They considered that the fracture of the valve could be accounted for by the effects of the fire. A full description of the engineroom, which was amidships, is to be found on pp 79-80 of the judgment. For present purposes it is sufficient to say that it consisted of three levels connected by steeply sloping metal stairways or ladders. The main engine was a six cylinder two stroke engine directly coupled to the propeller shaft. It extended from the bottom level of the engineroom up to the top level. On the port side at bottom level were the diesel oil purifiers. Diesel oil was pumped from a tank in the double bottom of the ship, through these purifiers, up to the DOST which was at the top level. The middle level platform on the port side was occupied principally by the ship's four diesel engine generators. The generators were surrounded by a walkway consisting of raised gratings and solid floor plates, about a foot above the middle level floor. Below the gratings was a bilge or saveall (which we shall refer to as the saveall), surrounded by a flat bar about 3 in in height, which formed the edge of the saveall and was designed to prevent objects getting into the saveall from the walkway between the main engine and generators. The top level of the engineroom was level with the cylinder heads of the main engine. On the port side at this level, aft and slightly inboard of the position of the generators on the level below, was the DOST, which contained the diesel fuel which fed by gravity through a drop line down to the generator engines. On its way from the DOST the oil passed first through a quick closing stop valve bolted to the inboard side of the DOST, just above the top level floor plates. The valve was fitted with a quick release mechanism capable of being operated in an emergency from outside the engineroom. The pipework of the drop line was connected to the outlet of the valve. It was this valve which was found by the experts to be fractured on its inlet side between the flange connection to the DOST and the main body of the valve. A full description of the valve and its connection to the DOST is to be found in the judgment at p 95. It is also illustrated in a number of photographs taken by the experts. For present purposes it is sufficient to note that it was made of cast iron. From the valve, the drop line ran a very short distance inboard before turning down to pass through the floor plates of the top level platform. From there it ran back under the DOST at a slight downwards incline until it 'levelled out and ran horizontally for about 35 cms before turning at a right angle and forward to a vertical pillar or kingpost located between and slightly inboard of the forward and aft generator sets. The line passed down the outboard side of this pillar to a clamp which held the pipe in place against the pillar. Above the clamp a small diesel pipe led off from the drop line to a boiler (the boiler feed-line). Below this clamp the drop line passed through the gratings surrounding the generator engines and turned outboard towards the port side of the ship. Four branch lines led to the separate engine fuel pumps. From the pumps some of the diesel would be pumped to the generator engine fuel injectors, and the remainder would be returned via return lines to the DOST. Below the clamp and about 2 ft above the gratings, a tap fitted with a ball valve had been connected to the drop line. It was this tap which was found by the experts to be 80 per cent open. Although there was severe burning at lower level, particularly in the area of the waste bin, and there was also very severe burning at the middle level forward of the generators in the vicinity of the lube oil tanks, it was common ground between the experts that there had been one seat of fire, and that was in the generator saveall in the vicinity of the aft inboard generator. The fuel for the fire was the diesel oil from the DOST; the tap was immediately above this area, and oil from the tap would flow into the saveall. (B) The rival contentions as to the cause of the fire. Any explanation as to the cause of the fire seeks to account both for the broken valve and the open tap. The underwriters' case is that the fire was started deliberately by someone, probably Mr Tsakiridis. This would involve turning on the tap until a sufficient pool of diesel had accumulated, lighting waste or rag partly soaked in the diesel to act as a wick and leaving the tap running to fuel the fire. The fire then quickly developed and by means of hot gases passing up the stairwells spread to the accommodation. While asserting that they did not have to give an explanation for the fracture of the valve, and that in any event on the evidence it was not possible to prove conclusively how it fractured, the underwriters advanced three possible explanations for the fracture of the valve. Two of these, however, the heating of the pipe and the heating of the valve, were abandoned in the course of the trial. The third explanation can be referred to as movement of the floor plates. The underwriters maintain that this is the most likely explanation for the fracture, namely differential movement caused by heating or cooling after heating of parts of the structure. In particular it is suggested that movement of the floor plates, through heat expansion or subsequent contraction in related steel structures, exerted pressure on the pipe where it passed through the floor plates, thus fracturing the valve. On this theory the DOST was empty, or virtually empty, at the time the valve fractured. But the underwriters do not necessarily pin their colours to this theory. They say there may be other possible, albeit speculative, explanations. The owners' theory is far more complex, since it has to account for an accidental opening of the tap as well as the fracture of the valve. It runs like this: during the efforts to refloat the vessel, vibrations were set up in the propeller shaft; these were caused by the fact that the upper part of the propeller was out of the water, thereby setting up tilting moments which were transferred to the propeller shaft and thence to the main engine and the structures of the engineroom, including the drop line. These vibrations were at or about the resonant frequency of the pipe of the drop line, with the result that the amplitude of the vibrations was greatly increased. These vibrations started a fatigue crack in the valve. But the valve did not fail at this stage. Thereafter the crack was propagated to failure by one of two mechanisms, or possibly a combination of both, namely vibrations caused by the generators themselves and/or by a shock load caused by the vessel pounding. This fracture must have taken place very shortly after Mr Tsakiridis left the engineroom. As a result of the fracture, oil would spew out of the DOST. Some of this oil would run along the underside of the pipe to the point where it turned through a 90 deg angle. From there it would fall onto one of six hot spots on the exhaust manifold of the aft inboard generator which was hot enough to ignite it. From there the flaming oil would run down the side of the generator and ignite oil in the saveall, this oil either being there in the ordinary course of things or having come from the fractured valve. At this stage there would be fire in the saveall which would spread to the area under the tap. The flames from this fire then heated the tap itself, so damaging the packing of the tap that it no longer afforded any or any significant resistance to turning by means of vibration. This vibration was also provided by the generator engines. It was common ground that this theory involved eight distinct stages or steps, which had to take place in sequence, and some at least within a very tight time schedule. The owners advanced two further alternative explanations. The first is that the fire was started by a carelessly discarded cigarette which fell into the saveall and there ignited waste which in turn ignited oil in the saveall. It was suggested that Mr Tsakiridis, who was a heavy smoker, may have been the culprit; alternatively, that waste had fallen on to a hot spot on the generator exhaust manifold and thence fallen to ignite oil in the saveall. These alternative mechanisms accept that the fracture of the valve was a consequence of the fire; but the mechanism for the opening of the tap is the same as in the fatigue/vibration theory. The Judge concluded that the underwriters had not proved to the relevant standard that Ikarian Reefer was deliberately set on fire. Effectively he reached an open verdict. He said he could not determine how it started; he did not find that it was accidental, though the owners, by their counter-notice, argue that he should have done. We shall have to examine his reasoning in due course; but it is clear that a significant factor was that he had already concluded that the grounding was accidental and not deliberate. He said that it was possible for the tap to vibrate open under the influence of the generators. He said he could not rule out the vibration/fatigue mechanism for the fracture of the valve; if he was wrong on that, he would not rule out the alternative explanations of smoking or burning waste. He considered that there were objections to the underwriters' explanation for the broken valve. He was impressed by Mr Tsakirides, who he did not think was the sort of man who would be prepared to set fire to the vessel; and he was not prepared to reject his evidence on the basis of the expert technical evidence. (C) The underwriters' contentions Mr Pollock criticises the Judge's reasoning and submits that he ought to have been satisfied that the fire was started deliberately. In outline, his submissions are these: (1) The Judge should not have made up his mind that the grounding was accidental before considering the evidence relating to the fire. If the grounding was deliberate, this would cast a flood of light on the explanations for the fire. The only motive for a deliberate grounding would be as the first step in achieving a loss of the vessel. But grounding alone would be little use, since the vessel would be likely, as she did, to float off, with little damage, insufficient to bring about a constructive loss. It is necessary, therefore, to set fire to the grounded vessel, an established method of scuttling a ship and one which in the circumstances involved little risk to the crew. (2) The owners have to give a plausible explanation for the open tap. It they fail to do so, the obvious inference is that it was turned on by human hand for the purpose of setting fire to the ship. The explanation put forward, namely that a partly fire damaged tap vibrated open is not plausible. He submits that the existence of the open tap casts an evidential burden on the owners, which they failed to discharge, and they cannot rely on the disappearance of the ship in support of the proposition that the underwriters have failed to discharge the legal burden of proof. (3) The owners have to put forward a plausible explanation for the fire in the saveall, which is a necessary precondition for the accidental opening of the tap by vibration. (4) Of the three explanations put forward by the owners for the existence of a fire in the saveall, none is plausible. In particular: (i) The fatigue/vibration theory requires a sequence of improbable and implausible events and coincidences such that, although it is theoretically possible, in practice it is so unlikely that it can be discounted. There is no positive evidence to support it and it requires a large number of favourable assumptions to be made, for which there is no evidence and no justification. (ii) The smoking theory is contrary the evidence and is highly implausible. (iii) There is no evidence to support the rag theory, and it is highly implausible. (5) If the underwriters succeed in steps 2, 3 and 4, they are entitled to succeed on this issue and will have established to the required standard that there must have been arson. It is not incumbent upon them to provide an explanation for the fractured valve. This is not a case where there are only two possible explanations for the fracture of the valve, namely the fatigue/vibration theory and the moving of the plates theory. Even assuming that the valve was not fractured by the arsonist or his accomplice, which appears to have been the assumption on which the trial proceeded, there are other possible explanations of which moving of the floor plates is the most likely, or least improbable. The fact that these possibilities cannot be proved because the evidence disappeared with the sinking of the vessel should not rebound to the underwriters' disadvantage. This is particularly so since the owners were told in clear terms by the solicitors of the need for further inspection of the vessel and preservation of the valve, whereas the underwriters were not aware until August, 1989, when certain further and better particulars of the points of claim were served, of the importance attached to the broken valve by the owners. (6) The Judge held that the moving of the floor plates theory was a possible explanation of the fracture of the valve, or if he did not, he should have done so. The underwriters submit that although there are difficulties with this theory, it is more probable than the vibration/fatigue theory. (7) The Judge considered each of the eight steps of the vibration/fatigue theory and considered that they were each possible, and accordingly he could not rule out this explanation. But he made no attempt to assess the degree of improbability of each step. Had he done so, he should have concluded that although each step was theoretically possible, each step was either improbable or highly improbable. Since the theory required each step to occur in sequence and often within tight time constraints, it should have been rejected as incredible. (D) The owners' contentions Mr Steel, on behalf of the owners, on the other hand makes the following submissions: (1) If the grounding was accidental, as the Judge found, it is less likely that the fire was deliberate. (2) This Court should pay due regard to the Judge's assessment of the two potential arsonists, Mr Tsakiridis and the chief engineer. The Judge believed they were honest and of good character. (3) The underwriters do not dispute that the vibration/fatigue theory is possible; the Court therefore should concentrate on the underwriters' explanation for the broken valve. Mr Steel submits that at the end of the day only one explanation should be considered, namely the moving of the plates. The Judge either found that this was impossible, or if he did not, he should have done. This would then bring into play what is known as the Sherlock Holmes dictum when he said to Dr Watson in "The Sign of Four": How often have I said to you, that, when you have eliminated the impossible, whatever remains, however improbable, must be the truth. (4) Even if the moving of the plates theory is possible, it is very improbable. It was introduced late in the day, only when the underwriters' primary case of heating the valve was abandoned. No attempt was made to analyse the mechanism by which the plates might be moved so as to impact on the pipe and fracture the valve until a very late stage of the trial. The evidence showed that the initiation of the fracture was on the aft side of the valve; but any movement was from forward to aft, and that cannot account for the fracture. (5) If the fracture of the valve was a consequence of the fire, the alternative explanations of smoking and burning rags are possible and more probable than arson. (E) The Court's approach Before we consider the detail of the evidence in relation to the various issues, we think it right to set out what we consider should be the proper approach of this Court. (1) The Court is not assisted by considerations of evidential burdens of proof which may arise in the course of the trial. The sole question at the end of the trial is whether the underwriters have discharged the burden of proving that the fire was deliberately started with a view to causing a constructive total loss of the vessel and that this was done with the connivance of the owners. The burden of proof is the balance of probabilities but "commensurate with the gravity of the charge". (See The Filiatra Legacy, [1991] 2 Lloyds Rep 337.) The burden is a heavy one; but the mere existence of the possibility that the fire was accidental does not mean that it has not been discharged. There must be a real or plausible explanation which is supported by the evidence, or at the least is not inconsistent with it. (2) In considering whether the burden of proof has been discharged, the Court must consider all the evidence in the case, including that relating to motive and particularly that in relation to the grounding. If the Judge made up his mind that the grounding was accidental before he considered the fire evidence, he was in error. The two parts of the case are inextricably linked. If the grounding was deliberate, it is difficult to think of any purpose in doing so, save as a preliminary step to the destruction of the vessel; indeed none has been suggested. It is still conceivable that an accidental fire supervened before the conspirators destroyed the ship; but it is in the highest degree unlikely. If the grounding was accidental, then it is less likely that the crew would deliberately set the vessel on fire. It is conceivable, but very improbable, that the owners had given some general instruction: "If you find yourself on a sandbank it would be convenient for us if you set fire to it." We regard the suggestion that a member of the crew who had been ordered to stay on the vessel might have deliberately set fire to the ship on his own initiative, so that it would have to be abandoned, as fanciful. (3) There are two crucial questions on this aspect of the case. The first is whether the owners have advanced a plausible explanation for the open tap. If they have not, the obvious inference is that drawn by the fire experts who found it, namely that it was turned on deliberately to fuel a fire lit by the arsonist. Secondly, since the owners' explanation for the open tap depends on the existence of an accidental fire below the tap, there must be a plausible explanation for the existence of such a fire. (4) We have already set out what the proper approach should be by this Court to the Judge's assessment of the witnesses. Most experienced Judges recognize that it is not easy to tell whether a witness is telling the truth, particularly if the evidence is given