FIRST DIVISION, INNER HOUSE, COURT OF SESSION
OPINION OF LORD COULSFIELD
in the causes
CALEDONIA NORTH SEA LIMITED
Pursuers and Reclaimers;
LONDON BRIDGE ENGINEERING LIMITED;
PICKUP NO. 7 LIMITED (formerly NORTHERN INDUSTRIAL & MARINE SERVICES COMPANY LIMITED);
BRITISH TELECOMMUNICATIONS plc;
WOOD GROUP ENGINEERING CONTRACTORS LIMITED;
NORTON (NO. 2) LIMITED (In liquidation)(formerly EASTMAN CHRISTENSEN LIMITED) and DAVID JOHN PALLEN, Chartered Accountant, the Liquidator thereof;
KELVIN INTERNATIONAL SERVICES LIMITED (formerly KELVIN CATERING LIMITED); and
COFLEXIP STENA OFFSHORE LIMITED (formerly STENA OFFSHORE LIMITED)
Defenders and Respondents:
Act: MacAulay, Q.C., Batchelor, Q.C., Hofford; Paull & Williamson
Alt: Currie, Q.C., Keen, Q.C., Wolffe; Simpson & Marwick, W.S.
17 December 1999
1. Summary and background.
2. (a) Issues arising under the contracts.
(b) The terms of the contracts.
(c) Summary of contract variations.
3. Construction of the indemnity clauses
(a) Summary of arguments as to the extent of the indemnities.
(a) Summary of arguments in relation to contribution.
5. Consequential loss.
6. Scottish levels of damages.
(a) Effect of a Scottish decree in the settlement process.
(b) Construction of the indemnities in relation to levels of damages.
7. Tax and Interest.
8. General introduction to the factual issues.
(a) Summary of platform lay-out and production process.
(b) Safety procedure.
(c) Reclaimers’ theory of the accident.
(d) Respondents’ objections to the reclaimers’ theory.
9. Novus actus interveniens.
10. The available direct evidence.
11. Vernon’s knowledge.
12. Two particular issues.
(b) The evidence of Grieve.
13. Gas dispersion.
(a) Approach to the factual issues.
(b) Balancing factors.
1. Summary and Background
At about 10.00 p.m. on 6 July 1988, a disastrous explosion and fire occurred on the Piper Alpha Oil Platform located in the North Sea about 110 miles north-east of Aberdeen. There were 226 persons on board the platform, of whom 165 were killed while others were injured. Other persons lost their lives or were injured in the course of rescue operations. At the time of the accident, the platform was under the control of Occidental Petroleum Caledonia Limited (OPCAL) as the operating company on behalf of a consortium concerned in the exploitation of the Piper oil field under the relevant licence. Thirty one of those who were killed on the platform were employees of the operating company; the remainder were employees of one or other of a number of contractors who were engaged, under contracts with the operating company, in the provision of services on board the platform. At the time, OPCAL was one of the Occidental group of companies. Since that time, it has ceased to be part of the Occidental group and become attached to the Elf group of companies and has changed its name to that under which it appears in the present actions.
Following the disaster, a large number of claims were made against OPCAL and were eventually settled by them, the costs being met by the insurers of OPCAL. Many of these claims arose from the deaths of or injuries to employees of contractors. Subsequently, actions were raised by the pursuers and reclaimers against the contractors with the object of recovering from them the sums paid out in settlement of the claims relating to contractors’ employees. There are a large number of such actions. The seven actions with which this reclaiming motion is concerned have been selected as representative, it being understood that the decision of the issues in these actions should govern the remaining cases. The actions were sent to proof, which occupied 391 days before the Lord Ordinary. On 2 September 1997, the Lord Ordinary issued his decision, the effect of which was that the pursuers failed in respect of all their claims, except for one relatively minor claim in respect of an injury to a contractors’ employee which did not fall under OPCAL’s insurance arrangements. In the course of his decision, the Lord Ordinary considered and made findings upon a large number of factual issues, many of a complex and technical nature. The pursuers have reclaimed against the Lord Ordinary’s decision against them, which was based on issues of law. The defenders have marked a cross reclaiming motion in which they challenge the Lord Ordinary’s decision on questions of fact, as well as on some issues of law.
The details of the membership of the consortium, or joint venture, and of the relationship between the consortium and the operating company are set out in the Lord Ordinary’s opinion and, like a great many matters of detail which will be referred to in the course of this opinion, do not need to be repeated. It is a requirement of the legislation applicable in the North Sea that each platform should be under the control and direction of a single operator, who bears responsibility for everything that happens on the platform. It was apparent to everyone concerned, from the moment that the disaster occurred, that OPCAL would have to carry responsibility towards the families of those who had been killed and to those who had suffered injury as a result of the disaster. Almost immediately after the disaster, OPCAL instructed Scottish solicitors to act on their behalf in respect of the claims which would obviously follow. Very shortly after the disaster, it appeared that lawyers from the United States were interested in undertaking the pursuit of claims and were looking to encourage the families of the persons who had been killed to instruct them to make their claims there, in particular in Texas. OPCAL’s advisers took the view that they should encourage claimants to make their claims in Scotland and in that approach they were supported by Scottish solicitors and by the Law Society of Scotland. As a result, before the end of July 1988, a group of claimants and their solicitors had been organised, under the name of the Piper Alpha Disaster Group (PADG), which eventually represented the great majority of the claimants. There were three other small groupings of claimants, all represented by solicitors in the United Kingdom. Only one claimant, a person who had suffered injury in the disaster, actually made a claim in Texas. However, the legal representatives of the claimants in the United Kingdom were well aware of the possibility that claims might be pursued in Texas and that such claims might, if successful, lead to awards of damages considerably higher than those available in the United Kingdom. That possibility figured very largely in negotiations carried out between a steering group on behalf of the PADG and representatives of OPCAL and the other members of the consortium and their insurers. From an early stage, the steering group made it clear that they would not be interested in offers in settlement of their claims based only on Scottish levels of damages. In the process of negotiation, the parties selected a number of sample cases and began by agreeing what would be the Scottish levels of damages appropriate to those claims. They then proceeded to negotiate factors which should be applied to increase the Scottish levels of damages to a level which might be acceptable to the claimants, always bearing in mind the possibility that claims might have been made in the United States. By the end of 1988, it had been agreed in broad terms that a multiplier of 12 should be applied to the Scottish levels of damages for loss of society and a multiplier of 2.5 to the Scottish levels of patrimonial loss. There were a number of other matters, such as the amounts payable to the parents of deceased victims, which also required to be settled. Multipliers were also agreed to be applied to the levels of solatium payable to survivors who had suffered injuries. An elaborate procedure was set up for the verification of claims, for the settlement of disputed issues and for the form in which settlements were to be given effect. These arrangements included provision for arbitration in disputed cases. It was also part of the arrangements that actions should be raised in the Scottish courts and that decrees should be obtained for the sums agreed or settled by the arbitration process, that payment should be made on such decrees and that full discharges should be granted by the claimants. The purpose of these arrangements was to endeavour to ensure that the settlements should be final and should be not capable of being re-opened in any jurisdiction. In view of the importance of the threat of proceedings in the United States, and in particular in Texas, in the negotiations, it is relevant to mention here that one of the issues discussed at length in the proof and the submissions before the Lord Ordinary was whether the claimants could have succeeded in obtaining jurisdiction in Texas. The Lord Ordinary held that they could have succeeded, and that decision was not challenged in the reclaiming motion; a question does, however, remain as to whether, in view of the reasons which the Lord Ordinary had for finding that jurisdiction could have been established in Texas, enhanced levels of compensation are recoverable in these proceedings.
Steps were taken to keep the contractors and their insurers informed as to the progress of the negotiations and the proposals to settle the claims. In particular, it was explained at a meeting on 2 September 1988, at which all the contractors involved in these actions were represented, that the representatives of OPCAL and the consortium considered that they would have to offer sums substantially above Scottish values to obtain a settlement. The contractors reserved their position. Correspondence followed in which attempts were made by OPCAL to obtain the support of the contractors to a settlement. The details of the correspondence and certain meetings in that connection are set out by the Lord Ordinary. The discussions were conducted under some pressure, because of the anxiety of OPCAL and the consortium to forestall proceedings in the United States, and, at one stage at least, there was what was described as a heated exchange. I think that what took place can fairly be summarised by saying that the contractors did not seriously or positively suggest that it would be unreasonable for OPCAL and the consortium to settle at levels above those of Scottish damages, but that they did not commit themselves to accepting the settlement proposals and reserved all their rights to contest any claims made against them. As the settlement process proceeded, details of the settlements made were sent to the legal representatives of the contractors who were invited to contribute to the settlements but declined to do so.
In the actions, the pursuers and reclaimers maintained that the initial explosion which caused the disaster occurred as result of a release of hydrocarbons from the site of a pressure safety valve in a section of the platform known as module C. The process on the platform, which all parties agreed was extremely hazardous, involved, firstly, the separation of hydrocarbons coming onto the platform from undersea wells into oil and gas fractions. The oil was dispatched through a main oil line towards the Flotta Terminal. The gas underwent further treatment, the purpose of which was to recover further hydrocarbons from it in liquid form: the liquid so recovered was referred to as condensate. The condensate eventually went to two condensate injection pumps, situated on the deck known as the 68-foot level, under module C. These pumps raised the pressure to a level at which the condensate could be injected into the main oil line. Each of the two condensate injection pumps was equipped with a pressure safety valve. On the day preceding the disaster, one of the condensate injection pumps, known as pump A, was not in operation. Employees of one of the firms of contractors, Score (U.K.) Ltd., who were engaged to recalibrate valves, had taken advantage of the opportunity to remove the pressure safety valve for recalibration. The gap left by the removal of the pressure safety valve had been closed by fixing blind flanges to the pipework. That work was done by one of the Score employees named Terence Sutton. The valve could not be refitted on that day. At about 9.30 p.m., the condensate injection pump which had been in operation up to that point, pump B, tripped and could not be restarted. The reclaimers’ case was that the lead production operator, Robert Vernon, decided to restart the condensate injection pump which had been out of service, in ignorance of the fact that the pressure safety valve had been removed and not replaced. In order to do so, the reclaimers allege, he introduced hydrocarbon under pressure into the pump. There was then an escape of hydrocarbon from a blind flange, because Sutton had not fitted the flange properly. It was not disputed by the reclaimers that that narrative involved negligence on the part of Vernon in failing to keep himself informed as to the work which had been done and perhaps in other respects, but they maintained that the accident was also caused by negligence on the part of Sutton in failing to fit the blind flange properly. Both Sutton and Vernon were killed in the disaster.
The reclaimers’ theory of how the accident happened was in substance accepted by the Lord Ordinary. In the cross reclaiming motion, the respondents maintain that the Lord Ordinary erred in finding the reclaimers’ theory of the accident established, on three main grounds and a number of subsidiary ones. The first of the three main grounds is that Vernon must have known that the PSV had been removed and not replaced and that, with that knowledge, he would not have repressurised condensate injection pump A: the respondents maintain, as an alternative, that if Vernon did repressurise the pump with that knowledge, the accident was due to wilful misconduct on this part. The second main ground is that there was no basis for holding that Sutton had acted negligently. The third is that the evidence available about the behaviour of condensate released from pressure, in the manner that the reclaimers’ theory requires, did not fit with the evidence about what happened on the platform at the time of the disaster. The respondents also criticise the Lord Ordinary’s approach to the evidence and maintain that the evidence available was circumstantial and fragmentary and could not justify drawing the conclusions which the reclaimers sought and which the Lord Ordinary accepted; and there are a number of other grounds on which they maintain that the reclaimers’ case must fail, which can only be explained after some more detailed consideration of the evidence.
The reclaimers’ claims against the respondents are founded on indemnities granted in the various contracts. One of the provisions common to the contracts is that a party is not entitled to be indemnified if the event giving rise to the claim was caused by the sole negligence or wilful misconduct of that party (or an employee of that party). It is for that reason that the issue whether Vernon was guilty of wilful misconduct. The respondents maintained that the indemnities granted by the contractors only came into operation in the event of the contractors themselves being negligent, it being well established and agreed that none of the contractors (leaving aside Score) was negligent. On that issue, the Lord Ordinary held in favour of the reclaimers. The respondents further maintained that the reclaimers had been reimbursed by their insurers for the claims made against them and that the indemnity contracts were contracts of indemnity and indemnity only; as a result, they fell to be approached in the same way as insurance contracts. From this, the respondents drew the conclusion that the contractors were not liable to reimburse the reclaimers but were liable, if at all, only in a claim for contribution by the insurers who had paid out to the reclaimers. On that issue, the Lord Ordinary held in favour of the respondents. Further issues which were raised under the contracts were whether in terms of the indemnities anything more than the Scottish level of damages was recoverable and whether the claim for recovery was excluded by certain provisions related to consequential loss. On these issues also the Lord Ordinary decided in favour of the respondents. The terms of the contracts and the issues arising under them are set out in more detail in the next section of this opinion. The result of the Lord Ordinary’s conclusions was that, although the reclaimers had established their case as to how the disaster occurred, the respondents were successful in all the actions except one, and decrees of absolvitor were pronounced. The one exception was a case in which the claim, which was at the instance of an employee injured in the disaster, was, through an oversight, not covered by the reclaimers’ insurance arrangements, a situation referred to as the "Oxy gap". In that case, the Lord Ordinary granted decree for a modest sum.