through an interpreter. Where, as in this case, there is clearly evidence upon which a charge of arson can properly be advanced by the underwriters, the Judge should be careful to keep an open mind as to the honesty of those impugned. It imposes too high a burden on the underwriters to say that such witnesses must be telling the truth unless the underwriters prove that their accounts are impossible. Their evidence has to be tested in the light of the probabilities and the evidence as a whole. (5) Where the owners' explanation requires a series of steps to happen in sequence, each of which is improbable or highly improbable, the explanations may become incredible, especially if some or all of the steps have to take place within a tight time-scale and involve one or more remarkable coincidences. As Mr Justice Greer pointed out in The Ionna, (1922) 12 Ll L Rep 54: One improbability would not be sufficient to justify one coming to the conclusion that the event did not happen. But when there are two improbabilities the likelihood of it happening is still more remote, and when there are three it is more remote still. (6) Although the Judge's conclusion as to the moving of the floor plates theory is not altogether clear (p 120, par 9) we think that he held that it was a possible explanation. This is reinforced by his further alternative conclusion that the smoking and burning rag theories could not be excluded. These theories accept that the broken valve was a consequence of the fire. (7) We do not accept Mr Steel's submission that there are only two possible explanations for the broken valve, namely the fatigue/vibration theory and the moving of the plates. This is not a case where the Sherlock-Holmes dictum can be applied. There are possibilities which cannot be excluded because the evidence is not available and all the relevant facts are not known. (See The Popi M [1985] 2 Lloyd's Rep 1 per Lord Brandon at p 6) (8) It is not profitable to consider why the evidence is not available. The underwriters have not attempted to prove that the wreck of the vessel was sunk on the instructions of or with the connivance of the owners. It is true that the owners were aware before the vessel sank of the importance of further inspection of the vessel and particularly the valve and its immediate vicinity, while the underwriters were unaware of the importance attached to this till much later. But in our judgment, no adverse inference should be drawn against the owners to fill any gaps in the underwriters' case. But equally, the underwriters do not have the burden of showing that the owners' explanation is impossible. Even in a criminal trial that is putting the burden too high. The Court must make up its mind on the evidence available. (F) The factual issues We turn then to a consideration of the factual issues relating to the fire. We do not intend to analyse the conflicting expert evidence at length since it would make this judgment inordinately long. Much of it is set out in the Judge's judgment. What we shall attempt to do is consider the probability or plausibility of the various explanations. (G) The open tap The explanation for the open tap lies at the heart of the case on the fire. The owners must put forward a plausible explanation of how it might have opened accidentally and in the course of the fire if they are to resist what is otherwise a clear inference that it was opened deliberately for the purpose of fuelling the fire. The theory is that given a fire in the saveall below the tap, the packing of the tap can be sufficiently damaged -- and the packing of the tap on Ikarian Reefer was found to be damaged -- that it is possible that resistance to rotation is totally or almost totally destroyed so that it vibrates open as a result of vibrations set up by the generators until they stopped. We consider the question of a fire in the saveall hereafter. It is plain that there are a considerable number of difficulties in this theory. Thus although it is obviously possible for a tap to rotate open under vibration where the resistance to rotation is virtually non-existent, it does not follow that this is at all likely to happen in practice. The owners' case depends to a large extent on the first Piraeus tests. Yet it is plain that these were very unsatisfactory from a scientific point of view. Indeed, Mr Cook accepted many of the criticisms in cross-examination (see the Judge's judgment 114); he regarded the tests as preliminary only and intended to carry out further and better tests in England; but these proved impracticable. Mr Pollock made the following criticisms: (1) It is difficult to know the true state of the tap taken from Ikarian Reefer. This had been dismantled on board the vessel by Dr Bound and Mr Cook; it then had to be reassembled and then a subjective judgment made, quite possibly after considerable further manipulation, to assess its resistance to rotation. (2) Mr Cook was not present when the tap used in the test rig of the Piraeus tests was subjected to flame. No one gave firsthand evidence as to how this was done. This was a singular omission. It was said that it was subjected to a blow torch for 20 minutes with diesel oil in the pipe line above. Yet a tap which was similarly treated with water in the line had sustained no significant damage. There does not appear to be any explanation of this difference. When the first tap was examined the damage to the packing was not similar to that in the tap taken from Ikarian Reefer. Dr Taylor made several attempts to reproduce the damage to the packing of similar taps, by use of a blow torch. He was unsuccessful. It is a cardinal principle of a properly-conducted scientific experiment that it should be reproducible. The tap in the test rig had been rotated both manually and by vibration a number of times before the initial test. This is likely to have affected any resistance to rotation as compared with the state of a tap which had simply been affected by fire. (3) The handle of the tap on Ikarian Reefer was made of aluminum and had partially melted in the heat of the fire, subsequently solidifying and fixing the position at 80 per cent open. If the aluminum is in a liquid state, Mr Cook thought it would not have much binding effect as it came into contact with the variable geometry of the valve. But if it was in the stage between the solidus and liquidus it might induce some measure of resistance. A substantial quantity of the handle is still adhering to the top of the tap, so it seems unlikely that it ever reached the liquidus state. (4) The tap was tested without any liquid in the line. Dr Taylor was of the opinion that the presence of a head of liquid in the line above the tap would have a double effect. First, it would have a dampening effect on the vibration of the ball and handle of the tap; Mr Cook agreed that it might have some effect on the natural frequency of the pipe. Secondly, that the head of oil in the drop line imposed a force on the ball of the valve which would constitute a significant frictional resistance to rotalion. Mr Cook accepted that this might have an effect. It is unfortunate that this matter was not pursued further in cross-examination, no doubt because at that stage Dr Taylor had not done the necessary calculations or carried out the experiments which he subsequently did. These showed that the force required to rotate an undamaged tap increased from 1 kgf (without liquid in the pipe line) to 9 kgf with a small head of liquid (less than l metre). The Piraeus tests showed that an undamaged tap would not rotate open under vibration. Dr Taylor found that the force required to rotate a damaged tap with a similar head of liquid was I kgf. From this he deduced that if there had been liquid in the pipe line at the test, it would not have vibrated open and that on Ikarian Reefer, where the head was considerably more than one metre -- being about three metres -- it would not have vibrated open. In a long passage in cross-examination Dr Taylor explained the proposition. He concluded with this answer at D70/27F: [Q.] So you are not able to help the court, really at all, are you, if you are not a vibration specialist when what we are considering is what the effects of vibration will be? How can you help the court at all? [A.] By analogy. If the vibration cannot open an undamaged tap on a dry rig and I could open that tap with a certain torque, or force pulling on the tap, measured and recorded, and then we go through the sequence of damaging the tap and testing it again but with a column of diesel oil asserting a pressure on the tap, and if it requires a greater torque after that than it did before then to me that is crystal clear that tap is offering a greater resistance to rotation, subsequently with the diesel oil's help, than it did before. This seems to us to be a formidable argument. There is nothing very esoteric about it. The liquid simply exerts a mechanical force on the ball which inhibits its ability to rotate, whether by vibration or manual turning. Why then did the Judge reject it? First, he referred to the fact that Dr Taylor accepted that no conclusions whatever could be drawn from the tests that he had carried out in November, 1992. Dr Taylor did indeed say this; but it is difficult to suppose that he truly intended to concede that all the evidence he had given as to the effect of a head of oil on the tap was misconceived. This was based on first principles, coupled with his calculations which he considered were supported by his experiments. It may be that in confining his answer to the tests he did not intend to accept more than that the tests did not carry the matter any further. In the course of cross-examination (at D71/5-6) where his answer appears, Dr Taylor refers both to the damping effect of oil on vibration frequency of the pipe and to the mechanical effect of loading on the tap. He made it plain that he was not expert on the first point. It may be that when he again said (at D7 1/16A) that no conclusion could be drawn from the tests as to whether the tap would have vibrated open, he was intending to refer to the first of these matters and not the second. He returned to the point in re-examination and confirmed his calculations; but this matter was not further clarified. Dr Taylor was obviously anxious not to stray outside his expertise; but it does not appear to us that he was doing so, when he gave evidence as to the mechanical forces exerted by a column of liquid. This was a proposition with which Mr Cook agreed, although unfortunately at that time no calculation or experiment had been done to assess the forces concerned. Secondly, the Judge said that he was confident that Dr Ward had in mind the effect of a column of liquid when he prepared his second report. We have been unable to discover on what the Judge's confidence was based. Dr Ward never dealt with the matter expressly. The point was raised by Dr Taylor in his second report when he criticised the Piraeus tests. Dr Ward sought to answer a number of other criticisms, but he never dealt with this point. (5) If the owners' explanation for the open tap is correct, it is necessary to postulate a remarkable coincidence. The Piraeus vibration tests show that in certain circumstances where the tap is free to turn it will either turn from open to close or from close to open, depending on the position of the engine producing the vibrations. There is no suggestion that the tap will move in both directions, first to open and then to close under the influence of the same source of vibration. It will be either one way or the other. The experiments also showed that the tap changed its position very rapidly either to the fully open or fully closed position. For example at 2800 rpm, which presumably is thought to be typical of what is to be expected of the generator engines, the tap moved from fully closed to fully open in 20 seconds. Yet the tap was not found in the fully open position; it was only 80 per cent open. There would appear to be only two possible explanations for this, first, that the generators stopped at the very moment that the tap reached this position. This of course is possible, but highly unlikely even if the time-scale for the opening of the tap is to be measured in a few minutes rather than seconds. Alternatively, Mr Steel suggested that the inhibiting effect of the partly-molten handle may have become effective at that precise moment. This postulates that the generators were still running. The effect of opening the tap would obviously be to fuel any fire that existed beneath it; if the aluminum was already sufficiently molten but still in the intermediate stage between solidus and liquidus, it seems likely that the increased fire would further liquify it and hence present no impediment to further turning. And again it would be a remarkable coincidence that it happened at that precise moment of a very short time sequence. As Mr Cook said, the Piraeus tests show that with a suitably fire damaged tap, such that there was virtually no resistance to rotation, vibration of the kind that might be expected from the generators on the vessel could cause the tap to rotate open or close. They provide no basis for thinking that this was at all likely to happen in practice; on the contrary, in our judgment, it was very unlikely, even assuming a fire in the saveall beneath the tap. (H) Was the fire in the saveall accidental? The owners' primary case in support of an affirmative answer to this question is the vibration/fatigue theory. It is common ground that this requires eight sequential steps. They are set out and considered in the judgment at pp 121-132. But for convenience we set them out again with minor modifications. (1) The refloating attempts must create vibrations at precisely/about the resonant frequency of the pipework attached to the DOST. (2) The resulting vibration must be of an amplitude sufficiently large to impose a stress on the valve which is equivalent to the fatigue strength of the valve relative to the number of stress cycles caused by the refloating attempts. (3) Vibration at precisely/about the resonant frequency must be sustained for a sufficiently long period of time for there to occur sufficient stress cycles, generating sufficient strength in the valve, to cause a fracture to initiate. (4) The crack must propagate to a sufficient extent at the moment when the refloating attempts come to an end. (5) Thereafter the vibration caused by the generators must be sufficient to cause the crack to propagate to failure; alternatively there must be a sufficient shock load caused by the vessel pounding. (6) Oil escaping from the fractured valve must within a short time find its way to the point of ignition. (7) Burning oil must fall from the point of ignition into the saveall or bilge under the aft inboard generator. (8) The burning oil must create a fire in the saveall, fuelled either by oil pouring into the saveall from above, or by oil and other materials already in the saveall. We must consider each of these steps with a view to deciding how probable it is that each step could have occurred and whether the theory as a whole is plausible. Step 1. The refloating attempts must create vibrations at precisely/about the resonant frequency of the pipework attached to the DOST. The problem here is that vibrations set up by the propeller and the resonant frequency of the pipework could only be established within very wide parameters. Clearly it is possible that they would coincide, but there is a correspondingly small chance that they did. In the view of Dr Palmer, they do not have to coincide exactly; but the magnification factor is greater the closer the coincidence. If the propeller was turning at 60 rpm, and no higher speed could be maintained, the frequencies of vibration would have been 4 Hz or less (blade rate) or 8 Hz or less (twice blade rate). Dr Palmer's computer analysis (FEA) showed that the first mode of vibration of the pipework was 5.91 Hz, and the second mode of vibration was 8.59 Hz. But allowing for inevitable inaccuracies and error, the real natural frequencies might be within 2 Hz of the calculated one. This meant that the actual frequency would have been between 3.91 Hz and 7.91 Hz in the first natural mode and 6.59 Hz and 10.59 Hz in the second natural mode. This was on the basis that the pipe was "pinned" rather than "fixed" at the clamp. The owners' experts accepted that it was somewhere between the two; this would have the effect of raising the natural frequency; but no one could say by how much. Everything would depend on the degree of restraint imposed on the pipe and the clamp (assuming that it was the only clamp). Since no one could say what this was, it means that the range of natural frequencies must be even wider. The Judge appears to have overstated the case when he said (at p 123, col 1) and again (at p 126, col 1) that movements set up by the propeller: . . . would excite resonant vibration in the DOST pipe in the second mode. No one went as far as this, which implies likelihood at the least. He correctly stated the position in his conclusions (at p 124, col 2): It is possible that the refloating attempts created vibration at about the resonant frequency of the pipework. He then said that the evidence of Dr Palmer and Dr Ward provided material support for the plaintiffs' case. Their evidence certainly supported the proposition that this was a possibility; but they did not seek to suggest it as likely. As Dr Baker observed, in everyday life resonant conditions are occasional things, few and far between. In our judgment, the 1 degree of probability must be considered as low. Step 2. The resulting vibration must be of an amplitude sufficiently large to impose a stress on the valve which is equivalent to the fatigue strength of the valve relative to the number of stress cycles caused by the refloating attempts. This involves considering two questions; first, what must the amplitude of the resonant vibration have been if the vibration was caused in the manner suggested? This involves a consideration of the amplitude of the forced (ie non-resonant) vibrations and the extent to which the forced vibrations would by magnified under resonant conditions. Secondly, what amplitude of vibration would have been needed to create a stress which would cause a fracture of the valve within the number of stress cycles available? What would the amplitude of forced vibration have been? It was common ground that the transmission of periodic tilting moments from the propeller to the engineroom via the propeller shaft might cause vibration