2.(a) Issues arising under the contracts
In each of the seven test cases, the contract between the reclaimers and the contractor included indemnities given by each party to the other and provisions limiting liability for what may briefly be described as consequential loss. There are four major issues and a number of subsidiary issues in relation to each of these contracts. The first major issue is whether the indemnities granted by the contractors in favour of the reclaimers come into operation only in the event that the contractor has been guilty of some fault, negligence or breach of statutory or other duty: this issue is critical because it is common ground that none of the contractors (apart from Score) was guilty of any such fault or breach of duty. This issue can be referred to as the issue of construction of the indemnities. The second issue is whether the indemnities granted by the contractors are to be considered as "contracts of indemnity and indemnity only" with the consequence that they fall to be treated as being on an equal footing with the insurance contracts entered into by the reclaimers: since the reclaimers have been reimbursed by their insurers for the claims made against them, apart from the limited class falling into the Oxy gap, it is contended that the contractors are not liable to reimburse anything to the reclaimers but that, if they are liable at all, they are only liable to a claim for a contribution by the insurers who have paid out the claims made against the reclaimers. This issue is conveniently referred to as the contribution question. For reasons which will be developed later, the first and second issues seem to me to be closely related. The third issue is whether, on the assumption that the reclaimers can recover something under the indemnity provisions, the amount actually paid out by the reclaimers’ insurers cannot be recovered, at least in whole, because such recovery is excluded by certain contract provisions excluding liability for consequential loss. This issue is referred to, therefore as the consequential loss question. The fourth major issue is whether, in any event, the indemnities were framed with reference to Scottish levels of damages and therefore anything beyond Scottish levels of damages cannot be recovered. Again, the third and fourth issues are to some extent related. In order to consider all these issues, and the further subsidiary issues, it is necessary to have in mind the whole relevant provisions of the various contracts. In many respects, the provisions of the contracts are identical or very similar but there are differences between some of the contractual provisions and these differences are at least capable of affecting the result. Cumbersome as it may be, therefore, it seems to me necessary to set out in detail the nature of the contracts in the seven test cases and the provisions relevant to the contribution question. The terms relevant to the consequential loss issue and their variations are set out in section 5 of this opinion. In setting out the contracts, it is convenient to begin with those which raise the first major issue in its starkest and, in a sense, simplest form.
(b) The terms of the Contracts
London Bridge Engineering Limited
The London Bridge contract was a contract by which the contractor agreed to supply a safety operator to work on the Piper Alpha platform during the whole of 1988. The scope of the work included in the contract is defined as the carrying out of the work of a safety officer under the direction of the reclaimers. It appears that in order to perform the contract the contractors had three employees of whom only one would normally be on the platform at any given time. The contract value was in the region of £150,000.
The relevant provisions are found in clauses 17, 18 and 20. Clause 17 is headed "Indemnities" and is divided into three sub-clauses. Clause 17.1 is headed "Contractor’s Indemnities" and provides:
"The Contractor shall indemnify, hold harmless and defend the Company and its parent, subsidiary and affiliate corporations and Participants, and their respective officers, employees, agents and representatives from and against any claim, demand, cause of action, loss, expense or liability (including but not limited to the costs of litigation) arising (whether before or after completion of the Work hereunder) by reason of:
a. Non-compliance with Laws
Claims by governmental authorities or others of any actual or asserted failure of the Contractor to comply with any law, ordinance, regulation, rule or order of any governmental or judicial body; and
b. Intellectual Property Infringement
(Including Patents and Copyrights)
Actual or asserted infringement or improper appropriation or use by the Company, Participants or Contractor of trade secrets, proprietary information, know-how, copyright rights (both statutory and non-statutory) or patented or unpatented inventions or for actual or alleged unauthorised imitation of the work of others, arising out of the use of methods, processes, designs, information or other things originating with the Contractor, its employees, agents, vendors or sub-contractors, and furnished or communicated to the Company by the Contractor or used by the Contractor in connection with performance of the Work and which have not been specified by the Company; and
c. Injury to Employees and Damage to Property of Contractor
Injury to or death of persons employed by or damage to or loss or destruction of property of the Contractor or its parent, subsidiary or affiliate corporations, or the Contractor’s agents, sub-contractors or suppliers, irrespective of any contributory negligence, whether active or passive, of the party to be indemnified, unless such injury, death, damage, loss or destruction was caused by the sole negligence or wilful misconduct of the party which would otherwise be indemnified; and
d. Third Party Injury and Property Damage
Injury, death or property damage, loss or destruction (up to a limit of Five Hundred Thousand Sterling (£500,000) in respect of performance of this contract at an on-shore location and up to a limit of One Million pounds (£1,000,000) in respect of performance at an off-shore location) other than such as is described in Article 17.1(c) and 17.2(a) and b and arising directly or indirectly out of the acts or omissions of the Contractor or its sub-contractors, suppliers or their respective employees or agents, irrespective of any contributory negligence, whether active or passive, of the party to be indemnified, unless such injury, death, damage, loss or destruction was caused by the sole negligence or wilful misconduct of the party which would otherwise be indemnified; and
Waste, debris, rubbish, liquid or non-liquid discharge or pollution of whatever nature which is dropped, seeped, discharged, spilled, blown out or leaked from equipment, apparatus, machinery, facilities or other property of the Contractor or its sub-contractors, suppliers, employees or agents, irrespective of any contributory negligence, whether active or passive, of the party to be indemnified except to the extent that the foregoing was caused by the sole negligence or wilful misconduct of the party which would otherwise be indemnified".
Paragraph 17.2 is headed "Company’s Indemnities" and provides:
"The Company shall indemnify, hold harmless and defend the Contractor and its parent, subsidiary and affiliate corporations and their respective officers, employees, agents and representatives from and against any claim, demand, cause of action, loss, expense or liability (including but not limited to the cost of litigation) arising (whether before or after completion of the Work hereunder) in relation to this Contract by reason of:-
a. Injury to Employees and Damage to Property of Company
Injury to or death of persons employed by or damage to or loss or destruction of property of the Company, Participants or their respective parent, subsidiary or affiliate corporations, irrespective of any contributory negligence, whether active or passive, of the party to be indemnified, unless caused by the sole negligence or wilful misconduct (in the case of injury or death) or wilful misconduct (in the case of property damage, loss, or destruction) of the party which would otherwise be indemnified.
Waste, debris, rubbish, liquid or non-liquid discharge or pollution of whatever nature which is seeped, discharged, spilled, blown out or leaked from any underground reservoir or underwater pipeline or from any cargo, barge, vessel or other equipment or facility of the Company, including fuel, lubricant or the like, irrespective of any contributory negligence, whether active or passive, of the party to be indemnified, except to the extent that the foregoing was caused by the wilful misconduct of the party which would otherwise be indemnified".
Clause 17.3 provides:
"For the purpose of this Article ‘wilful misconduct’ shall mean an intentional and conscious disregard for:
a. good and prudent practices normally associated with the type of
operations envisaged herein, or
b. of any of the terms of this Contract, not justified by any special
circumstances, but shall not include any error of judgment or mistake made either in acting or failing to act by any director, officer, employee, agent, contractor or sub-contractor of the party to be indemnified provided such party acted in good faith".
Clause 18 is headed "Insurance" and clause 18.1 requires the contractor to maintain suitable insurances in force covering the risk defined in clause 18.2. Clause 18.2 provides:
"The following insurance coverage is to be maintained by the Contractor:-
i Employer’s Liability and Workers Compensation Insurance to comply
with the statutory requirements.
ii General Public Liability Insurance in respect of the persons and
property of third parties to the extent of at least £500,000 for each incident arising out of the performance of the Work. In addition, if applicable, the Contractor shall provide automobile insurance to not less than the statutory requirements".
Clause 18.3 deals with the obtaining and exhibiting of certificates of insurance.
Consequential loss is regulated by clause 20 which provides:
"Notwithstanding any provision herein to the contrary, in no event shall either the Contractor or the Company be liable to the other for any indirect or consequential losses suffered, including but not limited to, loss of use, loss of profits, loss of production or business interruption".
Northern Industrial Marine Service Co. Ltd. (now Pickup No. 7 Ltd.)
The period of the contract in this case was something in excess of one month and the scope of the work was the provision of certain plant and an operator of that plant to assist, under the directions of the reclaimers, in installing a particular riser on the platform. Only one employee was involved and the contract sum was £35,000. In this contract the indemnity provisions were found in clause 13 and were identical with the provisions in the London Bridge contract with only two differences. The first, which is of less significance, is that in clause 13.1.d (the contractor’s indemnity for third party injury and property damage) the limit of liability is stated to be £500,000 per occurrence. The second difference is that clause 13.3.2 contains a definition of sole negligence in the following terms:
"‘Sole negligence’ shall mean the exclusive negligence of the party to be indemnified and shall not apply where any other party bears a proportion of the negligence".
The consequential loss provision is found in clause 16 and is identical to the provision in the London Bridge contract.
Stena Off-Shore Limited (now Coflexip Stena Offshore Limited)
This contract covered the period from January 1988 until October 1989 and the scope of the work was the provision of diving services at the platform and elsewhere, including the Flotta Terminal and work from service vessels. The estimated value of the contract was £3,000,000. In this case, the indemnities are found in clause 17 and are identical with those in the London Bridge case except that the limit of liability for third party injury and property damage is stated to be £2,000,000 per occurrence. In addition, there is a definition of sole negligence which is in clause 17.3.c and is identical to that in the Pickup No. 7 contract.
By clause 14.2, the contractor in this case is required to maintain employer’s liability and workers’ compensation insurance to comply with statutory requirements; and general public liability insurance in respect of the persons and property of third parties to the extent of at least £500,000 for each incident arising out of the performance of the work. Clause 14.1 provides that insurance policies to be maintained by the contractor other than workers’ compensation and employer’s liability policies must include company and participants as additional insureds. It also provides that the policy will indemnify the additional insured against any claims brought by any other of the insureds; and that these insurances should be primary and not contributing with any other insurance available to the company or participants. There is a consequential loss provision, as in the other cases.
Reference was also made to clause 16.03 which is headed "Administration & Communication" and provides inter alia that each party shall give prompt notice to the other of claims relating to the contract and shall not settle any claim for which it holds another party harmless without giving notice to and receiving the consent of the other.
Eastman Christensen Limited (now Norton No. 2 Limited)
The contract between the company and Eastman Christensen covered the period from January 1987 until December 1988 and the estimated value of the contract was £3,300,000. The scope of the work was the provision of personnel and equipment for directional drilling, a specialist and expert service. The indemnities in this contract are contained in clause 15. Clause 15.1 is headed "Contractor’s Indemnities" and provides:
"Contractor shall indemnify, hold harmless and defend the Company and its parent, subsidiary and affiliate corporations and Participants and their respective officers, employees, agents and representatives from and against any and all suits, actions, legal or administrative proceedings, claims, demands, damages, liabilities, interest, costs (including but not limited to the cost of litigation) and expenses of whatsoever kind or nature whether arising before or after completion of the Work hereunder and in any manner directly or indirectly caused, occasioned or contributed to in whole or in part, by reason of omission or negligence whether active or passive of Contractor, or of anyone acting under Contractor’s direction, control or on Contractor’s behalf in connection with or incidental to the work. Provided always that the Contractor’s total liability arising pursuant to this indemnity shall not exceed One Million Pounds (£1,000,000) per occurrence.
Without prejudice to the foregoing generality, the Contractor shall indemnify...".
In argument, the foregoing provision was often referred to as a preamble and this, and the other contracts containing an identical provision were referred to as "long preamble contracts". The description of the provision as a preamble, however, might be taken to imply some view of the relationship between this provision and other provisions in the contract. It is useful to have a way of referring to this type of provision, but it seems to me better to call this a general indemnity.
After the general indemnity, clause 15 sets out the same five heads of indemnity, under headings (a) to (e), as are found in the London Bridge and the preceding contracts. The company’s indemnities are found in clause 15.2 and the general provision and heads (a) and (b) are identical to the terms of the previously considered contracts. Clause 15.2.c provides:
"Loss of or Damage to the Well
The loss of or damage to the well, including subsurface damage or injury to the well in connection where appropriate with the services provided by Contractor under this Contract, except where such loss or damage is due directly or indirectly to the negligence or wilful misconduct of Contractor or its servants, agents or sub-contractors".
Under clause 16, the contractor is required to maintain insurance, in terms substantially the same as those in the previous contracts except that the amount specified for general public liability insurance is $5,000,000 and there is a requirement that there should be professional negligence cover for at least £1,000,000 per occurrence. Once again there is a consequential loss provision, in the terms previously set out.
Wood Group Engineering Contractors Limited
The contract in this case covered the period of three years from November 1987 and the estimated value was £7,000,000. The scope of the work was the provision of personnel for platform support duties in accordance with exhibit C to the contract from which it appears, without going into detail, that personnel would be required for the performance of a number of duties of varying sorts including, for example, electrical and engineering and plumbing services.
The indemnities in this contract are found in clause 15. Clause 15.1 sets out the contractor’s indemnities and is substantially in the same terms as the contract in the Eastman Christensen case except that the words which introduce the second paragraph of the indemnity, which contains the five specific sub-heads, after the general indemnity in the first paragraph are:-
"In addition but without prejudice to the foregoing generality, the contractor shall indemnify...".
The company’s indemnities in the Wood Group case are identical with those in the Eastman Christensen case. The sum required to be insured under the general public liability heading is £5,000,000 for each incident arising out of the performance of the work. The consequential loss provision is again identical with previous contracts.
Kelvin Catering Limited (now Kelvin International Services Limited)
The contract in this case covered the period from 1988 until 1991 and the estimated value was £5,000,000. The scope of the work was the provision of labour and services for catering, housekeeping and laundry work. The indemnities are found in clause 15 and clause 15.1, the contractor’s indemnities, is substantially identical to the clause in the Eastman Christensen case except that under head 15.1.d there is a limit of £2,000,000 imposed on the indemnity in respect of third party injury and property damage. The sum for which general public liability insurance is required to be maintained by the contractor is also £2,000,000. The company’s indemnities are identical with those in the Eastman Christensen and Wood Group cases. The consequential loss provision is again identical.
British Telecommunications plc
The contract sum in this case was £40,000. There was no specified period for the work but since the scope of the work was the supplying, installation and commissioning of replacement telecommunications antennae the period would clearly be limited and was probably in the region of six to eight weeks. The terms of the contract in this case appear to have been based on standard terms of B.T. and are quite different from those in the other cases above set out. The first relevant provision is clause 8 which is headed "Limitation of Liability" and provides, inter alia:
"8.1 B.T. accepts liability for faults in the Equipment and its own negligence to the extent stated in paragraph 7, 8.3, 8.4 and 8.5 but not otherwise. Except where expressly contained in these Conditions B.T. has no obligation, duty or liability in contract, delict (including negligence or breach of statutory duty) or otherwise.
8.4 B.T. does not exclude or restrict liability for death or personal injury resulting from its own negligence.
8.5 In any event B.T’s liability in contract, delict (including negligence or breach of statutory duty) or otherwise arising by reason of or in connection with this Contract or howsoever otherwise shall be limited to £1,000,000.
8.6 In any event in no circumstances shall B.T. be liable in contract, delict (including negligence or breach of statutory duty) or otherwise for loss (whether direct or indirect) of profits, business or anticipated savings, or for any indirect or consequential loss or damage whatever".
Clause 8.7 provides that each provision of paragraph 8 is to be construed as a separate limitation. Paragraph 7, to which reference is made in paragraph 8.1, is concerned only with responsibility for faults in equipment provided by B.T.
Clause 10 is headed "Indemnity for Off-Shore Work". Clause 10.1 provides that notwithstanding clause 8.1 but subject to the remaining limitations of clause 8 and certain conditions of sale attached, the contents of which are not material, clause 10 is to apply to the work on Piper Alpha. Clause 10.2 is in the following terms:
B.T. shall indemnify, hold harmless and defend the Customer and its parent, and Participants and their respective officers, employees, agents and representatives from and against any and all suits, actions, legal or administrative proceedings, claims, demands, damages, liabilities, interest, cost (including but not limited to the costs of litigation) and expenses of whatsoever nature arising during performance of the work by reason of:-
a. Non-compliance with Laws
Claims by U.K. governmental authorities or others of any actual or asserted failure of B.T. to comply with any law, ordinance, regulation rule or order of the U.K. Government or competent authority; and
b. Injury to Employees and Damage to Property of B.T.
Injury to or death of persons employed by or damage to or loss or destruction of property of B.T. or B.T’s agents, sub-contractors or suppliers irrespective of any contributory negligence whether active or passive, of the party to be indemnified, unless caused by the sole negligence or wilful misconduct (in case of injury or death) or wilful misconduct (in case of property damage, loss or destruction) of the party which would otherwise be indemnified".
Clause 10.3 which is headed "Customer’s Indemnities" provides for indemnities substantially the same as those included under "Contractor’s Indemnities" in the other contracts and clause 10.4 defines wilful misconduct, again in substantially the same terms as those in the other contracts.
2(c) Summary of Contract Variations
The variations in these contractual provisions can be set out, in summary form, as follows:-
Contracts with no general indemnity
1. London Bridge Engineering Ltd.
a. limit of indemnity in subparagraph (d): £500,000 in respect of
performance of the contract at an onshore location and £1 million in respect of performance at an offshore location.
b. public liability insurance obligation: at least £500,000 for each incident
arising out of the performance of the Work.
c. no definition of "sole negligence".
2. Northern Industrial & Marine Services Co. Ltd.
a. limit of indemnity in subparagraph (d): £500,000 per occurrence.
b. public liability insurance obligation: at least £500,000 for each incident
arising out of the performance of the Work.
c. definition of "sole negligence".
d. subparagraph (c) does not refer to property damage.
3. Stena Off-Shore Ltd.
a. limit of indemnity in subparagraph (d): £2 million per occurrence.
b. public liability insurance obligation: a minimum of £2 million in
respect of incidents arising out of the performance of the Work.
c. definition of "sole negligence".
d. additional insurance obligations.
e. clause 16.03.
Contracts which include a general indemnity
4. Eastman Christensen Ltd.
a. "Without prejudice to the foregoing generality...".
b. financial limit in general indemnity: £1 million per occurrence.
c. limit of indemnity in subparagraph (d): none.
d. public liability insurance obligation: at least US $5 million for each
incident arising out of the performance of the Work.
c. no definition of "sole negligence".
5. Wood Group Engineering Contractors Ltd.
a. "In addition but without prejudice to the foregoing generality...".
b. financial limit in general indemnity: £1 million per occurrence.
c. limit of indemnity in subparagraph (d): £5 million [in total].
d. public liability insurance obligation: at least £5 million for each
incident arising out of the performance of the Work.
e. no definition of "sole negligence".
6. Kelvin Catering Ltd.
a. "Without prejudice to the foregoing generality...".
b. financial limit in general indemnity: none.
c. limit of indemnity in subparagraph (d): £2 million [in total].
d. public liability insurance obligation: at least £2 million for each
incident arising out of the performance of the Work.
e. no definition of "sole negligence".
7. British Telecommunications plc
a. no third party indemnity.
b. no insuring obligation.
c. no definition of "sole negligence".
d. introductory paragraph applying limitations set out in clause 8.
3. Construction of the indemnity clauses
(a) Summary of arguments as to the extent of the indemnities
The terms of the contracts in six of the test cases are to a large extent identical or very similar but there are differences between them which may be important. The British Telecom contract contains rather more differences from the other contracts but there are similarities between some of its provisions and those of the other contracts. The remaining six similar contracts all contain provision for indemnities given by the respondents to the reclaimers and by the reclaimers to the respondents and all of them contain provisions excluding liability for consequential loss in identical, or virtually identical, terms. The indemnities given by the reclaimers are also identical except that in two of the cases (the Wood Group and Eastman Christensen contracts) there is an indemnity against damage to the well which is not present in the other contracts. This difference may, perhaps, be explained on the ground that the work comprised in the two contracts which contained that indemnity involved a particular risk of damage to the well through the work of the contractor.
The indemnities granted by the contractors all contain the five individual indemnities, or heads of indemnity, sometimes referred to as the "core" indemnities. These are the indemnities in respect of non-compliance with laws; patents or similar items of intellectual property; injury to or death of employees of the contractor or damage to the contractor’s property; damage to third parties; and pollution. In five of the cases there is a limit on the amount of the contractor’s liability in respect of the indemnity against damage to third parties, varying from £500,000 in the case of the Northern Industrial contract and, in some circumstances, the London Bridge contract, up to £5,000,000 in the case of the Wood Group contract; in the Eastman Christensen case there is no limit on liability. Three of the contracts contain the general indemnity provision preceding the five core indemnities. These are the Wood Group, the Eastman Christensen and the Kelvin Catering contracts. In the case of the Wood Group and Eastman Christensen contracts, there is a limit of £1,000,000 placed on the general indemnity; in the Kelvin Catering case there is no such limit. In the Eastman Christensen contract and the Kelvin Catering contract, the indemnities set out under the five specific heads follow the general indemnity and are introduced by the words "Without prejudice to the foregoing generality"; in the Wood Group case, the words are "In addition but without prejudice to the foregoing generality".
All of the contracts contain a definition of wilful misconduct in identical or virtually identical terms. Only two of the contracts contain a definition of sole negligence, namely the Coflexip Stena and Northern Industrial contracts.
Parties were agreed that, despite the fact that many of the relevant provisions of the contracts are identical, the differences between them were such that it was necessary to look at each contract individually, and that the proper construction would not necessarily be the same in each case, notwithstanding the fact that it is reasonably obvious that most of the provisions must have been used by the reclaimers as standard contract terms. Parties were also agreed that it was not possible to see any development in the pattern of the contracts over time nor any other helpful pattern in the way that the contracts had come to be made. Further, parties did not suggest that the amounts of, for example, the required insurance or the limits of liability imposed in certain provisions of the various contracts were of assistance in their construction, although reference was made, in general terms, to the role of the provisions in regard to insurance as relevant to construction.
Parties also agreed that the question of the proper construction of the contracts must be considered against the background of the environment in which the contractors were expected to operate and, in particular, the regulatory regime under which North Sea operations were carried out. As regards the environment in general, all work carried out on the platform was under the control of the reclaimers as the operators of the platform. The reclaimers did have employees of their own on the platform (at the time of the disaster 37 out of the total of 226) but a great many, possibly the majority, of operations on the platform were carried out by employees of contractors working under the direction of the reclaimers’ supervisory staff. Each of the contractors would therefore know that the work required to be carried out in an environment in which not only employees of the reclaimers but also employees and property of other contractors were liable to be affected. Activities in the North Sea oil industry are closely controlled by regulations made under section 6 of the Mineral Workings (Off-Shore Installations) Act 1971 and these regulations may give rise both to criminal and, in terms of Regulation 11(1), civil liability. Apart from any more detailed provisions, Regulation 32(1) of the Off-Shore Installations (Safety, Health and Welfare) Regulations 1976 provides that it is the duty of the installation manager and of the owner of the installation and of the concession owner to ensure that the provisions of the regulations are complied with in respect of any off-shore installation; and Regulation 32(2) provides that it is the duty of the employer of an employee employed for work on or near an off-shore installation to ensure that the employee complies with any provision of the regulations imposing a duty or a prohibition on him. Further, under Regulation 32(3) it is the duty of every person while on or near an off-shore installation not to do anything likely to endanger the safety or health of himself or other persons on or near the installation or render unsafe any equipment used on or near it; and also to co-operate with his employer in securing compliance with the regulations and to report any defect likely to endanger safety. Consequently, a scheme of strict liability is imposed on operators and employers and also on employees. The extent of that liability is illustrated by the decision in MacMillan v. Wimpey Off-Shore Engineers and Contractors Limited 1991 S.L.T. 515 in which it was held that the responsibility placed on an employer under Regulation 32(2) was an absolute one, imposing not only vicarious liability for the default of an employee but also a direct duty to guarantee compliance with the regulations by employees. It was further held that the duty extended to the protection of workmen at all times while they were on or near the installation, not only while they were actually working, so that the employer might be liable for an assault committed by an employee during a time when he was not at work.
In presenting their argument on this topic, the respondents began with the Eastman Christensen contract, one of those which contains the general indemnity provision. The reclaimers, on the other hand, began from the London Bridge contract, which lacks the general indemnity provision. On both sides, the argument naturally tended to emphasise those words and phrases in the contracts which were most favourable to the broad position of the party advancing them. In addition, in the course of argument, a considerable amount of ingenuity was expended in examining examples of situations which might, with greater or less probability, have been supposed to be in the contemplation of parties and which might illustrate the effects of particular constructions of the contract terms. It would be burdensome to attempt to follow all the twists and turns of the arguments presented and it is doubtful whether to attempt to do so would be of any particular help in resolving the problems which arise. It is sufficient, in my view, to set out in broad terms the general position of the parties on the major issues which separated them.
Generally, the respondents submitted that the indemnities should be construed contra proferentem and therefore against the interests of the reclaimers who were both the persons who stipulated the relevant terms and those who relied upon them. Reference was made to Youell v. Bland Welsh  2 Lloyds L.R. 127 and to Canada Steamship Lines Limited v. The King  A.C. 192. The principles of the latter decision had been accepted in Scotland in North of Scotland Hydro Electric Board v. D. & R. Taylor 1956 S.C. 1 and Smith v. U.M.B. Chrysler (Scotland) Limited 1978 S.C. (H.L.) 1. The latter case decided that the principles which applied in the construction of a clause excluding liability also applied in the case of an indemnity. In both cases it might be said that the original basis of the reasoning which led to the restrictive approach to such clauses was the inherent improbability of a party accepting liability for someone else’s negligence, but the principle of strict construction had been established by a tract of authority and it was not necessary to consider the question of improbability. The same approach could be seen to have been applied in Ailsa Craig Fishing Company Limited v. Malvern Fishing Company Ltd. 1982 S.C. (H.L.) 14, in which Lord Fraser had referred to specially exacting standards, and Bovis Construction (Scotland) Limited v. Whatlings Construction Limited 1995 S.C. (H.L.) 19. The result was that there were rules of construction which must be applied; these rules derived from the Canada Steamship case and, in particular, the third rule mentioned by Lord Morton in that case, the result of which was that negligence could not be covered by an exemption or indemnity clause unless specially mentioned. Even after construing the contract, and deciding that it was intended to deal with the question of negligence, for example where there was specific reference to negligence, it was still necessary to apply a strict construction, so that, if there were any doubt about the construction which the court was inclined to favour that doubt should be resolved against the party relying on the indemnity. Reference was made in this connection to Murray v. Caledonian Crane & Plant Hire Ltd. 1983 S.L.T. 306. In practice, the result was that it was enough for the respondents to be able to point to a sensible construction which the provision in question might be capable of bearing and which was contrary to the submission put forward by the party founding on the clause. If such a construction led to redundancy, that was not an argument against adopting it: see Beaufort Developments (N.I.) Limited v. Gilbert Ash (N.I.) Limited  2 W.L.R. 860.
In response to these general propositions, the reclaimers submitted that the role of the court was to ascertain the intention of the contracting parties from the language which they had used in the light of the surrounding circumstances taken to be in their knowledge at the time of contracting, taking account also of the fact that the indemnities relied on were part of a set of contractual provisions which included indemnities given by each party to the other in respect of the death of or injury to their respective employees. The propositions advanced by the defenders went far beyond the normal rules of contra proferentem and strict construction as they had been discussed in Smith v. U.M.B. Chrysler supra and Ailsa Fishing Company v. Malvern Fishing Company supra. Reference was also made to the The Raphael  2 Lloyds L.R. 42.