in the engineroom, and the Judge accepted the evidence of some of the crew members that there was such vibration. But it was very difficult to calculate the magnitude of the forced vibrations. Dr Ward accepted that the calculation of the moments in the propeller shaft was very difficult; that there would be some attenuation of the tilting moments by reactions at the stern gland and the shaft supports (of which there were 11 in all); that relative to a ship with an engineroom aft, this vessel had a very long propeller shaft; that an important feature in attenuation was the number of supports for the propeller shaft and this vessel had more than most. He accepted that deflections of the shaft at the engine caused by the tilting moments would be quite small and not noticeable by the human eye; this would need to be measured with a sensitive instrument. Some vessels operate in ballast with their propellers not fully submerged; and indeed there was evidence that in rough weather Ikarian Reefer's propeller was sometimes partly out of the water. It would be surprising if the thrust and tilting moments caused by the refloating operations at 60 rpm (a quarter of what they would be at full speed) should cause such serious vibrations. It was common ground that forced vibrations alone would not be sufficient. There had to be magnification due to resonance. But there was no satisfactory evidence as to what the magnification factor in the pipe work might be. Dr Ward said he did not know and Mr Evans had no experience of the magnification factor in pipe work. Dr Palmer accepted that there would be some hydrodynamic damping due to oil in the pipe, in addition to any clamping. It is necessary for the owners' theory to assume resonance because otherwise the valve would not have fractured. But this does not enable the Court to conclude that it is probable either that there was resonance or that the magnification factor was sufficient to achieve the required amplitude. What magnitude of cyclic stress was likely to cause fracture of the valve within the stress cycles available? The plaintiffs' experts started with the proposition that to fracture the valve in about 20,000 cycles it would be necessary to apply a cyclic stress of about 100 Mpa. Dr Baker described this as a typical middle of the road figure. This means that 50 per cent of valves of this kind could be expected to fracture at or below this stress and 50 per cent would not. However, it became clear that 20,000 cycles were not available. There would be a maximum of 9600 at the first harmonic, assuming that the engine was run at 60 rpm for 20 minutes. But it is quite plain that it could not have been. During the 20 minutes of the refloating efforts, the engine was stopped at least twice, revolutions would have to increase and decrease gradually and no doubt there would be intervals in which communications with the bridge took place. At best it seems that about half this number of cycles could be achieved. If one reduces the number of cycles, the stress needed to cause the failure increases. Because of "scatter" there would obviously be failure in some cases below the average for that type of valve; but the extent of scatter is less in early stages of fatigue. Dr Baker put the lowest figure at 100 Mpa over 5000 cycles or 50 Mpa over 10,000 cycles. Although the effect of the number of cycles is logarithmic and not linear, it is obvious that if the stress and the number of cycles are reduced to these sort of figures, the likelihood of a fracture is much lower. We do not accept Mr Steel's submission that the experts failed to take into account the stress concentration factor at the fillet of the valve in assessing the fatigue strength. How large would the amplitude of vibration have to be to produce stresses of sufficient magnitude to fracture the valve? The owners' case, as advanced by Dr Palmer, is set out in the second table at p 125 of the judgment. Only the last two lines are relevant. At the second mode, ie 10,000 cycles, 159 mm displacement of the pipe causes 100 Mpa at the valve, or 119 mm produce 75 Mpa. It should be borne in mind that these are displacements in one direction only; they have to be doubled because the pipe will vibrate in both directions. It is plain that these are quite unrealistic, since even a cursory examination of the photographs shows that it would be impossible for the pipe to move these distances. Dr Baker considered that vibration of 50 mm was on the very high side and it would be more credible if it was reduced to 30 mm. Mr Corlett said that he had seen pipework components in enginerooms resonating with an amplitude of 8-16 mm under extreme conditions such as crash stop manoeuvres. In our judgment, although the possibility cannot be entirely excluded, the likelihood of vibrations of sufficient amplitude for a sufficient number of cycles is minimal. Step 3. Vibration at precisely/about the resonant frequency must be sustained for a sufficiently long period of time for there to occur sufficient stress cycles, generating sufficient stress in the valve, to cause a fracture to initiate. There are three problems here: (1) As we have already pointed out, the actual number of cycles is substantially less than 9600, and if the material number of revolutions of the propeller to correspond with the resonant frequency of the pipe is 60 per minute, which of course no one knows, but which is probably the most favourable from the owners' point of view, the number of cycles at this speed will be much less still. (2) Resonant vibration in the pipe will only be set up when the engine revolutions are at or about the right speed. If the correct speed is, for example, 55 rpm for the propeller, as opposed to 60, this number of revolutions will only occur very briefly as the engine accelerates or decelerates. (3) There was a good deal of evidence that the weld where the boiler feed line came off the drop line was the weakest point in the pipework. If this were so, then it could be expected to fail in fatigue before the valve; but it did not. (I) Was there a defect in the valve? Appreciating the force of the arguments made on behalf of the underwriters, Mr Steel accepted that it was necessary for the fatigue/vibration theory that there should be a defect in the valve so that failure could occur at stress less than that necessary to fracture a sound valve. The defect that is postulated must be a macroscopic one, and not a microscopic one which is inherent in cast iron because of the graphite flakes in the steel. Dr Baker postulated some inherent porosity in the casting process. Although the Judge refers to a defect in the valve as one of the uncertainties of the case (pp 125, col 2 and 126, col 2) he made no finding that there was a defect. Indeed, it was impossible to do so. Nor did he assess the likelihood of such a defect existing. Dr Deegan, who was the only expert who had any firsthand experience of casting, said that in his opinion it was very unlikely. Defects due to porosity were more likely to occur in the thickest and not the thinnest part of the metal. Cast iron was a good casting metal; that is why it is used. Any surface defect is likely to have been detected by the fettling and the inspection process at the factory. It would, in any event, have to be a substantial defect, which was all the more likely to be detected. There is one further factor which has some bearing on this point. Dr Deegan and Professor Dover were of the opinion that the fracture path was inconsistent with a fatigue fracture. In their view such a fracture was likely to be at right angles to the axis of the valve, and not at about 75 deg, as it appeared to be. Dr Baker did not agree. But Professor Dover's experiments afforded some support for this view. All this suggests that the presence of a significant defect is unlikely. But there is a further factor that weighs against it, and that is the coincidence in the position where the defect must be presumed to be. It was the view of the owners' experts that the probable initiation site of the fracture was on the aft circumference at the fillet on the inlet side of the valve. We shall have to return to this question hereafter; but assuming that this is correct for reasons unconnected with the existence of a defect, it is pure chance that the defect should have been there. There is no greater likelihood of it being there than anywhere else in the valve; indeed Dr Deegan considered it was not the preferred site and was more likely to be in the heavy section. In fact, the valve was more likely to fail in fatigue on the outlet side of the valve, all other things being equal. Without the presence of defect at the initiation site of the fracture, the fatigue/vibration theory founders. There is no evidence that there was such a defect; its existence has to be presumed to support the theory, which is a bootstraps argument. The possibility of such a defect cannot be denied. But, in our judgment, the likelihood of its existence is remote. Step 4. The crack must propagate to a sufficient extent at the moment when the refloating attempts came to an end. Step 5. Thereafter the vibration caused by the generators must be sufficient to cause the crack to propagate to failure; alternatively, there must be a sufficient shock load imposed by the pounding. These two steps can be considered together. These two steps require some remarkable coincidences of time. The initiation of the crack must be caused by resonant vibration set up by the main engine. But it cannot have proceeded to failure when these manoeuvres stopped, since otherwise it would have been observed by the chief engineer and Mr Tsakiridis. Thereafter further vibration, which can only be attributed to the generators, or a shock load due to pounding, has to propagate the crack to complete failure at a very precise time. It must not do it until Mr Tsakiridis leaves the engineroom at about 00 45, since otherwise he must have seen it. But it must do so within a very few minutes thereafter for the remaining steps to occur and the fire to be detected about 10-15 minutes after he left. These features alone require any Judge to consider the fatigue/vibration theory with scepticism. The operation of the Paris law, which is a law of fracture mechanics which applies whether the valve had a defect or not, shows that it is exceedingly unlikely that a crack which had been initiated by the high stress regime caused by the vibrations from the main engine could have propagated to failure in the time available by the low stress regime due to the generators alone, if they were operating normally. In these circumstances the owners' experts, no doubt appreciating the significance of the Paris law, at a late stage in the case introduced the possibility that one or more of the generators was misfiring or that a shock load might be caused by pounding. (J) Faulty generators Because this point arose after the crew had given evidence, they were not asked much about it. The Judge summarizes some of the evidence at p 127, col 2. The chief engineer said that the generators were well maintained. It was Mr Abel's opinion that since the injectors had been cleaned after leaving Hamburg it was unlikely that there would be any recurrence of any problem. It would be apparent if the injectors were becoming progressively blocked, since this is a gradual process and leads to a drop in exhaust gas temperatures which should have been noticed. The proper running of the generators was clearly a matter of importance on the evening of Apr 12 since the refrigerating machinery were needed to cool down the holds. Neither the chief engineer nor Mr Tsakiridis noticed anything wrong with the generators. The Judge gives reasons why they might not do so; but Mr Tsakiridis had been told to go and check the generators and he reported nothing dangerous in the engineroom. There is therefore no evidence of faulty timing; the evidence, such as it is, is to the contrary. It is possible that there was faulty running of one or more generators, but it seems very unlikely. It is not entirely clear whether faulty operation of one or more generators is said to have been sufficient on its own, or whether the vibration had to coincide with the natural frequency of the pipe work to set up resonant vibration. Faulty operation would no doubt increase the forced vibration; but by how much is not known. Mr Steel submitted that it might also produce resonant vibration, and if it did it would magnify the amplitude of the vibrations and increase the chance of propagating the crack to failure in the available time. But once again it would be pure chance and involves another remarkable coincidence, namely that the main engine vibrations coincide with or approximate to the natural frequency of the pipe work in the second mode, while those produced by faulty generators coincide with or approximate to the first mode. (K) Pounding The possibility that pounding might have introduced a single shock load on the valve sufficient to propagate a crack already started by vibration caused by the main engine was introduced by Mr Fyans for the first time on Day 59. It was not foreshadowed in any report. The previous day, Dr Ward had expressed the opinion that the only source of vibration relied upon was the generators. It is unfortunate that it was introduced so late because the evidence of the crew was confused and conflicting on the question of pounding. There was a considerable body of evidence that for a time after the grounding the vessel was rolling from side to side, the roll being arrested as the hull, probably the bilge keel, struck the ground. But the rolling tended to lessen as the tide rose and this is what would be expected as the GM of the vessel moved back into the positive range. It is clear that many of the crew refer to this motion as pounding, but it is not that which Mr Evans had in mind, which is vertical heave of the vessel, caused by the swell lifting her from the grounded position and then dropping her down to the ground. This might be expected to occur as the vessel started to float off. There is very little clear positive evidence that this was occurring to any significant degree and the master said there was no pounding. Much depends on the height of the swell. Although some witnesses estimated this as up to two metres, the preponderance seems to have been that the swell was slight, about 1/2-1 metre. Mr Corlett was of the opinion that because of the shape of the hull there would not be severe pounding; there would be considerable suction effect as the vessel was lifted and corresponding displacement of water which would cushion the effect of the vessel dropping back on the ground. It seems to us doubtful whether the Judge's finding that there was significant pounding was justified. Mr Fyans was not able to offer any quantitive assessment of the forces likely to be involved. He had not considered the heave response of the vessel, nor had he considered in detail how the shock loading might be transmitted to the valve, though he thought that it would be transmitted via the double bottom to the pillar or kingpost to which the pipe was clamped and thence to the horizontal beam which supported the DOST. This theory is at best in our judgment a speculative possibility that cannot be excluded, but is very unlikely, especially as it involves the coincidence of the initial shock load occurring minutes after Mr Tsakiridis left the engineroom. Step 6. Oil escaping from the fractured valve must within a short time find its way to a point of ignition. Step 7. Burning oil must fall from the point of ignition into the saveall under the aft inboard generator. Step 8. The burning oil must create a fire in the saveall, fuelled either by oil pouring into the saveall from above, or by oil and other materials already in the saveall. The only points of ignition put forward are the areas which are uninsulated in the exhaust manifolds. These are six in number and are about 1 in in length and 3 in in diameter. Most of the oil from the fractured valve will gush out away from the DOST; there will be some back splashing as some hits the fracture face of the valve or the valve itself, but it is difficult to see how this back splashing can carry it into the generators. It is more likely simply to drop down through the gaps between the DOST and the floor plates and, without any momentum, is unlikely to bridge the substantial gap to the hot spots, even with the rolling motion of the vessel. The owners' preferred route, therefore, is that oil would have travelled along the underside of the pipe where it went outboard under the DOST and would fall at the point where it turns at right angles. It is necessary that there is sufficient momentum for it then to carry 320 mm outboard and downwards so as to reach the hot spots. It is not altogether easy to see why any substantial quantity of oil should do this, rather than falling off the pipe before it gets to this point, especially at the place where it is no longer inclined downwards, or carrying on round the corner, though when the vessel rolled to port this might increase the chances. There is a good deal to be said for Dr Taylor's view that it would be a "pretty neat shot" if it happened to land on a hot spot. If the oil came in contact with the hot spot, there is no doubt that it would ignite. But it must do so in sufficient quantities to ignite any combustible material already in the saveall. Mr Cook accepted that you would need quite a lot. Dr Taylor said that the oil would have to get down to the saveall without losing its heat, without being quenched, so as to arrive in the saveall still flaming. There was no positive evidence that the oil had escaped from the broken valve before the fire started or that it had followed the proposed route to the point of ignition and thence to the saveall. There might have been such evidence in the following respects: (1) Signs of oil having collected on top of the generators. Mr Cook agreed that you might expect to find signs of damage to the aluminum rocker covers (similar to that which appears in Bound photo no 37), if there had been any substantial quantity of oil reaching the generators. (2) There was no sign of oil, whether burning or not, having run down the side of the generator into the saveall. This might have been masked by soot. (3) There was no sign of any substantial quantity of oil, whether burning or not, having run down the ladder or having collected on the floor plates on the inboard side of the generators. The heat damage to the floor plates was probably due to burning of