The respondents further submitted that it could not be suggested that the approach to construction should be any different because there were reciprocal indemnities between the reclaimers and the contractors. The indemnities were not, in any case or on any view, properly reciprocal. In Nelson v. Atlantic Power & Gas Limited 1995 S.L.T. 102 the view had been taken that contractual provisions under which each party was responsible for injuries to its own employees were reciprocal and should receive effect, but in that case the indemnities had truly been exact counterparts of one another as was clear from the opinion of the Lord Ordinary where the indemnities were set out in full. Similarly, in E.E. Caledonia Ltd. v. Orbit Valve Co. Europe  1 W.L.R. 1515 there was a bilateral provision which was in terms reciprocal; and in any event the decision had been partly affected by a concession noted at page 1522H. In the present case it could not be said that each party was simply accepting liability for its own employees, come what might. The indemnities were inherently qualified by reference to sole negligence and wilful misconduct and the company’s indemnities were qualified by the expression "in relation to this contract" for which there was no equivalent in the contractor’s indemnity. In regard to third party injuries, the contractor granted an indemnity irrespective of any negligence by the operator and the operator granted no such indemnity. There were cases in which the words "in relation to this contract" could make a difference, for example where a contractor’s employee injured an operator’s employee while off duty. In any event, the existence of reciprocal indemnities was not a reason for a different approach to construction. Lord Morton’s tests applied not merely in determining whether negligence was covered by the terms of an indemnity but also in determining the extent to which it was covered.
In reply to this argument, the reclaimers submitted that it was established by the authorities referred to in the respondents’ argument that the fact that there were reciprocal indemnities was a relevant consideration and went to eliminate or reduce the effect of the argument that it was unlikely that any person would accept liability for injuries caused by the negligence of another. The indemnities which were important for the present purposes, those relating to death or injury of employees and injury to property of the contractor and the operator, were reciprocal and this could therefore be a relevant factor in their construction.
As I have mentioned, in dealing with the central issue of the proper approach to the words of the indemnities, the parties approached the question from different ends. The respondents began by concentrating on the cases containing the general indemnity, and, having made submissions in favour of the construction which they put forward in those cases, went on to submit that notwithstanding the absence of the general indemnity, and notwithstanding the presence in certain of the cases of a definition of sole negligence, the same construction fell to be applied in all the contracts. On the other hand, the reclaimers presented their argument first by reference to the contracts without a general indemnity and went on to submit that the same construction should be applied in the cases in which there was a general indemnity. Parties did not approach the matter in these different ways without reason because the argument for the respondents is undoubtedly easier to present in relation to the contracts which contain the general indemnity while that for the reclaimers is easier to present in the contracts which do not. Since, however, each contract is a separate contract it is necessary to keep in mind that each requires to be considered separately and that the order in which the matters are approached should not influence the outcome. It should therefore be a matter of indifference which contract is looked at first and, with that caution in mind, I propose to follow the same order as did the respondents in presenting the cross-appeal.
The main argument was submitted in relation to the Eastman Christensen contract. It was pointed out that the contract envisaged that the contractor would bring his employees onto the operator’s platform to do work in an environment to which the off-shore regulatory regime applied. It was therefore contemplated that any act by an employee which endangered the safety of himself or others could render both the contractor and the operator in breach of duty. The general indemnity covered all claims due to omission or negligence on the part of the contractor or of anyone acting on his behalf and therefore covered all claims in which the operator was made liable by something that was done by the contractor. If the provision stopped there, there would be nothing to apply to any case in which the loss was contributed to by negligence on the part of the operator. The word "omission" in the phrase "omission or negligence" in the general indemnity meant a breach of duty and it must be possible to tie that breach back to something which the contractor had done or failed to do. Further, liability under the general indemnity was limited to the sum of £1,000,000 in that case. The specific heads of indemnity were introduced by the words "without prejudice to the foregoing generality" and therefore subhead (c), like the other subheads, was subject to the general indemnity, so that there was no liability unless the loss was caused by the contractor. In the present case there was no suggestion that any responsibility could attach to any of the contractors. The reclaimers’ construction was that there were two free-standing indemnities, the general indemnity and the specific subhead, while the Lord Ordinary had taken the view that the indemnities were alternative. Neither suggestion was convincing and the best commercial sense was achieved by the respondents’ construction. That would achieve a sensible result and sufficiently account for the indemnities in that it would deal with the case in which both contractor and operator were liable unless there was sole negligence or wilful misconduct. The clause would therefore eliminate any disputes as to apportionment of liability in a case of joint fault. The clause should be taken as a whole and regarded as a single composite indemnity; the specific subheads brought in matters which might have been thought to be at risk of being left out if the general provision stood alone. It was difficult to say what might fall under the general provision but not under one or other of the subheads but it was not necessary for the purposes of the argument to be able to point to any situation which might so fall. The reference to contributory negligence in subhead (c) supported that interpretation. In using that phrase, the draftsman clearly had in mind the rule that there had to be a specific reference to negligence, if negligence was to be covered. The drafting recognised that there was an inherent limitation on the general indemnity in that there was no reference in it to negligence and therefore negligence was not covered, whereas the reference to contributory negligence was introduced specifically in order to show that in the case where there was joint fault the indemnity did extend to cover it. The insurance provisions were consistent with that view. At least three of the subheads, namely heads (a), (b) and (e), supported the view that what was in issue was something due to the action of the contractor. The indemnity in subhead (c) extended not only to cases of death of or injury to employees but also to damage to property and since there was no statutory responsibility for such damage parties must have had negligence claims primarily in view. The reclaimers’ construction would have the effect that the operator could choose whether to go under the general provision, subject to a limitation of liability, or under one of the specific heads and that was not a sensible construction. It was also unreasonable, because the result might be that a contractor who properly should bear only a small share of blame for an accident might end up with the whole liability for its effects; and because a contractor might be minded into thinking that his liability overall was limited to the sum stated in the general indemnity.
The reclaimers submitted that the general indemnity was simply an additional free-standing indemnity and did not control the following paragraphs. It dealt with a liability which would fall on the contractor in any event, being one brought about by his omission or negligence, but imposed a limit in respect of it of £1,000,000. The clause, in each contract, was headed "Indemnities", whereas the general indemnity began with a reference to "This Indemnity" and the opening phrase was later repeated, all of which suggested that there were separate and independent indemnities. The general indemnity was wider than the later indemnities in that it included reference to administrative proceedings. The phrase "without prejudice to the foregoing generality" did not restrict the use of language later in the clause, although the use of the word "generality" was perhaps unfortunate. The respondents’ construction, according to which all the subheads were subheads of, and must fall within, the general indemnity involved construing the word "omission" as covering a breach of statutory duty. On their reading also, the wording of the general indemnity sat unhappily with the other individual heads, apart from head (c), since the word "omission" had to be read as covering breach of law or breach of copyright. On the respondents’ construction, there was a tautology because the expression "contributory negligence" in head (c) was redundant since negligence was already implied as a result of the general indemnity. In the last resort, the reclaimers would argue that the scheme of the clause as a whole made it necessary to read into the general indemnity a limitation to the effect that it applied only where the individual heads later set out did not or, in any event, that, looking at the clause as a whole, there was no limitation on the scope of the particular subheads. Even if the result was that in certain circumstances the operator had an option whether to proceed under the general indemnity or a particular heading, that simply must be accepted. There was no substance in the contention that the effect of the reclaimers’ construction was that a contractor might be misled into thinking that the limitation in liability extended to the individual subheads. Equally, there was no substance in the contention that the reclaimers’ construction was unreasonable in that it had the effect that a contractor might be liable for the whole of a loss in a question with the operator even though that loss had been caused very largely by negligence on the part of the operator and merely contributed to by negligence by a third party.
In relation to the contracts without a general indemnity, the respondents’ argument accepted that, ex facie, the liabilities set out in heads (a) to (e) appeared to be independent and unlimited. However, head (c) in all the contracts included a reference to contributory negligence and it was necessary to ask the question, contributory to what? The only possible significance of those words was that what was intended was negligence by the operator contributory to negligence or fault on the part of the contractor. All the other arguments advanced in relation to the nature of the bargain and the reasonable or unreasonable effects of the different constructions put forward were equally applicable to the case of the contracts without a general indemnity and the construction should therefore be the same.
In reply to that submission, the reclaimers submitted that there was no warrant for reading the plain words of the indemnity set out in head (c) in any restricted sense. The reference to contributory negligence was clearly a reference to negligence on the part of the party to be indemnified but did not carry the implication that negligence on the part of the indemnifier was a precondition of the application of the indemnity. The words were used in a causative sense rather than in a sense implying fault or responsibility. The respondents’ construction involved using the expression "negligence" in different senses within the same clause. It also involved tautology.
The remaining arguments advanced on each side were special to the particular terms of particular contracts. In those cases in which there was a definition of sole negligence, it was submitted by the reclaimers that the definition clearly pointed to the intention that the indemnity should operate where there was negligence on the part of the operator as well as of a third party and that the exclusion of the operation of the indemnity should only apply where there was sole negligence on the part of the operator. It was further submitted on behalf of the reclaimers that in the case in which the connecting phrase between the general and the specific indemnities was "in addition and without prejudice", it was clear that the respondents’ construction could not apply. The respondents submitted that notwithstanding these specialties, the general construction which they put forward had to be accepted, in line with the general rule of construction contra proferentem.
Finally, in this connection, it was submitted on behalf of the reclaimers that there was material to indicate that the construction which they proposed was in accordance with a general or common practice in the North Sea oil industry whereby each party undertook the responsibility for insuring against the death of or injury to their own employees and reference was made to textbook discussions of that practice.
Decision of the Lord Ordinary
The Lord Ordinary dealt with this issue in volume 5 of his Opinion, particularly at pp. 927H, 939H and at pp. 1102-1116. He concluded that the five core indemnities should be considered as separate and independent indemnities and were not confined in meaning by the general indemnity or "long preamble". On that basis, he held that there were no implications that the contractor was only liable under the indemnities where he had been guilty of fault in some respect.
The reclaimers’ claim, in each of the cases other than the B.T. contract, is founded upon head (c) of the clause of indemnity, that is, the clause by which the contractor undertakes to indemnify the company against injury to or death of persons employed, or damage to or loss or destruction of property of, the contractor. In all the six cases other than the B.T. case, that paragraph, and indeed all the individual heads of indemnity, are identical and it is clear that these were standard terms put forward by the reclaimers for inclusion in the contract. It is also the reclaimers who found upon the clauses and therefore the conditions for the application of the principle of construction contra proferentem are present. In the course of the argument, a number of authorities bearing on the proper application of that principle were canvassed in detail. In my view, however, it would not be useful to attempt to analyse all those authorities in the way attempted in argument. As I understand the position, the principle is one which dictates an approach to the reading of words in a contract and, just as there are no absolute rules for the construction of an English sentence which can be laid down as applicable in all possible circumstances, so, it seems to me, the precise formulation of the contra proferentem principle is liable to vary with the subject matter. It might, indeed, be questioned why the principle is applied at all in the case of commercial contracts entered into between parties dealing on a business, and for all that is known, on an equal footing but the authorities that the principle does apply are clear and binding. In the circumstances, it seems to me sufficient to bear in mind the enunciations of the principle by Lord Morton in Canada Steamship Lines Limited v. The King  A.C. 192, because of the particular bearing of that decision on the question of liability for negligence, and by Lord Keith in Smith v. U.M.B. Chrysler (Scotland) Limited 1978 S.C. (H.L.) 1. Lord Morton said, at page 208,
"Their Lordships think that the duty of a court in approaching the consideration of such clauses may be summarised as follows:-
(1) If the clause contains language which expressly exempts the person
in whose favour it is made (hereafter called ‘the proferens’) from the consequence of the negligence of his own servants, effect must be given to that provision...
(2) If there is no express reference to negligence, the court must consider
whether the words used are wide enough, in their ordinary meaning, to cover negligence on the part of the servants of the proferens. If a doubt arises at this point, it must be resolved against the proferens...
(3) If the words used are wide enough for the above purpose, the court
must then consider whether ‘the head of damage may be based on some ground other than that of negligence’ to quote again Lord Greene in the Alderslade case. The ‘other ground’ must not be so fanciful or remote that the proferens cannot be supposed to have desired protection against it; but subject to this qualification, which is no doubt to be implied from Lord Greene’s words, the existence of a possible head of damage other than that of negligence is fatal to the proferens even if the words used are prima facie wide enough to cover negligence on the part of his servants".
The Canada Steamship Lines case was a case which dealt with an exemption clause but it was made clear by the decision in Smith v. U.M.B. Chrysler that the same principles apply to an indemnity clause. In that case, Lord Keith of Kinkel said (at pp. 16-17):
"In North of Scotland Hydro Electric Board v. D. & R. Taylor 1956 S.C. 1 the Second Division accepted that these principles (Lord Morton’s principles) applied in the law of Scotland. It is to be stressed that they do not represent rules of law, but simply particular applications of wider general principles of construction, the rule that express language must receive due effect and the rule omnia praesumuntur contra proferentem. While they apply to the construction both of a clause bearing to exempt from certain liabilities a party who has undertaken to carry out contractual work and of a clause whereby such a party has agreed to indemnify the other party against liabilities which would ordinarily fall upon him, they apply a fortiori in the latter case, since it represents a less usual and more extreme situation".
There was some discussion as to whether Lord Keith’s use of the expression a fortiori implied that the approach to construction must be even more strict in a case of an indemnity than in other cases falling under the rule or whether Lord Keith simply meant that it was more clear in such a case that the rule applied. I am inclined to prefer the latter reading, since I think it would be difficult to formulate different degrees of strictness in the application of such a rule of construction, but I do not think that, for the present purpose, it is necessary to resolve that difference. I am prepared to accept that the indemnity clauses in the present case should be examined very strictly.
With regard to the general argument on matters of law, it is only necessary to add that the reclaimers submitted that the fact that there were reciprocal indemnities between the company and the contractors was relevant to the construction and referred to Nelson v. Atlantic Power & Gas Limited 1995 S.L.T. 102 while the respondents submitted that that fact made no difference to the proper approach to construction, founding on what was said by Steyn L.J. in E.E. Caledonia Limited v. Orbit Valve Co. Europe  1 W.L.R. 1515. Again, I do not think it necessary for the present purpose to go into the details of these cases and I am content to proceed on the basis that the fact that there are cross indemnities is irrelevant for the purpose of construction of the indemnity on which the reclaimers found.