the table immediately below. (4) There was no sign of flame impingement on the surface of the DOST. Dr Taylor thought that this was the one piece of evidence which was wholly inconsistent with the owners' case. Mr Cook considered that if there was only a thin film of oil in the saveall of the DOST, there would not necessarily be signs on the DOST itself. If the video of the second Piraeus tests is anything to go by, it would appear there would have been a substantial quantity of oil in the DOST saveall, at least in the early stages; although it is possible that the fire had not reached that point in the early stages. (5) There was no sign of oil staining in the area near the valve, for example on the cylinder heads of the main engine or the walkway. This matter was not put to Mr Cook; but Dr Taylor expected there would be some, and certainly oil must have got on to these places in considerable quantities. None of these points are decisive in themselves; but obviously if there is an absence of signs in five places where one might expect to find it, the improbability of it having occurred increases. Apart from any oil travelling along the underside of the pipe, it is difficult to see that any quantity would flow from the broken valve into the saveall. The majority of the oil would drop on to the floor plates inboard of the generators or on to the main engine. There is a substantial lip round the edge of the saveall which would prevent diesel getting from the floor plates to the saveall. Mr Steel accepted that only very small quantities of oil from the fractured valve would get into the saveall. If this is right, as we think it is, it means that the fire in the saveall, which has to spread to the area of the tap and be sufficient to damage the packing, must be fuelled from oil already in the saveall. No doubt there would be some, since this is the purpose of the saveall. But it has a drainage system and it is not easy to see that there would be any substantial quantity there sufficient for the purpose. There is no doubt that both fire experts considered that the fire damage was more consistent with the tap being the sole source of the fuel. Mr Cook considered that so far as the fire evidence was concerned the fracturing of the valve was an unlikely, or very unlikely, mechanism (D4O p 51); though he did not find anything inconsistent with it. Dr Taylor agreed that the tap was the most likely source of the fuel. He regarded absence of burning on the surface of the DOST as inconsistent with the owners' case. The Judge considered whether he should rule out the fatigue/vibration theory in the light of Mr Cook's and Dr Taylor's analysis of the physical evidence. He did not do so because he held that the physical evidence was not inconsistent with the theory. He advanced a number of reasons for this, perhaps the most important of which was that he considered (p 116) that it was difficult to account for the limited nature and extent of the damage to the tap and the generator flat if the fire had been started in the manner alleged by the underwriters. He relied upon Mr Cook's reasoning. Mr Cook considered it was difficult to explain the limited nature of the fire if all the diesel was consumed in the saveall. But it was the view of both experts that there was evidence of spillage of diesel over the lip of the saveall to the lower level. Moreover, Mr Cook himself offered a number of explanations as to the limited nature of the damage; he thought that: The structural members across the generator bilge would provide some resistance to flame . . . that if the mouseholes in the structural members were unobstructed, the presence of the -- . . . structural members could certainly provide an explanation for a collection of a relatively smaller amount of fuel in a localised area . . . gratings and plates over the generator saveall . . . would have restricted back radiation from the flames into the pool and would have reduced the amount of vapour which is being generated . . . He thought it was this which . . . controlled the fire and that it is as a result of that that there has not been such significant damage either where the gratings are or where the plates are. He also accepted that it was probably right that the absence of more severe damage to the tap indicated that it was not closed at the beginning of the fire, and the cooling effect of oil passing through the tap during the fire. This was also Dr Taylor's view. Dr Taylor considered that not all the oil from the tap had burnt in the saveall because there was evidence of spillage. It was put to him in crossexamination that the somewhat limited nature of the damage in the generator saveall and to the tap was neutral, and he agreed. With all respect to the Judge, therefore, it appears that Mr Cook himself proffered an explanation which he regarded as satisfactory for the limited damage, and Mr Clarke did not suggest that this was inconsistent with the tap being the sole source of fire. Among his other reasons, the Judge was very critical of the inspection carried out by Mr Cook and Dr Bound; and he mentioned the fact that Dr Bound had not been called. We shall have to revert to these matters hereafter. He also concluded that he did not consider it safe to draw any conclusions from the time-scale of the spread of the fire. This reason ignores the fact that a diesel fire does not start very quickly because the oil has to vaporize before it will ignite. In our judgment, it is one of the most remarkable features of the fatigue/ vibration theory that the valve must fracture and all the subsequent steps take place within the 10-15 minutes that elapsed between Mr Tsakiridis leaving the engineroom and the fire being detected, by which time it had spread extensively. The Judge considered steps 6, 7 and 8 were all possible. He did not consider how likely it was that any one would have taken place. Had he done so, he could in our judgment only have concluded that it was very improbable. (L) A cigarette or burning rag It was common ground that for a cigarette to have been the cause of an accidental fire, the following steps had to take place: (1) Mr Tsakiridis (who is really the only candidate) must have been smoking a cigarette in the engineroom. (2) He must throw the cigarette through the grilles of the generator platform into the saveaill, either as he went down past the generator flat on his way to the lower level to start the diesel oil purifiers, or on his way back, or when he checked that the generators were operating correctly. (3) The cigarette must land on a piece of discarded rag or cotton waste in the saveall. (4) There must be a pool of oil in the saveall in the vicinity of the tap. (5) The rag or waste must not be completely soaked in oil. The part on which the cigarette lands must be sufficiently dry to smoulder and burst into flames when in contact with the burning cigarette end. (6) Another part of the rag or waste must be soaked in oil and lying in the pool of oil. (7) The cigarette must cause the dry part of the rag to smoulder and burst into flames after Mr Tsakiridis had left the engineroom, or at least at a time when he would not notice it. (8) The burning part of the rag or waste must ignite the oil-soaked part of the rag, which could then draw diesel from the surrounding pool, so as to burn long enough to heat the surrounding oil. After some time the immediately surrounding diesel will burst into flames, the fire will develop progressively through the remainder of the pool, creating a fire in the saveall below the tap. (9) There must be a sufficient fire below the tap to damage the packing and enable the tap to vibrate open. Although Mr Tsakiridis was a heavy smoker, there was no evidence that he was smoking at the time. He denied it and gave a convincing reason why he would not have thrown a cigarette end into the saveall, even if he was. Sand buckets were available in the engineroom for cigarette ends which Mr Tsakiridis was in the habit of using. There does not seem to be any reason to reject his evidence on this point; and unless one does, the cigarette theory is not a starter. Even if one does reject his evidence, it is difficult to disagree with Dr Taylor's assessment that although one cannot exclude the possibility, it is very unlikely, especially when one bears in mind the time-scale within which the steps must take place. The rag theory is a variation of the cigarette theory; it involves the following steps: (1) Rag or cotton waste should have been on the valve covers on the top of the aft inboard generator. (2) The rag or waste must be dislodged, probably either by the rolling or the vibration caused by the refloating efforts, so that it falls on to one of the six hot spots, where it will ignite. (3) The burning material must fall still flaming into the saveall. (4) There must be a substantial pool of oil in the vicinity of the tap. (5) The burning material must fall into the pool of oil and not be quenched by it. (6) The burning material must be large enough to be a persistent flaming source for about five to 10 minutes. (7) The burning material must then draw diesel from the surrounding pool so as to burn long enough. After some time the immediately surrounding diesel will burst into flame, the fire will develop progressively throughout the remainder of the pool, creating a fire in the saveall sufficient to damage the tap. It would appear that a substantial period of time must elapse between steps 2 and 3, since the refloating attempts stopped at least an hour before Mr Tsakiridis left the engineroom and the rolling of the ship had subsided by 00 45. Dr Taylor said that he would expect to have found residues of burnt rag or waste, if this was the explanation. That which was found was clearly not in a position where it could have fallen from a hot spot. The owners did not embrace the cigarette or burning rag theories with any enthusiasm. This is not surprising, since both explanations involve abandoning what is perhaps their strongest argument on this aspect of the case, namely the difficulties that the underwriters have in explaining the fractured valve. The Judge held that he could not rule out either mechanism. That may be so, but with respect to him, it is not the point. The question is: how likely was it? Once again, in agreement with Dr Taylor, it seems to us that the only answer is that it was very unlikely. (M) What other explanations are there for the broken valve, other than the fatigue/vibration theory? At the heart of Mr Steel's submissions on this part of the case was the proposition that at the end of the day there are only two mechanisms put forward to explain the fracture of the valve: the fatigue/vibration theory and the moving of the plates theory. If, as he submits is the case, the latter can be shown to be impossible, then the fatigue/vibration theory must be accepted, however implausible it may be, since it is conceded that it is not impossible. On the facts of this case, we do not accept this proposition for two reasons. First, the evidence is wholly inadequate to be able to say with certainty that these are the only two alternatives. Secondly, the moving of the plates theory, albeit there are considerable difficulties perhaps due to the inadequate material, cannot be said to be impossible. We will consider these reasons in turn. It is one of the striking features of this case that no one appears to have paid any regard to what was described as the croquet hoop. This was a kick guard which protected the valve from accidental damage or interference. It had four supports, two on either side of the valve; two horizontal ones were welded to the surface of the DOST, two vertical ones were welded to the saveall of the DOST. The second engineer's evidence was that it was undamaged before the fire, as one would expect. The photographs show quite clearly that it had been significantly displaced, apparently by a force from aft to forward, such that the two rear supports have become detached, the welds being broken and the front vertical support being bent forward. It is evident that this damage had occurred at some time after the grounding and before soot particles ceased to be available to conceal the original fixing spots of the supports on the face of the weld. None of the experts proffered an explanation of this damage. It cannot simply have been misshapen by heat. It seems very unlikely that it could have been so damaged simply by the valve fracturing and springing away about 1 in in an inboard direction. The direction of movement is wrong and it seems to us that the force of such spring is unlikely to be substantial. In these circumstances, we find it surprising that no one appears to have considered a possible explanation which occurred to all three members of this Court, namely that the valve had been struck by a sledge hammer, axe or other heavy object, either by the arsonist or an accomplice. It is perhaps not without interest that the chief engineer seems to have gone out of his way to give an explanation, albeit somewhat unconvincing as it seems to us, of his possession of an axe at the early stage of the fire. Impact damage of this sort would readily explain the damage to the croquet hoop, which would be struck by the handle or shaft as the head struck the body of the valve. It is also consistent with a fracture path from aft to forward and the limited degree of burning in the saveall and damage to the tap, to which we have referred. Doctors Taylor and Bound do not seem to have given the matter any consideration because, as their first report indicates: "There was no evidence of direct human intervention and it has been concluded that the damage resulted from the effects of the fire at the lower level." It is not clear what evidence they expected to see, especially as the photographs show that the soot deposit had not been removed from the aft side of the valve to see if there are any signs of mechanical damage. Mr Cook was asked whether he and Dr Bound had specifically looked at the valve to see whether it had been hammered to break it, and he said he could not recall. Had the matter been considered, certain questions would obviously arise. Why should an arsonist who had already started a fire in the saveall, leaving the tap turned on, need to fuel the fire from another source? Perhaps because it would spread the fire more quickly. Would a person hitting the valve get drenched with diesel as it broke? Perhaps, but not necessarily so. Is it possible that both the tap was open and the valve broken at an early stage of the fire? We do not think it is possible to say one way or the other, since much depends on how much oil was in the DOST at the outset, the rate of input by purifiers and the rate of outflow from the valve otherwise than into the pipe. Nevertheless, we are quite clear that it would be wrong for this Court to make any finding that this was the probable explanation of the fracture, since it was never considered at the trial, it being assumed throughout that the valve fractured through some non-human force. All that can be said is that it is a theoretical possibility that cannot be entirely excluded. Another possibility is that debris had fallen on to the valve and broken it. That might conceivably account for the damage to the croquet hoop. Less force is required to fracture a valve subjected to great heat; therefore the debris might not have been all that large. There was no sign of any such debris in this area when the experts examined it; but salvors had been on the scene previously. At best it is an unlikely explanation. At the outset of the trial, the underwriters took the view that they did not have to proffer an explanation for the broken valve. It was enough that they could show that the fatigue/vibration theory and the owners' explanation of the open tap were wholly implausible. We have a great deal of sympathy with this view. The only factual evidence upon which any explanation could be based was the photographs of the area. They are far from easy to interpret and frequently do not show the vital points. In spite of this, in the course of the case, and indeed in the course of the trial, the underwriters were drawn into putting forward possible explanations. The Judge was very critical of what he called the underwriters' experts' change of position on this aspect of the case (p 118). We shall have to revert to this matter later in this judgment. It is clear that the underwriters' experts did abandon the heating the valve and heating the pipe theories substantially, if not entirely, before they came to give evidence. But it is equally clear, in our judgment, that they considered that differential movement between various parts of the structure due to heating and expansion or cooling and contraction of metal could cause mechanical impact on the valve sufficient to break it. And, although the mechanism whereby this might have occurred was not considered in detail until the long vacation, and particularly by Mr Corlett, the principle was canvassed from an early stage. (N) Moving of the plates This theory, as it eventually evolved, became known as moving of the plates. If the floor plate through which the pipe passed moved sufficiently as a result of expansion to impact on the pipe the force could be sufficient to fracture the valve. A number of points can be made in favour of the underwriters. First, Dr Baker recognized the possibility of distortion of the structure causing movement of the floor plates which could cause loading to the valve. Dr Palmer was of the opinion that it was impossible to prove or disprove at what stage the valve fractured. Secondly, Dr Taylor had previous experience of valves in a similar position being fractured as a result of fire. He instanced four such valves. The Judge seems to have discounted the evidence (p 111) on the grounds that Dr Taylor could not say whether the valves, which were similar in size and purpose to that on Ikarian Reefer, were of cast iron or steel. He thought they were probably cast iron. But, since it was common ground that steel would be stronger than cast iron, it does not seem to us to be a valid criticism. Thirdly, the valve is situated at a point where two parts of the structure, separately supported, are in close proximity. The DOST is mounted on the beams and pillars which form this part of the engine room structure. The walkway, of which the plates form part, are supported by beams cantilevered out from the main engine. If there is differential movement between the two structures, the pipe is in a vulnerable position. Fourthly, there is