In this case it is necessary to consider a number of contracts, the provisions of which vary to a considerable degree, although all except the B.T. contract contain the same five "core" indemnities. As a result, it is not easy to know what is the best way to approach the problems. It is obvious that the variations in the terms of the contracts might lead to different results in the different cases, even though, as I have mentioned, it is fairly obvious that the core indemnities must have been standard indemnities put forward by the reclaimers. As has been mentioned, the parties started from different points in their arguments and did so for obvious reasons. There is, therefore, a danger that the choice of the contract with which to start may influence the result at which one arrives. However, it is necessary to start somewhere and the same risk arises, one way or another, whatever starting point is chosen. I would say, in passing, that both parties accepted that it was necessary to consider each contract individually and that there was at least a possibility that different conclusions might be arrived at with regard to different contracts. That might seem paradoxical in the circumstances, but there is no alternative to attempting to consider each contract individually.
I propose to start with the London Bridge contract because it is the simplest and has the fewest provisions to be considered: it also presents sharply the question of the construction of head (c) of the core indemnities. The London Bridge contract contains a straightforward clause setting out the five indemnities with no general indemnity clause and no definitions. As has been remarked earlier, the argument followed a course full of twists and turns as attempts were made to give a particular slant or a particular weight to different phrasing in different parts of the contract or in different contracts. Having attempted to consider and keep in mind all these arguments, it nevertheless seems to me that in the London Bridge case it is possible to stand back from the detailed arguments and look at what head (c), taken by itself, says. So considered, it appears to me that head (c) says, in straightforward terms, that the contractor will indemnify the company against any claim, demand, cause of action, loss, expense or liability arising by reason of injury to or death of persons employed by the contractor irrespective of any contributory negligence of the company, unless the injury was caused by the sole negligence or wilful misconduct of the company. So looked at, it seems to me that the paragraph says, simply and directly, that if the contractor’s employees are injured or killed, the contractor will indemnify the company for any liability in that respect unless the exception for sole negligence or wilful default applies. I find it difficult to see that it can really be suggested that there is any other meaning for the words or that there is any form of responsibility or liability that could be imagined to exist which would give room for the application of Lord Morton’s third principle of construction. This may, as Steyn L.J. said in E.E. Caledonia supra, be a matter of impression not open to lengthy argument. There are, however, a number of considerations which seem to me to support that reading of the clause and there is one particular argument put forward by the respondents which does require to be dealt with.
The first point is that it is plain that the paragraph was designed to deal with liability for negligence. It expressly does so in two ways, by reference to contributory negligence and by reference to sole negligence. The second point is that it is equally clear that there are five core indemnities all of which are separate and all of which deal, differently, with different matters. The third point is that it is clear that the intention of parties was that the company should only lose the benefit of the indemnity if there was sole negligence or wilful misconduct on its part. That being so, the only issue which has to be further dealt with, in my view, is the argument for the respondents, to which I earlier referred. That argument comes to be whether the use of the word "contributory" in the phrase "irrespective of any contributory negligence on the part of the party to be indemnified" carries the implication that the negligence referred to must be contributory to negligence on the part of the contractor; or, in other words, that the clause can only apply where the contractor has been shown to have been negligent. It is true that the expression "contributory negligence" is commonly used in a legal context to refer to negligence on the part of a claimant which is contributory to negligence on the part of the person against whom the claim is made. It is, therefore, possible to argue that there is an implication that the person indemnifying under paragraph (c) may have been assumed to have been negligent. In my view, however, the context in which the phrase is found displaces any such inference. Looking to the whole of the paragraph, it seems to me that "contributory" is used in contradistinction to "sole" and that what the drafter of the contract had in mind was to make it clear that only sole negligence on the part of the party to be indemnified would defeat the indemnity, not negligence which was only a partial cause of the loss or damage.
I propose to consider next the Eastman Christensen contract. That is the contract in which there is a general indemnity, containing a limitation of liability to £1m per occurrence, and in which the general indemnity is connected to the five core indemnities by the words "without prejudice to the foregoing generality". If in this case also one attempts to consider the meaning of the clause as a whole, the first point which comes to mind is that the words "without prejudice to the foregoing generality" are used in an unusual sense, or at least an unusual context. Normally, the point of including such words between a general provision and a series of detailed provisions is to try to ensure that the construction of the general provision is not limited by reference to the details which follow. In the present case, however, the general indemnity is limited in two respects, by reference to omission or negligence and by reference to the limit of liability, while the heads which follow are not, at least on their own terms, subject to limitation in the same way. What the respondents’ argument seeks to do is to import the limitations which were contained in the general indemnity into the indemnities which succeed the words and, in my view, that is not consistent with what the words on their face normally mean. Looking at the words themselves, therefore, without regard to the substance of the contract or its possible effects, I would be very much inclined to say that the expression "without prejudice to the foregoing generality" was nothing but a connecting link and gave rise to no inference or restriction of meaning.
There is, however, another important factor to be taken into account and that is the effect of the clause as a whole, if the reclaimers’ construction of it is adopted. If the reclaimers are correct, then if the respondents are negligent and thereby cause loss they are liable to indemnify the company but subject to a limitation of £1m. If, on the other hand, the reclaimers come under a liability to employees of the contractors, the contractors must indemnify the reclaimers without any limit, even if they have not been negligent. In this context, it is relevant to add that while the contractors are required to insure to cover liability towards third parties, in terms of head (d) of the indemnities, there is no requirement for insurance to cover the obligation to indemnify in respect of injuries to the contractor’s own employees or property. There is, of course, an obligation to maintain employer’s liability insurance but that would not cover the risk of having to indemnify the company. There is, therefore, it seems to me, quite a powerful argument that the contract as construed by the reclaimers does not make business or commercial sense.
There are, however, also difficulties in the respondents’ proposed construction of the contract. What the respondents suggest is that the word "negligence" as it occurs in head (c) of the core indemnities means something like common law negligence. In a situation in which there may be statutory liability on one or both of the parties, irrespective of any fault, arising from injury to or death of an employee, it might make quite good commercial sense to agree that, as between the parties, ultimate liability would depend on the presence or absence of common law negligence. However, the respondents were constrained to accept that if, as they contend, the individual indemnities are subordinate to the general preamble, then the arrangement could only work if the general indemnity were construed as covering not only common law negligence but also statutory liability. It follows, therefore, that the phrase "omission or negligence" in the general indemnity must cover statutory liability. If so, the word "negligence" is used in different senses in the general indemnity and in head (c) of the particular indemnities.
I have not found this an easy question. I have been attracted to the argument that the wording of the contract, considered by itself, does not clearly favour either the reclaimers’ or the respondents’ position, and therefore that the strict application of the contra proferentem principle would require that the respondents’ construction should be preferred. In the end, however, I have come to the conclusion that the words of head (c) in themselves are clear, that head (c) is the central and predominant and that the linking phrase is not apt to subordinate them to the general indemnity. Further, in my opinion, the respondents’ argument as to commercial sense does not carry sufficient weight, even coupled with the contra proferentem principle, to require a different result.
Having reached these conclusions about the Eastman Christensen and London Bridge Contracts, it seems to me that the remaining contracts fall into place quite simply. The Kelvin Catering contract contains the words "without prejudice to the foregoing generality" but has no financial limit stated in the preamble. The result is the same as in the Eastman Christensen case. The Wood Group Engineering contract does contain a financial limit in the general indemnity but the words connecting it to the particular indemnities are "in addition but without prejudice to the foregoing generality". Those words are, in my view, favourable to the reclaimers, and suggest that the particular indemnities are to be looked at without any reference to the general indemnity. Even if the respondents were to succeed in the Eastman Christensen and Kelvin Catering cases, the reclaimers would, in my view, prevail in the Wood Group case. The Northern Industrial Marine contract has no general indemnity and also contains a definition of "sole negligence" which makes it clear that what is in contemplation is the negligence of any party not just negligence on the part of the respondents. The Coflexip Stena contract is similar. In both these cases, the result should, in my view, be the same as it is in the London Bridge contract. The relevant clause in the British Telecom contract is in different terms, but the differences are not, in my opinion, sufficient to lead to a different construction, and the clause has, in my opinion the same effect on the clauses in the other contracts. In my opinion, therefore, the reclaimers’ arguments prevail in all the cases.
(a) Summary of Arguments in relation to Contribution
At a very late stage in the hearing before the Lord Ordinary, the respondents put forward an argument to the effect that since the reclaimers’ liabilities towards the injured persons or the families of the deceased had been met by their insurers, the reclaimers themselves no longer had any title or interest to sue for recovery of the sums which had been paid and that if there was any claim against the respondents under the indemnities it should be a claim for contribution at the instance of the insurers. The Lord Ordinary upheld that submission. The Lord Ordinary said, firstly, at p. 1434 of his opinion:
"In my view the principle behind contribution is very succinctly stated by Professor Gloag. The important factor is that parties should have undertaken the same risk to the same common creditor. However different the genesis of the contracts there can be no doubt that the pursuers’ insurers and the Contractor, if they have any obligation to OPCAL and the Participants, have it under contracts of indemnity. No doubt there is a difference regarding the consideration which prompts an indemnity between a case where the indemnity is given because of the payment of a premium and the case where it is granted because of the benefits of a Contract to provide services. An insurance indemnity will have features specific to it like the obligation uberrimae fidei just as a contract for services on an oil rig will have many features which will not be found in a contract of insurance. However it is clear from the authorities that the contracts which give rise to the joint debt need not be identical. The question is whether in relation to the creditors have the debtors obliged themselves for the same debt?".
The Lord Ordinary went on to say that in a case of this kind where no provision had been made for double recovery a creditor should not be able to recover twice and that while the insurers who had paid in the first instance should be able to recover either the whole, through subrogation, or a proportionate part, through contribution, they obviously could not have both. He went on to distinguish cases where subrogation could operate from those in which contribution was appropriate and expressed his conclusion, at p. 1437 as follows:
"I think the questions that arise on the matter under discussion ought to be settled on the basis of principle and not by reference to any rigid classification such as insurance and non-insurance. Initially attempts were made to confine contribution to particular categories of insurance and this was rejected by the Court as artificial. My conclusion therefore is that the Insurers of OPCAL and their participants do not have any right of subrogation in respect of the indemnities granted by the Contractors. The pursuers no longer have any title or interest to sue the Contractors. This means that if the insurers want to recover their outlay this would have to be by way of a separate action based on contribution. Thus only in respect of the Stena Offshore action is there a relevant claim under the Contractor’s indemnity and that to the limited extent of the Oxy Gap payment to the claimant Andrew Carroll less the tax benefit".
The Lord Ordinary expressed concern that the matter had been raised so late but held that, while that consideration might have consequences in expenses, it did not afford a reason for not sustaining the defenders’ argument.
The argument in the appeal ranged very widely and every relevant, or possibly relevant, authority was exhaustively discussed. As with other chapters of the argument, it would not, in my view, be profitable to attempt to follow all the details of the argument. It is sufficient to indicate the main lines of the submissions and the principal authorities on which these submissions were rested. The reclaimers advanced their principal argument under five main propositions. The first three of these were as follows: firstly, they submitted that under the remedy of subrogation in indemnity insurance an insurer, on payment, was placed in the position of the insured and entitled to all rights and remedies of the insured which might go to reduce the loss. Secondly, where there was subrogation, the rights and liabilities of third parties not connected with the contract of insurance were not affected, obligations owed by third parties to the insured were not discharged and the insurance contract was disregarded. Thirdly, an indemnity insurer was entitled to be subrogated to the rights and remedies of the insured against another party who was primarily liable for the loss and that entitlement extended to the insured’s rights and remedies against persons who were not wrongdoers or authors of the loss. These propositions were supported by reference to a number of authorities, particularly Mason v. Sainsbury (1782) 3 Doug. K.B. 61, Dickinson v. Jardine (1868) L.C.C.P. 639, Simpson & Co. v. Thomson (1877) 5 R. (H.L.) 40, North British Insurance v. London, Liverpool and Globe Insurance (1877) 5 Ch.D. 569, Darrell v. Tibbit (1880) 5 Q.B.D. 560 and Castellain v. Preston (1883) 11 Q.B.D. 380. It was further submitted that there was nothing in the decision of the House of Lords in Esso Petroleum v. Hall Russell & Company 1988 S.L.T. 874 which rendered the reclaimers’ submission unacceptable. The result of these authorities was that when the insurance contract was entered into the insurer acquired a contingent right of subrogation to all rights of the insured which might go to diminish the loss and on payment that entitlement vested in the insurer as from the time of the occurrence of the casualty causing the loss. From the point at which he made payment, the insurer was entitled to every advantage of the insured which might go to diminish the loss and that payment did not affect the rights and liabilities of third parties. The payment by the insurer did not discharge any obligation owed to the insured except in the case of double insurance. The legal right to enforce rights and remedies however remained with the insured and must be exercised in the insured’s name. The right of subrogation extended to all rights and remedies whether arising under contract, delict or other ground and was not limited to rights against wrongdoers or authors of the loss. In respect of third parties, the insured and the insurer were treated as one. Consequently, the correct approach was not to ask whether a payment made by an insurer had discharged the loss but whether, in all the circumstances, the law permitted subrogation.