evidence that the I beam, which is below and inboard of the DOST, has been distorted by heat. This is clearly visible in photographs G, H and I. It is not clear to what extent, if at all, this beam impinges upon the thwartships' beams which support the DOST; but it may do. A very good deal of expert evidence was devoted to three questions: (1) From which direction did the fracture travel? (2) Was there any evidence that the plates had moved; if so, in which direction? (3) Was there any mechanism which would account for such movement? We do not think it necessary to analyse the evidence on these questions in great detail. But we must refer to the nature of the dispute. Where did the fracture initiate? Dr Baker at one time thought it might have gone from top to bottom. However, he later was of the opinion that it was from aft to forward. This was Dr Palmer's view. It had initially also been Dr Deegan's view; but in the course of giving evidence he considered that he might have been mistaken, and in fact it could have been from forward to aft. Dr Dover seems to have shared the majority view, although he considered that it was very difficult to draw any conclusions from the photographs. Three reasons were put forward for concluding that it ran from aft to forward: (1) A study of the photographs seemed to show that the aft side was at a point of minimum thickness of the neck of the valve, whereas the forward side ran out into a thicker section. But the photographs, particularly photograph D are deceptive. What appears at first sight to be a curvature of the exterior of the valve may well in fact simply be the shadow of the nut which is masking the shape of the valve. It was this which caused Dr Deegan primarily to consider that his first impression might have been wrong. In our view, it was a valid reservation and we can see nothing untoward in Dr Deegan's more cautious approach. (2) The position of maximum stress concentration would be in the fillet adjacent to the flange. This point also to a large extent is dependant on the correct interpretation of the photographs. While it is common ground, other things being equal, that the fillet would be the place of maximum stress concentration and hence the weakest point of the valve, it is not certain that the photographs demonstrate this. Mr Kenny's sketch, which was put in to show a possible alternative view, illustrates the problem. (3) This would be the position of maximum bending moment of the valve in respect of a force exerted on the pipe as it goes through the plates. While, therefore, we agree with the Judge that the probable initiation site is on the aft side, this is by no means certain. It is possible that it was on the forward side. Moreover, it must be borne in mind that all experts considered that a fatigue failure was somewhat more likely to occur on the outlet side of the valve and Dr Deegan and Dr Dover considered the angle of the fracture inconsistent with fatigue. Was there any evidence that the plates, and particularly the plate through which the pipe ran, had moved? It seems to have been common ground that there was evidence of thwartship movement. But this was inboard, and would not impinge on the pipe. On the other hand, if there is any movement of the plate, it suggests that it was no longer secured by its screws or bolts to the supporting steel. Mr Corlett was of the opinion that photograph F1 (which had not been attached to Dr Taylor's original report and which Mr Corlett only discovered during the long vacation) showed that the forward plate, through which the pipe ran, had moved aft and under-ridden the next plate. This can certainly be seen and it seems a possible, even probable, explanation. This seems to have been the Judge's view, since he said: If (as I find) any movement was from forward to aft (p 120). What mechanism could cause the plates to move? Two possible mechanisms were suggested by Mr Corlett. (1) Expansion of the ladder The ladder would have expanded by up to 20 mm as a result of heat. If the ladder was restrained at the top, then either it would distort if it was also restrained at the bottom, or it would move whatever it was secured to at the bottom in an aftwards direction, up to 40 mm. It is impossible to prove or disprove this suggestion, because the photographic evidence is inadequate. It is not clear how the ladder is fixed at the top and whether what it was fixed to would distort. It does appear to be bolted on the inboard side to a lug, which itself may have been secured to the beam; if so, it is understandable that it would not distort the structure at the top. There does not seem to be any bowing of the ladder along its length. It is not clear how the ladder was secured at the base, whether to the plate or the supporting structure. It is clear that there is a lug or upstand to which it is bolted. It seems perhaps more likely that on the inboard side this would be bolted to the plate, rather than any supporting steel; if so, then the outboard side nearest the DOST would have to be similarly secured. This would provide a mechanism for aftward movement of the plate. In order to fracture the valve, the movement has to be between 13.7 mm (underwriters' case) and 28.1 mm (owners' case). These figures are within the range of possible movement. (2) Racking This could occur if the fore and aft steel member adjacent to the DOST supporting the plates expanded aftwards due to heat, whereas the inboard one had not. The result would be a distortion of the support into a parallelogram, with consequent movement of the plates. Alternatively, if the ladder was secured to the steel supports and not the plates, a similar effect could result. There is no evidence that this is what occurred, other than what can be seen in photograph F1 of the movement of the plates. If the fracture initiated on the aft side, and the only movement of the pipes was from forward to aft, then this would be a fundamental objection, as the Judge stated. But, in our judgment, it is not possible to be certain of this, and we do not think the Judge thought otherwise. If he did, we think he was in error. If in fact there was movement of the plates aft to any significant degree, then this would account for the fracture, albeit one that must have initiated on the forward side. Even though this is less likely, other things being equal, it is a possible cause. (O) The Judge and the expert witnesses It is plain that the Judge was greatly oppressed by the volume of expert evidence, both on the navigation issue, but particularly in relation to the fire. He had ample cause for concern; expert evidence in relation to the fire alone lasted for 32 days. The experts had a field day. Professor Dover expressed the problem very well in a passage quoted by the Judge at p 110: I still would find it difficult to know what caused the failure of the valve . . . I do not see the fatigue vibration (mechanism) as possible, but the other thermal mechanisms are all possible to some extent . . . I don't know how the valve failed . . . the most likely looking area is the thermal distortion of the structure . . . I cannot prove how the valve failed . . . it is beyond all of our powers because we don't have the full information . . . we can conduct stress analysis and it will point to possible modes of failure. It will even in some cases suggest more likely ones than others but it won't tell us how the valve failed . . . it is because we have so little information that we have quite a few avenues of investigation . . . we have seen one side produce some work, the other side mentioned something about it and the first side would change it again. The underlying reason for this is that the experts instructed on behalf of the owners were required to put forward possible mechanisms to provide an innocent explanation for the open tap and the initiation of an accidental fire for which there was no positive evidence at all, other than Mr Tsakiridis' denial that he had started it. If you start with the assumption that the tap must have opened without human agency, then it is possible for the ingenuity of experts to think of a possible explanation, provided a number of assumptions are made which can neither be proved nor disproved because the evidence no longer exists. Each expert in his own speciality is able to say: "This step is possible if certain things happened." But no one on the owners' side seems to have stood back and considered the likelihood of such a sequence of events taking place. Not surprisingly, the underwriters responded to the fatigue/vibration theory by calling equally distinguished and qualified experts. They could not prove that any step in the fatigue/vibration theory was impossible. All they could attempt to do was to show that each step was improbable or very improbable, a task in which, in our judgment, they I plainly succeeded. At p 81 of his judgment the Judge gave an admirable resume of the duties and responsibilities of expert witnesses. We have no hesitation in endorsing it. We would, however, add one word of caution in relation to par 4: That an expert should make it clear when a particular question or issue falls outside his expertise. It is evident that in this case the Judge was concerned to confine each expert to his area of expertise; but it is not always possible to do so and where the subject of inquiry is fire, an experienced fire expert, when he is assessing the significance of certain evidence, must be entitled to weigh the probabilities and this may involve making use of the skills of other experts or drawing on his general mechanical or chemical knowledge. Between pp 97 and 113 the Judge gave a synopsis of the evidence of all the expert witnesses and his impressions of them. He was clearly impressed by Dr Palmer, Dr Walker, Dr Baker and Dr Ward among the owners' experts. Although he preferred Mr Cook's evidence to that of Dr Taylor, where they differed, he was very critical of the inspection of Ikarian Reefer carried out by Mr Cook and Dr Bound (pp 113 and 114). This is a matter of some importance, because Mr Cook's opinion as to the likely cause of the fire was favourable to the underwriters and the Judge advanced this criticism of Mr Cook as the first ground for rejecting the opinion of both fire experts (p 130). In our judgment, the Judge was not being entirely fair to Mr Cook and Dr Bound. No criticism was ever made of them by Mr Clarke. As fire experts, they were familiar with breakage of metal structures as a consequence of fire; there are other examples which can be seen in the photographs. We do not think they are to be criticised because they did not there and then work out exactly the mechanism by which this could have come about. They are certainly not to be criticised on the basis that they did not foresee the fatigue/vibration theory as explaining the fracture of the valve, or that the tap might have vibrated open under the influence of the generators. Although they did not make a rigorous scrutiny of the area of the valve, such as would be necessary or desirable to prove or disprove the fatigue/vibration theory, we do not think they can be blamed for this. They clearly made a fairly careful examination. Mr Cook contemplated that it was only a preliminary inspection. We have no doubt that the underwriters' fire experts would also have wanted to re-examine the valve and the area around it, had they known before the vessel sank of the importance attached to it or had any idea of the fatigue/vibration theory. Of the underwriters' experts, we do not consider that the Judge expressed any material criticism of Mr Corlett or Dr Taylor. He appears, however, to have been somewhat critical of Dr Deegan, since he said that he: "was concerned about the change of view as to the probable initiation region of the fracture path of the valve." But he did not express the nature of his concern. Certainly he cannot have regarded Dr Deegan's change of mind as being dishonest, because he said in terms that he did not impugn the honesty of any of the defendants' experts. As we have already pointed out, the likely initiation of the fracture path depends very largely upon the correct interpretation of the photographs, and particularly photograph D. This is far from easy. In our judgment, Dr Deegan is not to be criticised for thinking that he might have been mistaken in his first interpretation of these photographs, particularly since his drawing (Deegan figure 3) was designed to illustrate some other point. Dr Deegan is a highly experienced and reputable metallurgist; there is nothing that we have been shown in his evidence to suggest that he was biased or partisan, inflexible, or unwilling to make concessions if they were called for. The Judge was very critical of Professor Dover. He said he had a number of reservations about his evidence and he dealt with them at length (pp 107-110). Much of this criticism is deserved; there can be no doubt that Professor Dover should have made it clear much earlier than he did that in the light of the criticisms made by Dr Palmer and Dr Baker the heating the valve and the heating the pipe theories, which he had advanced as possible explanations for the fractured valve, could not seriously be entertained. But it is fair to Professor Dover to point out that much of his reports and evidence was directed to showing that the fatigue/vibration theory was not plausible and many of the criticisms that he made were accepted by Dr Baker and Dr Palmer as the case progressed. At pp 118-119 the Judge sets out what he calls the changes of position in the underwriters' case in relation to the fracture of the valve. He says that it has caused him considerable concern. But again he does not make the nature of his concern clear. There are a number of points that should be made in relation to this criticism: (1) It is not correct that the underwriters had changed their case at a late stage from heating the valve/pipe to movement of the plates or structure. The principle of mechanical loading due to heating of associated structures was adumbrated in outline in Dr Taylor's first report B7 (1) 5/23 and 24 and more specifically in Dr Deegan's first report B7 (1) 7/6 and 7. It is true to say that the mechanisms of expansion of the ladder and racking were not put forward until Mr Corlett did so in the long vacation. But the differential movement of the structure was always part of the underwriters' case; heating the valve and of a pipe were alternatives introduced by Professor Dover which were not eventually persisted in. (2) In this type of case it is almost inevitable, and is certainly a common experience so far as the members of this Court are concerned, that experts will change or modify their views in the light of the opinion of opposing experts and cross-examination. This is because in fire cases, whether at sea or on land, much of the evidence is destroyed by fire and for one reason or another inspection may not have concentrated on every point that is subsequently thought to be relevant, and however good photographs may be, they frequently pose problems of interpretation. (3) As Mr Pollock pointed out, while it is legitimate to point to changes in position, the Judge was being less than even-handed in confining his criticisms to the underwriters' experts on this score. There are many examples where the owners experts abandoned or radically altered their theories. Some examples will suffice: (a) Cavitation, which was initially put forward as a major force of vibration, was abandoned during the long vacation. (b) Periodic tilting moments set up by the propeller which ultimately became the principal source of vibration to initiate the crack in the valve was first evolved during the long vacation. (c) There was a radical reduction from Dr Baker's "middle of the road" figure of 20,000 cycles at 100 Mpa, involving amplitudes of 159 mm, when it was clear that these figures were quite unrealistic. (d) The possibility of faulty generators was introduced on Day 57, perhaps when the implications of the Paris law became clear to the owners experts; (e) The possibility of a shock load from pounding was introduced for the first time on Day 59, without any written report; (f) On the navigation aspect, Mr Beatty put forward a large number of possible explanations as to why the 19 35 and 21 41 satfixes were not received. The Judge rejected those in relation to the 19 35 satfix, and all but two, which in our judgment he should also have rejected, in relation to the 21 41 satfix. When the Judge was reviewing the evidence in relation to the various theories put forward, he expressed the view in every case where there was any conflict of opinion between the experts, that he preferred the views of the owners' experts. He gave no specific reason for doing so, other than the general matters to which he had already referred. We find it somewhat remarkable that honest, distinguished, experienced and qualified experts called on behalf of the defence in relation to the fire should always be wrong, while those called on behalf of the plaintiffs should always be right. On two occasions in the course of his judgment, the Judge referred to the fact that Dr Bound was not called as a witness. At p 111 he expressed his surprise and at pp 130 and 132 he advances it as one of the reasons for concluding that the physical evidence was not inconsistent with the fatigue/ vibration mechanism. The Judge had raised the point during final submissions. It is not at all clear to us what his concern was; but it is clear that in some way he seems to have put it into the balance against the underwriters. He should not have done so; the fact that Dr Bound was not called was wholly irrelevant and there was a perfectly satisfactory explanation for not calling him. The position was this: underwriters had obtained leave to call two fire experts, against the owners' opposition, but only on the basis that there might be some difference between Mr Cook and Dr Bound as to what was observed on the inspection. In fact there was no difference whatever. Dr Taylor was far more experienced than Dr Bound; he would have inspected the vessel, but he was unavailable. The factual basis on which Dr Taylor was originally commenting in his reports was the information provided by Dr Bound contained in their joint reports. It was wholly understandable that the underwriters wished to call Dr Taylor to interpret and give his opinion on the significance of the factual evidence. There was therefore no reason to call Dr Bound. The owners never criticised underwriters for not calling him; and, as Mr Steel very frankly admitted, they would probably have objected had the underwriters sought to do so; there was after all no reason why they should have the benefit of calling two fire experts. All this the Judge should have appreciated. What he was not expressly told was that when the question of calling Dr Bound at a late stage was raised by the Judge, Mr Clarke told Mr Tomlinson that he did not wish him to be tendered for cross-examination. But that was the position. The Judge could have called Dr Bound himself, since he was available; but he did not indicate what he wanted to ask. 4. PRIVITY and MOTIVE (A) The case for the underwriters is that the stranding of Ikarian Reefer occurred at the owners' behest. This does not imply that some kind of conference took place at which the master and some members of the crew agreed to scuttle the vessel off the coast of Africa on this particular voyage. The submission of the insurers is that the wish of the owners to dispose of the vessel at some convenient time and place in circumstances giving rise to a claim for her insured value had been informally conveyed to the master before the voyage to Abidjan or even earlier. The question whether the owners had a motive for scuttling the vessel therefore becomes significant. Proof that a motive existed is not of course conclusive that the owners were party to the scuttling, as the learned Judge pointed out. There are no doubt many occasions when a shipowner would prefer to be paid the insured value rather than retain his ship without conniving at its destruction. Yet proof of a motive is of course relevant when such an accusation is made. The presence of a motive is a matter to be taken into account when deciding whether the owners were privy to the ship's destruction. Equally the absence of motive will assist the owners to rebut an accusation of this nature. The motive alleged is, of course, entirely financial, but even so the character of the Comninos brothers is significant in determining the presence or otherwise of such a motive on their part. In 1985 they were well-known shipowners with a good reputation. Costas Comninos had held a number of important appointments in the industry including that of second vice president of the Union of Greek Shipowners. He was also a director of the Board of the Greek Chamber of Shipping, of Lloyds Register of Shipping Greek Committee, of Norske Veritas, the Superior Council of Maritime Accidents (ASNA) and of the Sunderland P and I Club. It has been a central plank of the brothers' case on motive that they would not jeopardise their reputation by scuttling. Ikarian Reefer represented only a small part of the tonnage which they owned. Quite apart from the damage to their status in the industry they would be severely prejudiced in obtaining insurance cover for their other vessels or further loans from their bankers. They rely on their high reputation in all their commercial dealings and it is significant that after the casualty the London insurance market continued to insure their fleet. Before considering the evidence relating to Ikarian Reefer in greater detail, we should turn to the overall financial position of the group. We accept Mr Steel's submission that only the period of some weeks or months before the loss is directly relevant, because the underwriters' allegation is that the master and others of the crew who were implicated in casting away the vessel must have received their instructions or some other indication of the owner's authority or approval beforehand, which in practical terms must have occurred, if it occurred at all, during the period between, say, December, 1984/January, 1985 and the vessel's last voyage which began on Apr 3, 1985. It is necessary, however, first to say something about the background, beginning in September 1983. The group consisted of the whole of the' brothers' shipping interests, which they owned jointly in the proportion 80 per cent (Costas) to 20 per cent (Anthony). The group had a conventional structure in that each of its vessels was owned by a single-ship company, incorporated in Liberia, Panama or elsewhere, but with a central management at its headquarters offices in Piraeus. By 1983, there were in effect three operating divisions: tankers, bulk carriers and reefer vessels; and the group was prominent in the reefer market, which is highly specialised. In 1982 a joint venture with P & O involving six reefer vessels was disbanded when P & O withdrew. The group acquired P & O's 50 per cent share in these vessels which continued to be employed in what was known as the Lauritzen pool. The three- or four-year period ending in 1982 was highly profitable for the group but the shipping market in general and the reefer market in particular were entering a period of decline. It was in these circumstances that the group had its first dealings with the Chase Manhattan bank, through the bank's office in Piraeus. The group was interested in financing the purchase of two new-building bulk carriers, Hulls 295 and 296, which it had ordered from Hyundai, the Korean yard, at a price of US$14.1m each for delivery in August and November, 1984 respectively, and in refinancing loans totalling US$21.7m which were secured on five vessels, including Ikarian Reefer. The four others were two bulk carriers and the Athenian and Attica reefers. In the result, the proposals for refinancing the five-ship loan and for one only of the two new-buildings were accepted by Chase in August/November, 1983. The brothers had been introduced to Chase by one of the latter's employees Mr Alexiou who gave evidence at the trial. He was friendly with the Comninos brothers on a social basis. Indeed they had both been present at his wedding, Anthony Comninos being the best man. It is clear from the Chase file that the introduction was welcome as it was anxious to consolidate its position in the shipping market particularly at Piraeus. The financial director of the group, Mr Poulman, who also gave evidence, was the person most directly concerned in dealing with the bank. It is the owners' case and not disputed that Mr Poulman was in regular contact with the bank and it was customary for him to supply it with copies of the group's audited accounts. Although responsible for the introduction Mr Alexiou was not given the handling of the group's account. That was the task of Mr Papoulias. He, like Mr Alexiou, was based at Chase's Piraeus offices and was responsible for making recommendations about the loan to the London or New York offices of Chase. The documentation regarding the loan is incomplete because while the bank's London file has been disclosed the important file kept by its Piraeus office has not been produced. Effectively, no documents apart from routine communications between themselves and the bank have been produced by the owners themselves. The different departments of Chase communicated with each other in relation to a particular loan through the medium of a form of document entitled Senior Officers Approval Memorandum (SOAM). The Comninos brothers received enthusiastic support from Mr Papoulias in these documents. He refers to them for example in a SOAM dated June 10, 1983 in these terms: . . . the principals involved Costas and Antonias Comninos have an excellent reputation in the market place for being conservative owners who have come up the hard way and are not about to lose everything . . . and describes them as having a "prime name" in the Piraeus market. Subsequently more senior officers sound a note of caution. On Sept 13, 1983 Mr Dahms of the New York office sent a telex to Mr Papoulias in these terms: While I gather that you are very high on this group I must advise you that the most recent approval here did not come easily and we feel that it is best if you make sure that we are walking, well before we start to run with this company . . . The documents produced by Chase's London office contain detailed references to the group's financial state at this time and to its plans for the future. There were reported to be cash reserves of US$23.6m (at May, 1983) and it was agreed that cash deposits on average of US$5m would be maintained with the bank. There were minimum value clauses in the loan agreements which required that the security would be not less than 125 per cent of the sum outstanding on the loan. The group's strategy in June, 1983 was to modernise its fleet and maintain its involvement in bulk carriers, tankers and reefers. The market was described by the bank as "soft" but there was considered to be some prospect of an upturn and the intention was to dispose of 10 of the older vessels, out of a total fleet of 26 which included 13 reefers, "as soon as market values pick up". The freight market, however, and with it the values of the trading vessels continued to decline. The bank's internal valuation of the five vessels on which the US$21.7m loan was secured was only US$20m and it was recorded that an independent valuation had produced a figure of US$255m which although greater than the amount of the loan was marginally below the 125 per cent minimum value requirement. Some evidence was directed to the bank's attitude towards this and subsequent breaches of the loan agreement. The owners contend that this factor of itself, in a depressed market where many shipowners and their bankers were in a similar condition, was unlikely to be decisive of the bank's attitude towards them, and this in our view can be accepted. Despite a worsening market situation, the group continued to enjoy the support of Chase throughout 1984. The group's financial position, however, was undoubtedly weakening still further. The interest cover (profit before interest against net interest payable) declined from 3.6 in 1983 through 1.2 in 1983 to 0.6 in 1984. The liquidity ratio (current assets against current liabilities) was 1.9 in 1982, 0.9 in 1983 and 0.8 in 1984. During 1984, the delivery date for the second new building Hull 296 was postponed by agreement until January, 1985 and the group's development plans were reported by Mr Papoulias in May, 1984, following a meeting with Mr Poulman, as follows: Following the initiation of the scrapping programme and the disposal of 3 of the fleet's vessels, the Group has started to implement a diversification into tankers, always with employment. Significantly, a major reefer operator, Salen, collapsed during the autumn of 1984. This, according to ML Papoulias (SOAM dated Jan 28, 1985): Brought about a substantial decline in the reefer vessel values and a price cutting war in the major reefer routes . . . to face this situation the group has implemented a program to shift away from the older B/Cs and reefers in its fleet, employed on a spot basis, to tankers, with confirmed employment to improve group and cash flow. By January, 1985 the group's position was critical. Costas Comninos accepted that it was "difficult" but the more stringent adjective in our view is justified. The stark facts were these. Bank loans, including US$33m from Chase, totalled US$75m. Interest payments of US$7.5m and principal amounts totalling US$4.5m were to become due in 1985, and the group's operating profits for 1984 were no more than US$1.2m, far short of the required figure if interest, let alone capital, payments were to be made as scheduled. It was only by disposing of vessels that the group's commitments had been met during 1984, and even so there had been an overall loss of US$8.6m in that year. The asset coverage held by Chase was reduced to 80 per cent according to its own estimates and 100 per cent by independent market valuations. The value of the three reefer vessels mortgaged to Chase was considered by them to be US$6.3m against US$17m previously. This included US$300,000 for Ikarian Reefer but it will be recalled that the value some months later is agreed for the purposes of these proceedings as US$450,000. Last, but by no means least, the group's cash held by Chase had declined to, it seems, US$3.2m. The SOAM dated Jan 28, 1985 recommended that the group's risk rating be reduced from 4 (on a scale of 10) to 6, and the documents show that negotiations began then which continued until November, 1985 when Chase agreed to reschedule the repayments of principal under its loans. A rescheduling agreement with Midland and Hill Samuel was reached much earlier, in March, 1985. It is the underwriters' case that the owners were under pressure from the bank during the period of January to March, 1985 and that they were under threat of the bank refusing to reschedule the outstanding loans if they were unable to maintain the payment of principal and interest. However, while the documentation reveals that the bank was imposing conditions of some strictness in its advances there is no indication of any threat by the bank to refuse a rescheduling of the loan. Indeed the general tone of the SOAM was that the bank in its own interests would be prepared to assist in a postponement of repayments until 1986. The market would hardly bear the collapse of another major figure in the reefer industry and if the bank called in its loan the likelihood was that it would recover less of its outlay than if it continued its support. Costas Comninos resisted suggestions made in cross-examination on behalf of the underwriters that he was under pressure from Chase and that this pressure, combined with the need to obtain Chase's agreement to rescheduling their loans, drove him to a position where he was prepared to authorize the casting away of his vessel. He protested that: You always imagine that I always was in conflict and I had problems with the banks. I had a very good cooperation with the banks . . . Later, he said that the bank and its clients, including the group, were partners. We consider that his protests were justified. As Anthony Comninos put it, they were not at war with the bank, but "the war [was] with the market, to survive". The significant fact, in our judgment, is not that the group was under pressure from the bank to change its plans, but that its own operational strategy, on the basis of which it sought to obtain co-operation and agreement from the bank, included the disposal, during 1985, of a number of vessels, including Ikarian Reefer. These plans are amply documented from the Chase London files. The first SOAM dated June 10, 1983 includes the general statement of intention, already quoted above ("10 of their older vessels"). Soon afterwards, on Sept 12, 1983, Mr Papoulias reported after a meeting with the Comninos brothers and Mr Poulman: Depending on market development they plan to dispose of another 3-4 vessels by June '84, most probably the IKARIAN REEFFR (built 1968) and 2-3 of their older bulk carriers. The loan package including Ikarian Reefer was refinanced after that date and there is no mention of Ikarian Reefer as a candidate for sale or scrapping during 1984, but by January, 1985 the position was clear: Starting in early 1984 the group has disposed of 2 B/Cs, 4 reefers and 2 tankers . . . this program is continued until 1985 with two vessels confirmed for scrapping and another eight vessels (4 B/Cs, 4 reefers) to follow." (SOAM dated 28th January 1985). The bank's annual review on the same date confirmed Ikarian Reefer was one of the four reefers: . . . that will go for scrap this year . . . the net proceeds of this scrapping will be around $10.3m. In a further SOAM dated Mar 6, 1985 Mr Papoulias recommended a further reduction in the risk rating to 8. This triggered a "criticised loan" procedure which called for a special report to Chase's head office in New York, dated Mar 13. So far as the group's financial position was concerned, figures for 1984 had become available which showed that the position had "deteriorated considerably", but projections showed that the group would be able to meet its 1985 debt service requirements "mainly through the disposal of vessels." The group's "action plan" included the disposal of 11 vessels for scrap, including Ikarian Reefer at an estimated sale price of US$550,000. This "Scrapping Program" was included as Appendix II to the SOAM and the heading shows that it was originally drawn up for 1984, altered to 1985 in manuscript. The total proceeds were estimated at US$11.7m against a LTD (long term debt) reduction of US$10.7m but the programme does not reveal that proceeds of US$550,000 would fall short of the loan outstanding on Ikarian Reefer by some US$800,000. Mr Steel submitted that the decision when and where, and to whom, to sell the vessel was one for the group in the exercise of its commercial judgment, and this we can accept. We must also assume, however, that the bank's reaction to the action plan was reported to the group and specifically to the Comninos brothers and this would have a considerable influence upon their decision. The bank's reaction was as follows. Mr Papoulias recommended that the group should be allowed to continue with its asset disposal plan "up to 30/6/85" but his seniors apparently were not content with this. The SOAM is minuted by one of them "Action plan to be completed by May 1, 1985" and another on Apr 2, wrote that his approval was subject to the Criticised Loan Report including a detailed action plan and being submitted within 14 days. The CLR (in fact dated Mar 13) states that the group has established a strategy which included the disposal programme referred to above. In our judgment, therefore, the group in January/March, 1985 was committed to disposing of Ikarian Reefer and was under considerable pressure from the market situation if not from the bank to do so sooner rather than later during the year. There is no evidence from any of the owners witnesses, including the Comninos brothers of any discussions as to what steps should be taken to dispose of Ikarian Reefer, whether by sale for scrap or otherwise, and at what stage during the year. Figures are produced to show that the vessel might have continued trading until the end of the reefer season, which is usually in the summer of each year. Thereafter it is usual for the vessels to be laid up until the new season begins sometime in the late autumn. It is suggested that two more trading voyages might have been carried out after the time charter voyage which she was performing at the time of her loss, and that these might have been profitable to the extent of an estimated US$72,000. Apart from the charter-party entered into on Apr 3, there is no documentary evidence of any decision taken with regard to her future, during the crucial period in February/March, 1985. The owners' witnesses were concerned rather to resist any suggestion