The last two of the reclaimers’ main propositions related to contribution. Their fourth proposition was that a right of pro rata relief between two or more obligants by way of contribution could only arise where the obligations were on an equal footing. There was no equality in law between a general liability insurance contract and a claim for indemnity in a commercial contract. Fifthly, and in any event, the indemnity obligation which the respondents had undertaken, properly construed in its context, meant that the respondents had undertaken to bear the primary responsibility for any loss arising from the death of or injury to their employees. In this part of the argument, reference was made to Gloag on Contract supra and to a variety of authorities including Stair Institutions I.8.9 and Erskine Institutes III.3.74. Reference was also made to Deering v. Earl of Winchelsea (1787) 1 Cox 319, Craigthorne v. Swinburn (1807) 14 Ves. Jun. 160, Stirling v. Forrester (1821) 3 Blyth 575, Caledonian Railway Company v. Colt (1863) McQueen 833, B.P. Petroleum v. Esso Petroleum 1987 S.L.T. 345, Sickness & Accident Assurance v. General Accident (1892) 19 R. 977 and Albion Insurance Company Limited v. Government Insurance Office of New South Wales 1969 121 L.C.R. 342. On this basis it was submitted that there were a number of reasons why there was no equality in law between an indemnity contract and one of insurance. Insurance was normally taken out for the insured’s benefit as a secondary protection regardless of duties owed by other persons. Parties to commercial contracts expected the obligations of the other party to be met in full and there was no reason to treat indemnities differently. The rules as to double insurance were well-known in the insurance industry, having been established for over 200 years, and there were tested methods of dealing with questions of contribution among insurers which was not a matter necessarily of a simple pro rata division. There was no similar background of usage or precedent in dealing with co-ordinate liabilities between an insurer and an indemnifier. The contractors in the present case were not regarded as insurers and the nature and object of the services contracts were different from those of contracts of insurance. Insurance, for example, was a contract of uberrima fides, the insurer relying on information derived from the insured as to the risk. The obligation of an insurer was general, arising from payment of the premium, whereas the service contracts contained multiple reciprocal obligations. Insurance contracts were regulated by legislation particular to them. As regards the final point, it was common in the business world for parties to allocate risks and responsibilities between themselves and there were indications in textbooks that there was a practice of doing so in the offshore oil industry. The Lord Ordinary’s conclusion would create an inextricable tangle because of the necessity for different insurers bearing different shares of loss all to pursue their own actions for contribution. In any event, looking at the contracts themselves, the respondents had undertaken the primary liability. There was no provision designed to give the respondents the benefit of the reclaimers’ insurance, expressly or impliedly. The contractors were required to take out certain insurances and the indemnities required them to hold harmless and defend the reclaimers against claims. The indemnities were part of contracts in which each party gave indemnity to the other, in particular in regard to injury to employees and damage to property. The reclaimers also referred to a number of American cases, which are dealt with later in this opinion. Lastly, the reclaimers submitted that in the absence of some notice or plea on record it was not open to the respondents to take this point at the stage in the case at which it had been raised.
The respondents submitted that it was clear from the terms of the contracts that the obligation which the reclaimers sought to found upon was one of indemnity and indemnity only. A contractual indemnity, such as those in the present case, was an obligation to pay money and the amount payable was equal to the amount of the loss which the person entitled to the indemnity suffered as a consequence of the risk against which the indemnity was given. On the facts of the present case, the reclaimers had come under a liability and had called on both the respondents and their insurers to meet it. The insurers had responded and had discharged the liability. If the respondents were under an obligation towards the reclaimers, that was a separate obligation from that of the insurers and hence it did not necessarily follow from the fact that each was due to meet the loss that there were no contribution rights between them. In any event, the effect of the contractual indemnity was that the person entitled to it could only recover the amount of his loss and if anything happened to reduce the amount of the loss sustained by the person entitled to the indemnity, then that reduced to the same extent the amount which the indemnifier had to pay. In the present case, the insurers of the reclaimers had paid the liability and obtained discharges from the claimants and so had prevented the reclaimers from sustaining any loss. Once an indemnifier had paid the person entitled to the indemnity, he would have a right to insist that the person indemnified should pursue any rights which, if enforced, would go to diminish the loss. The rights so pursued were pursued by way of subrogation, that is, the rights remained rights of the person indemnified which must be pursued in the name of the person indemnified. The difference between that situation and the situation of a double indemnity in which contribution was appropriate depended on whether the payment discharged the liability: particular reference was made to Simpson & Co. v. Thomson supra and to the speech of Lord Jauncey in the Esso Bernicia case. The authorities showed that an indemnifier, whether or not an insurer, was subrogated to rights which went to reduce the loss but none of the authorities supported subrogation to other indemnities. In that connection, particular reference was made to the North British case and to Mason v. Sainsbury in which Lord Mansfield had approached the issue by considering the true nature of the claims, namely a claim based on liability for a wrong, on the one hand, and an indemnity on the other. The fact that a claim was pursued wholly or partly in the interests of an indemnifier could not enlarge the rights of the person indemnified against the third party on whom the claim was made. The contrast between an obligation to indemnify and other kinds of obligation such as an obligation to make reparation for a wrong was critical. Subrogation and contribution were mutually exclusive remedies appropriate in relation to different kinds of obligation. It would be an error to regard the rules as special to the case of insurance contracts: properly they were the application of more general principles of indemnity and in the case of contribution what was applicable was an equitable principle parallel to that applied between co-obligants under one contractual obligation. The critical issue was that where each of the obligations under consideration was one of indemnity the person indemnified could not recover more than one reparation. Payment by one indemnifier was a defence to a claim against a second and it was not possible to draw a distinction between two obligations of indemnity because they occurred in contracts which were different and had different purposes and different terms. The only question was whether each contract was one of indemnity covering an identical loss. The reclaimers sought to draw a contrast between insurance obligations and other contracts of indemnity but insurance was itself a commercial contract and it had been made clear in the authorities, particularly Castellain v. Preston, that the relevant law applied to insurance contracts only because they were contracts of indemnity. That was borne out by the fact that there was no room for subrogation in insurance contracts which were not indemnity contracts, as was indicated in McGillivray on Insurance. It was not possible to make a distinction based only on the fact that one of the indemnities appeared in a contract which included other types of obligation because some insurance contracts did provide for other obligations such as an obligation to provide services. Equally it was no ground for a distinction that one of the indemnities was given by an insurer whose business was the assessment of risk and another by a company engaged in a different business. Again insurers might be said to be obliged to meet claims but the same applied to anyone who granted an indemnity. As the authorities connected with questions of contribution or relief between co-debtors demonstrated, where two debtors did come under an obligation to a common creditor and one discharged his own obligation, and thereby discharged the creditor’s claims against the other, then various possibilities could follow, depending upon whether the debtor who did not pay was to be regarded as the primary or ultimate debtor, in which case the one who paid might be entitled to relief of the full amount; otherwise if the one who paid was primarily liable he would have no right of relief or contribution; and if both were equally liable, contribution would apply. The cases illustrated that there could be relief up to the full amount of the obligation. The authorities demonstrated that the proper application of the principles of relief or contribution depended on the whole circumstances of both contracts and the questions which arose between persons who were possibly liable in contribution or relief could only be worked out in an action properly brought by the debtors who had paid to recover a proper contribution. Thus, although the reclaimers argued that the respondents’ obligation was primary or ultimate, the answer was that even if that was correct the issue could only be worked out in an action of contribution at the instance of the proper pursuer. The obligations in each of two contracts to indemnify could contain terms to make it clear which obligation was primary in relation to any other but the contracts in the present case did not support the reclaimers’ position in regard to the primacy of the respondents’ obligation to indemnify. Parties might make what contracts they pleased and it was possible to make up a contract which would have the result that one indemnifier was obliged not only to indemnify the creditor but any other indemnifier. Such a result might have been achieved in some of the contracts in relation to different indemnities from those under consideration in the present action. So far as the present actions were concerned, however, there was nothing in the indemnities founded upon to take them outside the normal rule of contribution. It might be argued, as the reclaimers had sought to do, that insurance was always treated differently and that the insurer and the insured were regarded as one: but these were only metaphors and did not affect the proper rule. As regards the argument that the point in issue had been raised late, a pursuer seeking to enforce an indemnity had to address the question of what loss had been sustained as well as the question what peril had occurred and in the present case the pursuers’ averments went on the footing that they, and the other participants, had paid the claims. It was not till evidence had been led that the actual position, that the payments had been made by insurers, had been disclosed. The defenders had had previously a general awareness that insurers were involved but the extent of their involvement had not been known. The point was, therefore, properly open to the respondents on the pleadings in the case.
There is it seems to me, a substantial amount of common ground between the parties which can be expressed in the following eight propositions.
1. A contract of insurance is a contract to indemnify against loss. An insurer is not entitled to refuse to meet a loss which falls within the policy (in the absence of some particular defence such as non-disclosure) and an insurer who pays in accordance with the policy is entitled to be subrogated to any rights of the insured which go to reduce the loss.
2. A party is not entitled to receive more than one indemnity in respect of one loss. It follows that if the whole loss is paid by one insurer, the insured cannot claim any part of the loss from another insurer. For the same reason, any recovery made by the insured from any other person goes to the benefit of the insurer.
3. The rules which apply in the situation which is described as "double insurance", set out in the previous proposition, apply equally in any other case in which there is more than one contract of indemnity provided, that those contracts are contracts equivalent to and on the same footing as contracts of insurance, or in other words, as it has been put, are contracts of "indemnity and indemnity only".
4. When an indemnifier (whether an insurer or any other indemnifier) pays a loss he is entitled to a contribution from any other person who undertook to indemnify in respect of the same risk.
5. An indemnifier who has paid a loss is entitled to be subrogated to any rights which the person indemnified has against any other person responsible for the loss, whether as a wrongdoer or under a contractual liability (other than a bare liability to indemnify). There are many examples of cases in which the indemnifier is entitled to pursue claims in this way; the cases include Darrell v. Tibbitts (1880) 5 Q.B. 560 and North British & Mercantile Insurance Co. v. London Liverpool & Globe Insurance Co. (1877) 5 Ch.D. 569.
6. Subrogation is a remedy to be pursued in the name of the person insured or indemnified because it is a means of exercising a right or rights of the insured or indemnified person. Subrogation, however, is not to be regarded as an assignation of any right by the insured to the insurer.
7. Contribution is a right of the indemnifier which is based on equitable principles and is to be pursued in the name of the indemnifier who pays the loss, at least in the ordinary case.
8. In a question of contribution between insurers or indemnifiers, it is relevant to take into account what the contracts provided so that, for example, any limits on the cover or indemnity will be given effect.
By way of commentary on these propositions, it may be observed in the first place, that it is not entirely clear, at least from the authorities canvassed before us, what the origin of subrogation is, either as a right or as a term. However, the historical origin of subrogation may not be of any real significance since, in substance, it is clear that what is meant by the term is that the insurer who has paid the loss is entitled to pursue in the insured’s name rights which could have been pursued by the insured, without having had those rights assigned to him.
Secondly, with regard to the fifth proposition, the central issue in this part of the argument concerns the extent of subrogation. It is clear that the indemnifier is entitled to be subrogated to rights against a wrongdoer. What has to be decided is how far the indemnifier is subrogated to contractual rights against a third party and how a distinction is to be drawn between rights to which the indemnifier is subrogated and those which are to be the subject of contribution.
A third comment is that it is clear from the authorities which were canvassed that a case of double insurance is treated as being in some respects parallel to cases of double surety or joint and several liability under a bond, even though there is no common agreement between the parties. Again, it is not quite clear what the historical origin of that rule is, but the rationale is reasonably clear and is found in a combination of two factors or considerations. These are, firstly, that it is necessary, as a matter of public policy and to prevent fraud, to restrict the right of any person to recover more under a contract of insurance or indemnity than he has lost; and , secondly that equitable principles should be applied to decide, in a situation in which there is more than one indemnity, how the loss should be distributed.
A fourth comment is that if it is true, as set out in proposition 8 above, that it is relevant to take into account the terms of the various contracts, it is difficult to see how the ultimate test in considering the effect of any arrangements can be other than the true intention of the parties as disclosed by the agreements which they have made, subject to the "grand principle" which applies to contracts which are contracts of indemnity and indemnity only.
If that is correct, there would seem to be three questions to be answered.
1. Is there any objection in principle or authority to an arrangement whereby one of two indemnifiers agrees to bear the ultimate responsibility for a loss; or, to put the question another way, by which a party agrees to bear the "insurance risk"?
2. If there is such an arrangement and one insurer pays the loss, is there any reason why the claim against the other indemnifier should not be pursued in the name of the insured - that is, why this should not be treated as an instance of subrogation rather than contribution?
3. Is the contract in the present case such an arrangement?
There is perhaps a sub-question in relation to question 2 above, namely, whether, even if such a claim should normally be pursued as a claim for contribution, that rule should be enforced if the issue of contribution as against subrogation has not been raised by the defender on record.
In my view, there is not really much room for doubt about the answer to the first question. I can see no reason why a person who has agreed, as part of a commercial contract freely entered into, to bear the ultimate responsibility for a loss should have an equity, or an equitable claim, against any other indemnifier; nor can I see any reason why it might be said that any other indemnifier would be unjustly enriched if the indemnifier who had undertaken the ultimate responsibility were required to bear the ultimate loss. The principle of contribution is not based on contract but on equity, as is made clear in Mason v. Sainsbury and a number of other cases (see also Brodies’ Supplement to Stairs’ Institution p. 942, in regard to contribution between sureties). Nevertheless, it appears to me that it is necessary to establish what it was that the indemnifier undertook to do, before applying the equitable principle. A contract by which a person agrees to bear the ultimate responsibility for defined risks does not seem to me to be objectionable and could, indeed, be regarded as positively beneficial from a public point of view. The authorities cited do not, in my opinion, disclose any ground for regarding such a contract as objectionable: indeed they display parallel arrangements in the case of sureties in which, for example, a surety may be regarded as undertaking a surety for the others or for another, rather than for the principal debt. (see Gloag Contract 2nd ed. 208-9, Bell Principles 267 Chalmers & Ross v. MacCuloch 1757 M. 11746). The respondents, as I understood the argument, really conceded that, if appropriate language were used, such an agreement could be validly made.
The second question is, it seems to me, the crux of the case so far as the general law is concerned. It is clear that subrogation extends to every right of the insured to pursue a claim for damages, whether that right arises from delict or from breach of contract. Subrogation also extends to certain rights under contract, for example, the right to have property repaired under a lease and the right to payment of the price of real property: or, at least, the authorities indicate that any such rights are held by the insured for the benefit of the insurer after payment under the indemnity, which would appear to come to much the same things as subrogation. (North British & Mercantile Insurance Co. v. Liverpool London & Globe Insurance Co. (1877) 5 Ch.D. 569, Darrell v. Tibbitts (1880) 5 Q.B. 560). Why, therefore, should subrogation not extend to a right to have payment enforced against the party who has accepted the ultimate responsibility?
Looking at the matter first of all from the point of view of principle it seems to me that there are two main reasons which might be suggested as operating against the extension of subrogation to the case figured, along with a possible supplementary reason. The first reason is that both the indemnifier who seeks to enforce payment from the ultimate indemnifier and that ultimate indemnifier are, by definition, indemnifiers and fall to be treated on the same footing. The second reason is that one payment of the loss extinguishes the claim of the insured or indemnified person, leaving nothing to which the indemnifier can be subrogated. The supplementary reason is that there is a lack of authority to support a claim by one indemnifier against what I have called the ultimate indemnifier.
Repeated reference was made in the course of the argument to the opinion of Lord Low as Lord Ordinary in Sickness & Accident Assurance Association Limited v. General Accident Assurance Corporation Limited (1892) 19 R. 977 at 980. Lord Low said:
"In marine insurance a rule which has been long recognised is that when the insured has recovered to the full extent of his loss under one policy, the insurer under that policy can recover from other underwriters who have insured the same interest against the same risks a rateable sum by way of contribution. The foundation of the rule is that a contract of marine insurance is one of indemnity, and that the insured, whatever the amount of his insurance or the number of the underwriters with whom he has contracted, can never recover more than is required to indemnify him. The different policies being all with the same person, and against the same risk, are therefore regarded as truly one insurance, and if one of the underwriters is compelled to meet the whole claim, he is entitled to claim contribution from the other underwriters, just as a surety or cautioner who pays the whole debt is entitled to claim rateable relief against his co-sureties or co-cautioners. There is no reason in principle in my opinion why the same rule should not be applied to other classes of insurance which are also contracts of indemnity and this has been recognised by high authority in cases of fire assurance - North British and Mercantile Insurance Company v. London, Liverpool & Globe Insurance Company 5 Ch. Div. 569".
Later Lord Low referred to Simpson & Co. v. Thomson (1877) 5 R. (H.L.) 40 and observed:
"That case, however, appears to me to belong to a totally different branch of law from the present case. It exemplified an application of the doctrine that where the insured has a primary right against third parties who have been the authors of the loss, the insurers on making good the loss are entitled to be put in his place and to enforce the remedies which he would have had against these third parties. That, however, is not the doctrine which lies at the root of the rule of marine insurance to which I have referred, but is a doctrine which would be destructive of that rule. The right of an underwriter who has indemnified the insured to claim contribution from the other underwriters cannot be founded upon the doctrine of subrogation, because an assignee can have no higher right than his cedent, and a ship owner who has received full indemnity from one underwriter can never make any claim against another underwriter. The answer, therefore, to the claim of an underwriter who had paid, if made only in the right and as assignees of the insured, would be that the contract was one of indemnity, and that the insured had already been indemnified".
Lord Low’s opinion certainly makes clear that there is a distinction between cases in which contribution is appropriate and cases in which subrogation can operate and what the consequences of the distinction are, once it has been applied, although it may be observed, in passing, that Lord Low appears to have regarded subrogation as a form of, or equivalent to, assignation, a view which is not wholly in accordance with later authority and in particular with the views of Lord President Emslie in the Esso Bernicia case. Lord Low also makes it clear that for an insurer to proceed by subrogation, he must be in a position to say that there is a primary right of the insured which the insurer is in a position to enforce. What the opinion does not, however, in my view, deal with at all is the question whether and in what circumstances a right of the insured under a contract with another person may be regarded as a primary right and how the distinction between primary and other rights can be drawn.
A second source which was repeatedly referred to as indicating the principle to was the speech of Lord Jauncey in Esso Petroleum Company Limited v. Hall Russell & Company Limited (the Esso Bernicia case) 1988 S.L.T. 874. In that speech Lord Jauncey reviewed a series of authorities from Castellain v. Preston (1883) 11 Q.B.D. 380 onwards. These authorities were themselves closely reviewed in the course of the debate but for the present purpose it is sufficient to refer to Lord Jauncey’s treatment of them. At 882, Lord Jauncey cited the opinion of Brett L.J. in Castellain v. Preston in which Brett L.J. said, inter alia:
"Now it seems to me that in order to carry out the fundamental rule of insurance law, this doctrine of subrogation must be carried to the extent which I am now about to endeavour to express, namely, that as between the underwriter and the assured the underwriter is entitled to the advantage of every right of the assured, whether such right consists in contract, fulfilled or unfulfilled, or in remedy for tort capable of being insisted on or already insisted on, or in any other right, whether by way of condition or otherwise, legal or equitable which can be, or has been exercised or has accrued, and whether such right could or could not be enforced by the insurer in the name of the assured by the exercise or acquiring of which right or condition the loss against which the assured is insured can be, or has been diminished".
Lord Jauncey then referred to the speech of Lord Cairns in Simpson & Company v. Thomson supra in which the Lord Chancellor referred to the well-known principle that the indemnifier
"will, on making good the indemnity, be entitled to succeed to all the ways and means by which the person indemnified might have protected himself or reimbursed himself for the loss".
Having reviewed a series of later authorities, Lord Jauncey went on to state his conclusion (at p. 883). That conclusion was that the authorities left no doubt of the existence of a general rule that an indemnifier subrogated to the rights of someone whom he has indemnified could only pursue those rights in the name of that person. The principle which was drawn from these authorities was that subrogation was available as a remedy for an indemnifier who had paid under the indemnity in regard to any right which would go to reduce the loss suffered by the person indemnified but was not available in the case of another indemnity. That formulation of the principle, however, seems to me to leave open the same question as did the previous one, expressed by Lord Low, in that it does not provide either a reason for saying that there cannot be a contract under which one indemnity takes priority over another or a means of distinguishing cases in which the right of the insured does fall to be treated as going to reduce the loss on the one hand or as being a co-ordinate indemnity on the other.
A further source which was referred to as indicating a principle which would tend to defeat the claim of the reclaimers in the present case was the Australian High Court decision in Albion Insurance Company Limited v. Government Insurance Office of New South Wales (1969) 121 C.L.R. 342. I would cite, in the first place, the opinion of Barwick C.J., with McTiernan and Menzies J.J., at p. 345. The opinion states:
"There is double insurance where an assured is insured against the same risk with two independent insurers. To insure doubly is lawful but the assured cannot recover more than the loss suffered and for which there is indemnity under each of the policies. The insured may claim indemnity from either insurer. However, as both insurers are liable, the doctrine of contribution between insurers has been evolved. It began in the second half of the 18th century with Lord Mansfield’s decisions with regard to marine insurers and there is no doubt that it now applies generally to insurance which provides the insured with an indemnity. There is no reason why the doctrine should not apply to insurance against liability to third parties and there is every reason in principle that it should. The doctrine, however, only applies when each insurer insures against the same risk, although it is not necessary that the insurances should be identical. Thus one insurer may insure properties A and B against fire and the other insurer may only insure property A against fire. Again one policy may be for a limited amount and the other may be for an unlimited amount. One policy may cover the risk of a whole voyage and the other may cover only part of the voyage. Differences of this sort may affect the amount of contribution recoverable but they do not bear upon the question whether or not each insurer has insured against the same risk so as to give rise to some contribution. The element essential for contribution is that, whatever else may be covered by either of the policies, each must cover the risk which has given rise to the claim. There is no double insurance unless each insurer is liable under his policy to indemnify the insured in whole or in part against the happening which has given rise to the insured’s loss or liability".
Later, Barwick C.J. referred to the opinion of Lord Low in Sickness & Accident Assurance Association Limited supra. Reference was also made in the course of argument to the separate opinion of Kitto J. who was concerned to demonstrate that the rule of contribution in cases of double insurance was not, as has sometimes been said, an arbitrary rule imposed by decisions of Lord Mansfield but can be derived from and justified by fundamental principles of equity. Kitto J. expressed the principles as follows:
"The first, a principle applicable at law no less than in equity, is that persons who are under co-ordinate liabilities to make good the one loss (e.g. sureties liable to make good a failure to pay the one debt) must share the burden pro rata. The second is that since a policy of marine insurance is a contract of indemnity only, so that where several insurers have separately insured the one person against the one loss that person, though he may upon suffering the loss sue any or all of the insurers, may not recover more in total than a single reparation for the loss, the relation between the insurers in such a case is analogous to the relation between several sureties for a debt".
Kitto J. went on to demonstrate the application of those principles in the series of cases in which the rule had been developed. Towards the end of that discussion, at p. 351, Kitto J. referred to the difference between the way the rule had been applied at law and in equity but observed that that difference was not material and continued:
"What is important is the reason, namely that payment by the one discharges not only himself but each of the others and qui sentit commodum sentire debet et onus.
What attracts the right of contribution between insurers, then, is not any similarity between the relevant insurance contracts as regards their general nature or purpose or the extent of the rights and obligations they create, but is simply the fact that each contract is a contract of indemnity and covers the identical loss that the identical insured has sustained; for that is the situation in which ‘the insured is to receive but one satisfaction’ (to use Lord Mansfield’s expression) and accordingly all the insurances are ‘regarded as truly one insurance’: Sickness & Accident Assurance Association Limited v. General Accident Assurance Corporation Limited".
The principle which the respondents sought to draw from that opinion was that, irrespective of the general nature or the purpose of any contract in which an indemnity was included, nevertheless the fact that the contract was one of indemnity and covered the identical loss attracted the operation of the principle of contribution. It does not seem to me that Kitto J. went quite so far. What he was doing, in my view, was laying down the rule for the application of the principle of contribution in cases of contracts which are contracts of indemnity; he was not laying down a rule that all contracts of indemnity, irrespective of their form, necessarily fall to be regarded in the same way. If he did intend to go so far as to say that the mere fact that a contract was expressed as a contract of indemnity was sufficient to create a situation in which contribution was the only remedy, then it respectfully appears to me that he gave no sufficient reason for that conclusion. As I said at the beginning of this discussion, there seems to me to be no requirement of public policy or of equity which would prevent a contract under which indemnities have an order of priority from being applied according to its terms.
In these circumstances, I do not find in the authorities any principle which excludes the argument which the reclaimers in this case put forward. In countering the argument for the reclaimers, the respondents made two additional submissions namely that there was a lack of authority to support the reclaimers’ proposition and that as a matter of practice, or good practice, the issue of the priority of insurances or indemnities could properly be resolved, indeed only properly resolved, in a contribution action in the name of the insurer. As regards the lack of authority it seems to me that it is sufficient to say that the question does not appear to have been raised previously in the present form and that, as I shall mention later, there are indications that it has been assumed that no question of contribution arose as between insurance, or another contract of "indemnity and indemnity only", on the other hand, and a contractual indemnity forming part of a different kind of contract on the other. As regards practice, it may well be that where issues of the extent of contribution are raised they can most conveniently be resolved in an action of the kind which the respondents submitted was the proper action, but if the reclaimers are correct in their submission in this case, they are entitled to recover 100% of the sums paid to the victims and no question which requires any special procedure arises.
If there is no consideration of principle which requires an approach different from that which I outlined at the start of this section of the opinion, the question nevertheless remains whether there is any authority specifically against it. Reference was made, in the course of the debate, to a wide range of authorities but there are only a relatively small number which, in my view, might be suggested as being directly against the reclaimers’ proposition. There was some reference in this context to the speech of Lord Cairns L.C. in Simpson Co. v. Thomson supra at p. 42, but it does not seem to me that anything in that speech bears on the present issue. Reference was also made to the discussion of contribution between cautioners in Gloag: Contract 2nd edition p. 208 but that discussion is concerned only with the working out of rights of contribution in a case in which contribution is appropriate and does not assist on the question whether particular obligations are to be treated as contributory or otherwise.
The respondents founded on Morley v. Moore  2 K.B. 359 and Austin v. Zurich General Accident Insurance Company  1 K.B. 250. Morley v. Moore was concerned with the effect of a knock-for-knock agreement between insurers. The plaintiff, whose car had been damaged by the negligence of the defendant, recovered a sum from his insurance company but the sum recovered did not represent the whole of his loss. Nevertheless, in pursuance of the knock-for-knock agreement, the plaintiff’s insurance company requested him not to pursue any action against the defendant. The plaintiff did pursue an action in which he sued for the whole of his loss. It was held that the plaintiff was not disabled from pursuing that action, although if he recovered the full amount of his loss from the defendant he would hold that part representing the payment he had received from his own insurers in trust for those insurers. The case, in my view, shows no more than that the insurers were not entitled to restrict the actions of the plaintiff in the absence of some express contractual provision. Austin v. Zurich was a case concerned with the question whether an authorised driver under an insurance policy in favour of another person was or was not subject to the conditions of the policy. The particular authorised driver had been subject to summonses for driving offences which had not been notified to the insurance company. The driver had been sued and the case had settled and an action was raised in his name against another party involved in the accident. The case is only important because both MacKinnon L.J. and Uthwatt J. said that the action should have been one of contribution, in the name of the insurers, rather than a claim in subrogation in the name of the authorised driver. If it shows anything relevant for the present purpose the case may be an indication that an action for contribution could extend to 100% of the sum paid out but, as MacKinnon L.J. observed, the matter is really, in the circumstances of that case, a technical one and if anything the observation indicates that in a case where 100% recovery is sought the difference between contribution and subrogation is not material.