that they were involved in any such decisions during this period, or that they had any conversation with the master or other persons on board the vessel. We find this absence of evidence surprising, unless the reason for it is that there were some discussions and a decision was reached which the witnesses now are unwilling to reveal. Given the plan to sell Ikarian Reefer for scrap sometime during the year, given the apparent dearth of employment opportunities during the remaining part of the season and the modest profits, if any, which the vessel might be expected to earn, and given the clear view of the bank that the action plan should be implemented promptly, it seems inconceivable that no decision was made or that no discussions took place. Added to which is the fact, referred to below, that the brothers were prepared in some cases to sell vessels, which the bank wished them to dispose of, to buyers in which they, the brothers, held some beneficial interest, if they considered that the vessel in question did have a profitable future. It may be supposed that this was at least a possible course of action which would have occurred to them in relation to Ikarian Reefer, unless they considered that it should be disposed of forthwith. Yet their evidence is that no special consideration was given to the vessel's future, except that it should continue trading for some indefinite period. Moreover, a sale of this particular vessel would give rise to a specific problem in the group's relations with Chase. The scrap value was (say) US$450,000. The outstanding loan in respect of this vessel, part of the US$21.7m facility, was in excess of US$1.3m. Accordingly, if the vessel was sold, not only would the sale proceeds go to the bank, but the fact of the sale would trigger off the repayment of the balance of the loan. It may be that the bank would agree to extend the loan and apply the balance to the repayment of principal and interest over a period but the advantages to be derived from the sale by the owners would be limited. If on the other hand Ikarian Reefer was lost by an insured peril the position would be quite different, by reason of her insured value being greatly in excess both of her scrap value and of the amount of principal remaining due to the bank: Insurance proceeds US$3 million Principal due to bank US$1.3 million Balance available to owners US$1.7 million. Instead of being met with an obligation to pay the bank US$850,000 in addition to the sale proceeds the brothers would have US$1.7m in hand. The group had recently received a payment under an insurance policy in respect of the total loss of another of its vessels Evangelia C so the brothers would be well aware of the short term cash flow advantages of such a claim. It was in these circumstances, the underwriters argue, that the brothers were motivated to scuttle Ikarian Reefer. The advantages of recovering the insurance monies compared with selling it for scrap have been pointed out above. The underwriters further submit that the relationship with the bank was not as satisfactory as the owners made it out to be at the trial. Mr Alexiou, the only witness called from the bank, was in reality a subordinate figure and in any case not directly concerned with the running of the group's account. By the beginning of 1985 the group was in its fourth year of crisis and facing a shortfall of US$12m. According to one of the bank documents an operating profit of US$8.4m was forecast, but neither Mr Poulman nor the brothers admit to being the source of that information. Their estimate of the operating profit for 1985 was US$3.4m and this figure was taken into account in calculating the US$12m shortfall. The reality was that the brothers were well aware that the group was facing a serious financial crisis. They sought to remedy the situation and restore their cash flow not only by selling off ships for scrap but also by squeezing down capital; ie by savings on inventories and no doubt by calling in debts owed to the group and taking steps to delay payment to its own creditors. Such a course would have a dire effect in the long term as reducing its capital was equivalent to eating the seed corn. This gave an indication of the brothers' awareness of the serious nature of the crisis, and the reason why it was necessary to take extreme steps. (B) Specific transactions Many other aspects of the group's financial history were relied upon by one party or the other as being relevant to the issue of connivance, either directly or indirectly as evidence of the owners' credibility and commercial morality. These can best be dealt with by reference to a number of individual transactions, first, the Esperanza loan. This related to the bulk carrier (Hull 296) built by a Korean company Hyundai. Although there was a 30-month holiday before payments of principal became due the plaintiffs took steps to avoid taking delivery alleging defects of manufacture. An arbitration on this dispute was taking place in London at the very time Ikarian Reefer was making its way to Abidjan. Hyundai sought to obtain an injunction restraining the owners from alleging these defects, which they said were trivial. They contended that the owners' case was not bona fide and was simply a delaying tactic necessitated by their critical financial position. Mr Jung swore an affidavit to this effect in support of the application. Mr Alexiou then swore an affidavit dated Feb 22, 1985 in which he said: The Comninos group and those they represent are in a satisfactory financial condition. The group meets all its obligations to the bank (Chase) and operates profitably. The difficulties which they face are created by the current market conditions which are the same for everyone. The Comninos group are coping better than most. I would be quite prepared to recommend a further loan to the group, on reasonable terms, if such a request was made. While the last sentence may be relied upon by the owners, this Court has difficulty in understanding how on the evidence before us the financial state of the group could then be described as satisfactory. It was not operating profitably because its obligations to the banks could not be met out of operating profits alone. No doubt as a result in part at least of this evidence, Hyundai's application was dismissed. The arbitration proceeded and the owners' complaints were dismissed. Another transaction at this time which may assist the owners, although to a limited extent, is the Harris loan. This was intended to provide a cash profit to be derived from a single voyage. The scheme was that the group would purchase an old tanker in Europe which would then be sailed to the Far East where the scrap prices were higher. If a cargo was available so much the better. It would then be sold for scrap and a profit made on the transaction. The bank authorized a loan for this adventure of US$2.55m, the advance taking place in March to April, 1985. Although this is evidence of the bank's support of the brothers it was taking only a minimal risk as it was a condition of the loan that the group maintained a deposit of US$1 m with the bank. In the event instead of generating a profit of US$50,000-$100,000 the group incurred a loss of US$272,341. This had the effect of depleting the group's cash reserves, held by the bank, by that amount; and the main significance of this apparently unusual and one-off transaction, in our judgment, is that the bank does not appear to have contemplated, when the advance was agreed, that the group would not in fact continue to trade. (C) Private assets The underwriters also dispute the owners' assertion that they had sufficient private wealth to inject into the group should it not be able to meet its obligations to the bank. In a letter from the plaintiffs' solicitors dated Aug 15, 1991 the writer considered that the group had a thriving shipping enterprise and went on to set out the plaintiffs' private assets: (1) Land in Virginia in which the brothers jointly held a 70 per cent interest. As between them Costas had an 80 per cent share and Anthony 20 per cent. Their interest had cost them US$5.1m. Although it was agricultural land it adjoined the Washington Dulles Airport, the thinking being no doubt that if the airport expanded the land would be purchased at a much enhanced price. The estimates of its value differed wildly but the plaintiffs' case was that it was worth US$20 m. Mr Poulman thought the total value was US$40 m. In the result the dispute is academic. First, the airport has not expanded and the land is something of a white elephant. Secondly, the land is charged for the benefit of Costas Comninos' wife and children so it was not available for the business anyway. It should not have figured in the solicitors' letter. (2) Villa in Florida (jointly owned). This was purchased in 1981 for US$334,800. Contrary to the assertion in the letter it was sold in 1983 so it was not available as an asset in 1985. (3) Villa in Florida. Held for the benefit of Costas by a Liberian company. This was sold in 1986 for US$230,000 which was US$55,000 less than the original purchase price. (4) Villa in Florida (Anthony). This also was sold at a loss but the proceeds of sale in 1985 were paid to the account of Anthony's wife and daughter. It was not available to Anthony as a private asset. In cross-examination it was suggested that this was done to put the proceeds out of the reach of the bank. (5) An historic house in Athens valued at US$775,000 (Costas). Apart from a suggestion that the value was a little on the high side not much seems to have been made about this asset in cross-examination. (6) Villa outside Athens (Anthony). Sold in 1985 but it was claimed that this was done to put it out of reach of the bank. (7) Both brothers had bank accounts in Switzerland either with Swiss Bank or Bank Luis. These were under fictitious names and in the case of Costas the deposits were substantial. The insurers suggest that the fictitious names had a sinister connotation but according to Mr Poulman the practice was conventional in Switzerland. The existence of the deposits appears not to have been revealed to Chase. (8) The yachts Alpega (Costas) and Billa (Anthony). These vessels were valued in the solicitors' letter at US$6 m and US$600,000 respectively. The valuations were much exaggerated, the brothers saying at the trial that the appropriate figures were US$1.7 m for Alpega and US$175,000 for Billa. Both vessels were group assets and appeared in the audited accounts. The complaint made by the insurers is that the brothers falsely told Chase that the yachts had been sold when in fact they had not. Alpega had simply been transferred to a company controlled by Costas while Billa had been sold to Anthony's brother-in-law. However, the moneys obtained on transfer were paid into the group's account. In fact Anthony enjoyed no benefit from the sale of Billa. Alpega continued to be used by Costas as a business asset. It was contended by the brothers that the bank knew very well that Costas had kept the yacht. Certainly Mr Alexiou was aware of it and the yacht was treated by everyone as "Costas' toy". The thrust of the suggestions made by the underwriters were to some extent inconsistent. On the one hand they were saying that the brothers were not rich or at any rate not so rich as they claimed to be and on the other hand that they were taking steps to dispose of their assets to avoid the bank taking possession of them. The reason why the bank was in a position to pursue these assets was that at the bank's request the brothers had each entered into an agreement to guarantee the repayment of the bank's loans to the group. The brothers' attitude was that the guarantees were only intended to demonstrate their moral commitment to the group and that the bank never had any real intention to enforce them. They were to ensure that the brothers did not "walk away" from the single-ship companies which formed the group, although there had never been any sign of their intending to do so. Additionally no list of assets was ever sought by the bank. On the other hand it is difficult to conceive of a bank with the benefit of a guarantee failing to enforce it, if necessary. Eventually, Chase did so against Costas Comninos in 1988 when a compromise was reached, although not against Anthony Comninos, who separated from his brother and in that sense left the group in 1986. The disposal of Esperanza may conveniently be dealt with at this point. This was the vessel purchased from Hyundai. In 1987 the bank gave instructions for its sale, at a price of not less than US$10.4m Costas Comninos claimed that he was against the sale but obeyed the bank's instructions to find a purchaser at that price. He arranged for the vessel to be acquired by a company in which he, Costas, held a beneficial interest. The bank was paid the sum it had asked. Two years later the vessel was sold for approximately US$15m. The underwriters claim that this demonstrates his dishonesty. The bank, they say, had stipulated that the original sale was to be at arms length to a third party buyer. This is supported by a letter dated Nov 24, 1987. They accuse him of dishonesty in the transaction by not telling the bank that he had sold it to one of his own companies thus enabling him to make a substantial profit at the bank's expense, though only because his prediction of the market and of the ship's later value proved to be more accurate than theirs. When he was cross-examined about the agreement with the bank to dispose of Esperanza, Costas Comninos claimed that he did not understand the meaning of the expression "arms length sale to a third party". We do not find this answer convincing. There were other examples of vessels being sold by the brothers for scrap to their own companies ie companies not in the group in which one or both of them had some interest, without the bank being informed. In 1985 Ionian Reefer was disposed of in this way and Aegean Reefer in 1986. The same strategy was employed with two more reefers, Athenian (later named Golden) and Attica (Silver). It is the underwriters' case that this deception was continued over a number of years. There is no dispute about it, the Comninos brothers admitting that they did not disclose all their assets to the bank. While accepting that there is a wide gulf between deception of a bank and the scuttling of a ship the evidence reveals that the two brothers were capable of falling short of the standards of honest and respectable businessmen when on occasions it suited them to do so. The learned Judge described Mr Alexiou as an impressive witness and had no hesitation in accepting his account of the dealings between Chase Manhattan and the group. In particular he accepted that it was the policy of Chase to reschedule loans rather than pursue a confrontational approach. The Judge found himself in no doubt whatsoever that the bank would have granted the Comninos brothers a loan rescheduling irrespective of the fate of Ikarian Reefer. There are some difficulties about Mr Alexiou's evidence. Apart from the affidavit already referred to the Judge appears to have accorded him a consequence in the bank which in reality he did not enjoy. It was perhaps unfortunate or unavoidable that the Chase witness was one who was not immediately concerned at the time with the Comninos portfolio rather than Mr Papoulias who was. The Judge also found Mr Poulman to be competent, responsible and reliable and accepted his statement that the discussions during 1985 did not give him an indication that the bank would be uncooperative. In determining the question whether the bank would have restructured the Comninos debt in 1985 the Judge also relied on the circumstances of the Harris loan. Although this was not a significant loan for the reasons already given and the bank refused two other applications for Comninos loans at about the same time, it appears to us that there was evidence before the Judge to support his finding that the bank would have restructured. The important feature however is not whether the bank would in fact have restructured the loans but the Comninos brothers' perception of what the current market position and the financial situation of the group required them to do in order to ensure the survival of their business. Once agreement had been reached with Midland and Hill Samuel, in March, 1985, the group's relations with Chase became paramount. (D) Credibility The Judge reviewed the evidence which the insurers relied upon to attack the brothers' credibility. He commented that some of the brothers' answers reveal an unsatisfactory approach to their dealings with the banks though pointing out that their cross-examination had been based on the bank's London files rather than those from Piraeus and went on to say: But even if dishonesty had been clearly estab1ished in one respect such dishonesty would not necessarily involve dishonesty in another and different respect. As Lord Sterndale MR said in The Elias Issias (1923) 15 Ll L Rep 186: It is a long step from deceiving a bank to scuttling a ship. Having seen them give evidence over a number of days and watched their demeanour in the witness box the learned Judge accepted their evidence that they did not give instructions to cast away Ikarian Reefer. Mr Steel submits that in the light of those findings, based on the learned Judge's assessment of the witness, this appeal ought to fail. While accepting that motive is relevant to the limited extent referred to above the real issue is whether the brothers connived at the scuttling, and the Judge found in terms that the underwriters had not proved that the owners in any way consented or were privy to that action. Additional matters There are a number of further grounds which were canvassed in argument and have not been dealt with in this judgment. (1) There would be no significant benefit to the group from a scuttling of Ikarian Reefer as the insured value was less than 4 per cent of its total indebtedness. While this is accurate as a statement of fact it takes no account of the substantial cashflow advantages of a successful insurance claim, compared with selling the vessel for scrap. The sums involved were not insignificant in the context of the group's relations with Chase during the important period from January to March, 1985. (2) Ikarian Reefer was an unlikely choice for a deliberate loss. At the trial the owners identified two other vessels Anna C and Anastasios C which they suggested were more likely candidates if any vessel was to be scuttled. Before us, Mr Pollock pointed out that the proceeds of any insurance claim in respect of those vessels had been assigned, not to Chase but to Hill Samuel, so that they would not have been available to strengthen the group's position in its negotiations with Chase. More generally, we consider that the essential question is not whether the owners decided to resort to scuttling a vessel and then in the light of that decision selected the most likely candidate from their fleet. Rather, Ikarian Reefer was already ear-marked for disposal and the owners had to decide when and how to achieve that result. As already explained the reefer season would end in May/June and there would be no possibility of further employment of Ikarian Reefer until the autumn. The bank required the "action plan" to be implemented by June, 1985. (3) The insurers make no suggestion of any complicity on the part of Mr Poulman, yet if there was a conspiracy to cast away Ikarian Reefer it would be difficult to exclude the financial director of the group. We must also bear in mind, however, that the decision which was required was primarily an operational matter and we do not understand that the finance director would necessarily be involved. (4) Ikarian Reefer was engaged on a profitable voyage and her loss would be likely to antagonise an important time-charterer, Chargeurs Reunis of Paris, in what was on any view a limited market. Moreover, the master who had an unblemished record was shortly about to retire. These are important considerations but in fact there is no evidence of the charterers' reaction when the vessel failed to present for loading at Abidjan and we were told only that they made no claim for any loss or expense suffered thereby. Moreover, the fact that the vessel was empty of cargo on the approach voyage was an advantage, rather than otherwise, if a scuttling was planned. Finally, the fact that the master was about to retire might make him more amenable to a suggestion of continued on-shore employment, and it could even be regarded in certain circumstances as an advantage if a scuttling was planned, because the consequence of his master' s certificate being suspended would be less serious for him than for others in mid-career. (5) The submission by the underwriters' that there was some kind of standing order to the master to scuttle the vessel at a convenient moment is unsupported by any evidence, apart from the fact of the grounding. However the absence of documentary evidence, even if there was such an instruction, is hardly surprising, and the nature of the allegation is such that any direct evidence supporting it is unlikely. There were ample opportunities before the vessel sailed on her last voyage to tell the master that if an opportunity arose to scuttle, this would be welcome to the owners. What is surprising, in our view, is the owners' insistence that there were no communications with the master, by them or on their behalf, as regards either the vessel's future or his own. (6) There is no evidence of any payment or other reward to the alleged conspirators, that is, the master, the chief engineer and Tsakirides, nor of any noticeable improvement in their life-styles after the loss occurred. Again, the absence of documentary or other direct evidence is not necessarily significant, and we bear in mind that the owners themselves have not received any payment from the underwriters or any other benefit in consequence of the loss. Moreover, the master's evidence was unsatisfactory in this respect. He, or the owners, failed to disclose two overseas bank accounts, and the amount of payments which he is now said to have made towards the purchase of his son's house are not known. No satisfactory explanation of the failure to disclose bank statements has been given, and in our view this cannot simply be glossed over, as the Judge was inclined to do. (7) Scuttling is an opportunist crime which will only be committed if it can be done without undue risk to those on board, and in the circumstances where a plausible explanation can be given. In the present case the risk to the crew was minimal. The vessel went aground on a sandy bottom and it was likely that there would be ample time for the crew to take to the lifeboats. The site was close to well-used sea lanes as was demonstrated by the response to Ikarian Reefer's distress calls. 5. CONCLUSIONS It is now necessary to draw together the various threads of this case and reach conclusions. (A) The grounding (1) For the reasons given above we have come to the conclusion that, notwithstanding that the Judge was clearly influenced by the demeanour and his assessment of the witnesses, this is a case in which we are entitled, and indeed bound, to make our own findings of fact where they differ from the Judge's. (2) Once it is established, as in our judgment it is, that the master knew, within a small margin of error, where the vessel was at 20 00 hours, the alteration of course at about 20 20 hours to 115 deg. was obviously one which would take the vessel very close to the southern edge of the Shoals, if not into them, especially if no allowance was made for the incoming tide. Unless he had temporarily taken leave of his senses, as to which there is no suggestion, it is inconceivable that such an experienced master could make so egregious an error. (3) We are satisfied that the master was aware of the 21 41 satfix. As is conceded by the owners, this would have shown him, if he did not already know, that he was sailing into shallow water at full speed. (4) At the time of the impacts the master knew he was not in deep water. He must have known that the most likely explanation was that the ship had hit bottom and not a log. His evidence on this was untruthful and must have been dishonest. (5) For the reasons given in the section headed "The Judge's assessment of the master" we cannot accept that the master was a truthful, honest or reliable witness. (6) We are driven to the conclusion and are convinced that Ikarian Reefer was deliberately run aground. (B) The fire (7) For, the reasons set out in the section 3G "the open tap", we are of the opinion that the owners' explanation for the open tap is very unlikely and not plausible. (8) From our analysis of the vibration/fatigue theory, we conclude that each of the required steps (or combination of steps where we have considered them together) is either improbable or highly improbable. These steps all have to take place in sequence and within a very specific and tight time schedule. Added to this is the fact that some remarkable coincidences have to occur, if the theory is correct and assumptions have to be made for which there is no evidence, save the denial of Mr Tsakiridis, that he started the fire. We have no hesitation in concluding that the theory is wholly implausible. (9) For the reasons set out in the section 3L "I cigarette or burning rag", we conclude that neither of these explanations is plausible. (10) In stating his conclusions on this part of the case, the Judge was content to say that the various steps in the vibration/fatigue theory were possible, and that the theory itself was possible. He failed to consider, as he should have done, how probable or improbable it was that each or all the steps could occur. It was likewise with the cigarette and burning rag theory. (11) In reaching his conclusions on this part of the case, the Judge's approach seems to have been that unless he could exclude the possibility that the tap opened automatically or that the fire started accidentally, he would be bound to accept the denial of Mr Tsakiridis. In our judgment, he was in error in this approach. His denial had to be tested against the probability or plausibility of the explanations put forward. (12) An important factor in the Judge's reasoning was that he relied upon the fact, as he had found it to be, that the grounding was accidental. If this was right, then obviously it is less likely that the fire was deliberate. The Judge should not have made up his mind on this point until he had heard all the evidence in the case. If the correct conclusion is that the grounding was deliberate, as we hold it was, this casts a flood of light on the cause of the fire. The only purpose that we have been able to perceive in deliberately grounding the vessel is as a preliminary stage in achieving an actual or constructive total loss of the vessel. Setting fire to it is the obvious way of doing so in order to achieve the necessary amount of damage. Moreover, it has the advantage of destroying much of the evidence. We do not accept Mr Steel's submission that a more suitable method would be to let the water in. This would involve the vessel being aground in shallow water and the ingress of water would not readily be explained because of the nature of the bottom. (13) Nothwithstanding the Judge's assessment of Mr Tsakiridis and the chief engineer, we are driven to the conclusion and are convinced that the vessel was deliberately set on fire, almost certainly by Mr Tsakiridis, and probably with the knowledge and connivance of the chief engineer. The coincidence of Mr Tsakiridis' presence at the seat of the fire, 10-15 minutes before it was detected, is astonishing and highly incriminating. It is exactly the sort of time delay that one might expect with a diesel fire, which is slow to get going. In his evidence, Mr Tsakiridis said that after the fire he had never discussed with anyone of the crew the fact that a fire had started in the engineroom so soon after he left it. In our judgment, this is suspicious and suggests that either the crew knew what had happened, as in the case of the chief engineer, or that in the case of innocent members of the crew, they considered it prudent to ask no questions. In fact, in one incautious answer, the chief engineer nearly gave himself away. At D35/p. 39D there is this passage in cross-examination: [Q] It did not strike you as odd that there was no fire when Mr Tsakiridis was down in the engine room but that a fire had broken out by ten minutes after he had returned? [A] (through the interpreter): It struck me as odd since I told him and he went by the generators and he had a look around. It did strike me as odd. [Q] Why was it particularly strange, bearing in mind that he had been [past] the generators? [A] (through the interpreter): Because he told me that he went by there and he did not see anything dangerous and I said to myself: Well how is from one moment to the next that this happened? [Q] Yes, and you automatically assumed, did you, that the fire had broken out around the generators? [A] (through the interpreter): I could not form a view immediately as to where this fire was. The mention of the generators is very odd, unless he knew that was where the fire was to be started. (14) We do not accept Mr Steel's submission that there would be any difficulties in setting fire to the vessel in the way postulated by the underwriters. Neither of the fire experts could see any difficulty. A grating could be raised and replaced; charred cotton waste was found in a position which would be entirely consistent with its having been used to start the fire. Alternatively, it could have been put through the gap at the side of the saveall which is clearly illustrated in Mr Cook's photographs 45 and 46. (15) There are possible explanations of the fractured valve which are consistent with arson. The evidence is simply not sufficient to exclude all other explanations other than the moving of the plates and the vibration/fatigue theory. In any event, for the reasons we have set out in the section headed "moving the plates", we conclude that this is a possible mechanism, a view which seems to have been shared by the Judge. (16) For these reasons, we are driven to the conclusion and are convinced that the fire was deliberately started. In reaching this conclusion, we make it plain that we have had regard to the evidence in relation to motive. (C) Motive (17) In our judgment, there was clear evidence of motive and a good reason to dispose of Ikarian Reefer by scuttling, notwithstanding that her insured value was not as high as some other vessels in the fleet or that the receipt of US$3m. would not solve the group's financial crisis. The scrapping of Ikarian Reefer was clearly seen in early 1985 as a preferred option to her continued trading. Yet this could have serious repercussions on the group's cash flow. The receipt of US$3m. insurance was greatly to be preferred to US$450,000, which would all go to the bank and trigger further repayments. (18) Notwithstanding the high reputation that the Comninos brothers enjoyed generally, and with their bank in particular, the evidence shows that they were capable of falling short of the standards of honest and respectable businessmen when on occasions it suited them to do so. (19) The Judge's alternative conclusion that, if the ship was scuttled, the owners were not privy to it was strictly academic, since he held that there was nothing to connive at. It cannot stand, in the light of our conclusion that the vessel was deliberately grounded and set on fire. (20) There is no evidence whatever that the crew had any reason to act on their own account or to the prejudice of the owners. Mr Steel's suggestion that the master and crew might have supposed wrongly that the owners would welcome scuttling is, in our judgment, untenable. They would have no knowledge of the insurance or financial position or of the special reason why this particular vessel was of substantially greater value to the owners as a constructive total loss, rather than being sold for scrap. Furthermore, the fact that there is a conspiracy between at least two, and probably three, members of the crew, militates against any conclusion that the scuttling was the work of a disgruntled employee, and points to the complicity of the owners. In our judgment, the overwhelming inference is that the owners authorised the scuttling. (21) Likewise, the Judge's alternative conclusion that if the ship was set on fire deliberately following an accidental grounding, this was an act of barratry, was also academic. For reasons which we have given, the suggestion is quite untenable and was not advanced by Mr Steel. (22) (D) The underwriters' alternative case that the grounding was the proximate cause of the constructive total loss. It is unnecessary in the light of our conclusions that both the grounding and the fire were deliberate, and were done with the privity of the owners, to consider this matter further. We merely record briefly the point. If the owners failed to prove that the grounding was fortuitous and therefore a peril insured against, but the underwriters failed to prove that either the grounding or the fire was deliberate, then success or failure would depend upon whether the proximate cause of the constructive total loss was the grounding or the fire. If the former, the underwriters would succeed; if the latter, the owners. On the pleadings and throughout the trial until Mr Tomlinson's final submissions on Day 74, the underwriters contended that the grounding was the proximate cause of the loss. However, on that day, Mr Tomlinson conceded that it was the fire. We do not understand why this concession was made, since on the owners' pleadings and case, at least so far as the vibration/fatigue and burning rag theories are concerned, these were consequences of the grounding. In these circumstances, Mr Pollock sought leave to withdraw the concession. The Court will normally allow a concession wrongly made to be withdrawn if justice requires it and it can be done without prejudice to the other party. Mr Steel very frankly accepted that there was no prejudice to the owners, since the concession was made at a very late stage. Had it been necessary to do so, we would have been disposed to permit the concession to be withdrawn, though a decision based upon this ground might have had repercussions in costs. In the event, it is unnecessary. (23) Finally, we wish to add this: although we have criticized the Judge's approach and disagreed with his conclusions, it is only right to say that we have considerable sympathy for him. He was overwhelmed by a plethora of expert evidence. Much of it was of little or no assistance; the immense amount of detail and technical complexity can only have deflected him from the essential facts of the case. So far as the navigation experts are concerned, neither appeared to pay any regard to the material evidence of the first officer as to where the vessel was in relation to her track at 19 35. Captain Cockcroft arbitrarily determined that the vessel altered course at 21 00 hours, although the evidence showed that it was nearer 20 20 hours; he also chose a grounding position that suited his theory, rather than any of the direct evidence from Ljubljana or the ship. Captain Third unrealistically appears to have concentrated upon explanations which ignored the effects of current and tidal streams and failed to highlight the question whether the master made any allowance for them, and if so, what. Mr Beatty spent a great deal of time trying to give reasons why the 19 35 and 21 41 satnav fixes were not received. None of his reasons was sustainable. As for the experts in relation to the fire, the owners' experts put forward immensely complex theories to try and afford an innocent explanation of the fire and the open tap. These were based on the most tenuous evidence, or no evidence at all, but rather assumptions. On the other side, Professor Dover spent a great deal of time pursuing the chimera of the heating the valve and pipe theories, long after it should have been apparent that they were untenable. It is clear, and it is much to his credit, that the Judge understood this evidence. But he appears to have become enmeshed in the web of detail to such an extent that he failed to stand back and ask himself whether these explanations and theories were remotely probable or plausible. In the result, the essential simplicity of the case, which was grasped by Dr Bound and Mr Cook on Apr 23, 1985, was overlooked. And the Judge had less opportunity than we have had to see the case in the round. (24) For these reasons, the appeal will be allowed and judgment entered for the defendants. Appeal allowed, with costs awarded in favour of the defendants, here and below, on a standard basis, but without indemnity costs. Trial costs to be backdated to the date of the judgment. Costs of four days of the trial to be deducted after identification of days by the parties. Leave to appeal to the House of Lords refused. DISPOSITION: Appeal allowed SOLICITORS: Ince & Co; Clifford Chance