In my opinion, therefore, there is no principle or authority which defeats the contention of the reclaimers. On the other hand, there is one case in which there is, as I mentioned earlier, a hint or an assumption that an obligation of indemnity in a contract, which was not one of insurance or "indemnity and indemnity only", should not be regarded as on all fours with an obligation to insure or subject to contribution in the same way as such an obligation. The case is Larrinaga Steamship Company Limited v. The King  A.C. 246. The case was not concerned with questions of contribution or the like at all. It was a case in which insurers had paid for damage caused to a ship and the purpose of the action was to recover under an indemnity in a charter party. The point about the case is that there was no suggestion that the insurance obligation and the indemnity in the charter party might be regarded as on an equal footing or as giving rise to contribution. The decision does not take the argument very far because, as the respondents pointed out, the question with which this case is concerned never arose, but the very fact that it did not arise is what may be significant.
The reclaimers referred to a number of decisions of courts in the United States, including Luke Hart & Others v. Western Rail Road Corporation (1847) H.F. & H. 99, Hall & Long v. The Railroad Companies (1871) 80 U.S. 367 and Chicago St. Louis & New Orleans Railroad Company v. Pullman Southern Car Company 139 U.S. 79 (1981), but I doubt whether these cases take the issue any further than do the decisions in North British & Mercantile Insurance Company v. Liverpool London & Globe Insurance Company and Darrell v. Tibbitts. Reference was also made to North Central Airlines v. City of Aberdeen 370F 2d 129 (1966) and Consolidated Freightways v. Moore 229P 2d 82 (1951) but although these decisions are, I think, consistent with the reclaimers’ position, they do not, in my respectful view, contain much in the way of analysis of the issues of law which could contribute to the present discussion. Another case referred to, Patent Scaffolding Company v. William Simpson Construction Company 256 Cal. App. 2d 506 (1967) was clearly decided on an entirely different approach from that in any of the other cases cited, and involved discussion of questions of superior equity. The remaining case, F.H. Vahlsing Inc. v. Hartford Fire Insurance Company 108 S.W. 2d 947 (1937) may give some support to the reclaimers. It was a case in which Vahlsing had leased a shed from a railroad company and had agreed to be responsible for damage by fire to railroad cars placed on a spare line adjacent to the premises. Some cars so placed were damaged; the railroad company received payment from its insurers under a policy which covered the cars; and the insurance company raised proceedings against Vahlsing. The company succeeded and Vahlsing appealed. The decision of the Texas Court of Appeals, so far as relevant to the present issue, states:
"Appellant makes the contention that its lease contract with the railroad company is a contract of indemnity with reference to its liability to the railroad company for damage to cars occasioned by fire, and claims that the appellee by its contract of insurance with the railroad company, being one of indemnity, would not be subrogated to the rights of the railroad company in virtue of the appellants’ relation of landlord and tenant to the railroad company."
That argument was rejected and the appeal was refused. The passage quoted suggests that an argument akin to the argument for the respondents in the present case was being advanced, but the reference to the relation of landlord and tenant leaves, to my mind, some doubt as to the precise ambit of the argument and the decision. However, to the extent that it was held that the insurance company was entitled to recover, the case supports the reclaimers’ argument. There is, in any event, nothing in the American cases, in my view, to cast doubt on the approach which, for the reasons which I have given, I regard as appropriate.
In my opinion, therefore, there is nothing in principle or authority to refute the proposition that the proper approach in cases of this kind depends upon the proper construction of the contract and the true intention of the parties or to suggest that there is any reason, policy or practice to prevent such an agreement having its intended effect. In my view, therefore, if the reclaimers can satisfy the court that it was the true intention of the parties that the contractor should bear the primary responsibility or "the insurance risk", there is nothing to prevent that agreement receiving effect according to its terms. The final question in this part of the case, therefore, is whether the contracts between the respondents and the reclaimers should be construed as imposing on the respondents the primary liability or insurance risk in relation to death of or injury to their employees on the platform, subject, of course, to the conditions in relation to sole negligence or wilful default on the part of the reclaimers. It seems to me that the answer to this question is closely linked with the answer to the question discussed under the heading of the construction of the indemnities. For the reasons given earlier, it seems to me that head (c) of the particular indemnities is clear and that its application is not limited or controlled by any other provision of the contract. In terms of that provision, the respondents, the contractors, undertook to indemnify the reclaimers against all claims in respect of death of or injury to the respondents’ employees, whether or not the respondents had been at fault, subject only to the exception set out in the contract. The indemnity is part of a scheme of cross-indemnities relating to various different subject matters, and there is no reason not to give the words their full effect. So read, in my view, their meaning is straightforward. What the contract says in direct terms is that the respondents are to pay for death of or injury to their own employees. There is nothing to indicate that it was the intention of the contracting parties that the respondents should benefit in any way from any insurance arrangements which might be entered into by the reclaimers (cf. Scholefield Limited v. Zyngier  1 A.C. 562).
There are a number of other points, perhaps of lesser significance, which in my view do add some support to this view. In the first place, there are statements which support the idea that a contract of this kind, to bear the insurance risk, is one which reasonable men might enter into and which, in particular, might be contemplated in the oil production industry. Indications to that effect are found in Daintith & Willoughby United Kingdom Oil and Gas Law 2nd edition para. 1-845, Hazelwood P. & I. Clubs - Law and Practice 2nd edition pp. 269-70 and Sharp Offshore Oil and Gas Insurance. (see also Fontenot v. Mesa Petroleum 791 F 2d. 1207 (1986)). Sharp supra states at p. 108:
"The position in respect of employers’ liability is invariably dealt with by the exchange of mutual indemnities in respect of injuries to or deaths of employees. There is perhaps a simple reason for this. If an individual is injured he will expect to have a right to sue any party who may have been guilty of negligence leading to the circumstances which caused the injury. This party may be another contractor, the principal or his employer, or any combination of all three. The issue can become complicated by reason of contributory negligence. Determining liability and awarding costs can be a lengthy process in these circumstances, and this can only add to the anguish of the injured party, or the dependants of the deceased who may have been the sole breadwinner. The employer therefore accepts a responsibility to provide for his employees and will generally give the party with whom he is contracting full indemnity in respect of any suit or action brought against that other party."
These are not works of authority and it is true, as the respondents pointed out, that since the indemnities in the present case are not absolute but subject to exceptions, they do not achieve a complete elimination of questions of liability. Nevertheless, these references do support the idea that a bargain of the kind envisaged in the reclaimers’ submissions is not a particularly surprising one, nor one which could be considered to be offensive to ideas of equity or of enrichment. Further, if the respondents’ argument in favour of contribution as the proper remedy were correct in principle, then the respondents might have paid the claims and sought a contribution from the reclaimers’ insurers: but the answer to that would surely have been that the insurance contracts were entered into for the benefit of the reclaimers, not that of the respondents. For all these reasons, it seems to me that the respondents’ argument is unfounded and that the Lord Ordinary should have rejected it.
I would add that, at various stages of the argument, it occurred to me to wonder whether the whole discussion as to the interrelationship of indemnity, contribution and subrogation was not, at bottom, an argument about procedure which risked getting in the way of a proper consideration of the substantive rights of the parties. As Professor Birks has pointed out (Introduction to the Law of Restitution p. 93) subrogation is in the nature of a metaphor and the whole field of discussion is fertile in figurative language. It is at least possible, in my view, that much of the discussion in the past authorities has been shaped by difficulties which arose, or were thought to arise, in connection with the assignation of rights and claims. However, the concepts of subrogation on the one hand and contribution on the other seem to be embedded in the law which must be applied in the present case. I sympathise very strongly with the observations made by your Lordship in the chair on the question of subrogation and its relationship to assignation but I have not found it necessary for the decision of the present case to take that issue any further.
The last point to be mentioned in this section of the opinion concerns the stage of proceedings at which the contribution question was raised. I am obliged to agree that, according to current law and practice, it seems to be open to defenders to raise a point of law, even of such critical importance, at the end of the case, without prior notice. I regard that as a serious defect in our system of pleading, even though there is a partial remedy in that the court can take the conduct of the defenders into account in dealing with questions of expenses.
5. Consequential Loss
All the contracts contained provisions dealing with consequential loss. In five of the cases, the relevant provisions are identical, but in the B.T. and Coflexip Stena cases, there were different provisions. The respondents submitted to the Lord Ordinary that the effect of these provisions was to exclude liability for what was described as "the Texas enhancement" of the victims’ claims. The Lord Ordinary accepted that submission in the five cases in which the provisions were identical and in the B.T. case. Because of the different provisions in the Coflexip Stena contract, the argument was not available to the respondents in that case.
The background to this part of the argument is that, after the disaster, the reclaimers were faced with the possibility that the families of the victims might claim damages in Texas and that awards made in Texas might be very much higher than awards made in any claims in Scotland. As a result of a period of intensive negotiation, fully described in the opinion of the Lord Ordinary and summarised earlier in this opinion, settlements were reached at levels which represented a compromise between the amounts which might have been awarded as damages in a Scottish court and amounts which might possibly have been awarded in a Texas court. The relevant figures were all the subject of negotiation and express agreement between the claimants and OPCAL and as a result it is possible to say exactly what the amount of the Texas enhancement was in each case. For example, in the London Bridge case, which concerns the amount paid to the dependants of a safety operator, Robert Carroll, who was killed in the disaster, the Scottish value was agreed as £231,750 and the settlement figure was £683,854.
The relevant provisions in the London Bridge contract were Articles 19 and 20. Article 19 is headed "Participants" and provides that the contractor recognises that the company is the operator on behalf of a number of participants in a consortium and waives any right to require that any claims for indemnity or damages should be pursued by the individual participants or divided up among them. The critical Article is Article 20 and it is convenient to repeat its provisions. The Article is headed "Consequential Loss" and provides:
"Notwithstanding any provision herein to the contrary, in no event shall either the Contractor or the Company be liable to the other for any indirect or consequential losses suffered, including but not limited to, loss of use, loss of profits, loss of production or business interruption."
The respondents’ submission, which the Lord Ordinary accepted, can be set out in the following eight propositions.
(a) An indemnity is an agreement to meet a loss arising from a specified peril.
(b) Article 19 of the contract envisaged claims by the company against the contractor for indemnity or damages.
(c) The contract made a clear distinction between an indemnity and damages.
(d) The phraseology of Article 20, especially the opening words, meant that the indemnity by the contractor given in the previous Articles was limited, by the agreement of parties, by the use of the term "consequential loss".
(e) Consequential loss is a term of art to be understood by reference to the second limb of the test of remoteness of damages for breach of contract as set out in Hadley v Baxendale (1854) 9 Ex. 341.
(f) Accordingly, on a proper construction of the article the reclaimers could not recover any indirect or consequential loss arising out of any claims.
(g) The enhancements above the Scottish levels of damages were not recoverable since they arose out of special circumstances particular to the reclaimers, namely a particular susceptibility to the jurisdiction of the Texas courts, of which the respondents were not aware when they entered into the contracts.
(h) The enhancements therefore represented a different type of loss which was indirect or consequential and not recoverable.
With reference to proposition (g), the Lord Ordinary heard detailed evidence about the reclaimers’ business connections with Texas and the effect which those connections would have had in relation to conferring jurisdiction on a Texas court to hear these claims. The reclaimers did have a number of connections with Texas. Most of these might, very broadly, be described as ordinary contracts for goods or services, but there was one particular arrangement involving a company known as OCSI through which the reclaimers conducted their sales of oil in Texas. The Lord Ordinary’s conclusions about the evidence and argument which he heard are set out at pp. 1294 and 1295 of Volume 6 of his Opinion. The Lord Ordinary observes that over many years the reclaimers did business with a wide variety of Texas firms and tied themselves to many business transactions there, some of them being extensive and continuing. He continues:
"With such regular and essential business contacts with Texas it may be that OPCAL purposefully committed themselves to a business connection with Texas. The problem is that this, although a perfectly reasonable argument, can in no way be regarded as decisive. Just what risk OPCAL and its Participants were under in respect of these dealings through Texas is a matter of fine judgment; and of course these contacts do not represent the totality of the risk even looking at the problem without considering OCSI. OPCAL and its Participants have on any view associations with many companies carrying out a variety of activities in Texas and in a valuable and protracted litigation the negotiating parties must as Mr Silva said be concerned about what surprises might come to light.
If I had to describe the precise risk OPCAL were under in numerical terms I should find this difficult. A view that the risk can be described as 75% or for that matter 30% must be somewhat arbitrary given the complexity, diversity and uncertainty of the American and Texas legal systems. If I had to form any view I should be inclined to believe that the pursuers were somewhat over-optimistic in considering that there was little doubt that they would succeed on jurisdiction on the contracts alone. The settlements offered would certainly have been somewhat generous had the OCSI transactions not been discovered. On the other hand I think the pursuers are right to claim that the settlement was at the time the best on offer and whether or not it was reasonable to settle on the best terms available rather than face a protracted, difficult and expensive series of litigations is a difficult question. Fortunately it is a question that I do not have to answer for I consider that the OCSI situation transformed the claimants’ prospects in respect of jurisdiction.
I cannot read the facts relating to OCSI as I have reviewed them as meaning other than that OPCAL were selling the whole of their production through and under the control of agents in Texas."
Later the Lord Ordinary, having reviewed the activities of OCSI, says that, looking at the whole evidence about contacts with Texas, there could be little doubt, in his view, that there were such contacts as would be needed to render OPCAL and its participants very vulnerable to litigation in Texas. Neither party attempted to re-open in the reclaiming motion the question of the degree of risk to which the reclaimers were exposed in relation to litigation in Texas, nor to challenge the Lord Ordinary’s conclusion that, given the risk that jurisdiction would be established in Texas on the facts as he found them, the settlements were reasonable, although the respondents did emphasise that what was important was the reasonableness of the settlements, not whether the reclaimers had acted reasonably in entering into them. The respondents also emphasised that the Lord Ordinary had not made any finding as to what the position about jurisdiction would have been if the arrangements concerning the sale of petroleum through OCSI had not existed or had not come to light, and therefore that his finding that there was a substantial risk that jurisdiction would be established in Texas was based upon findings in relation to circumstances which could fairly be described as unusual.
The reclaimers’ submissions in the appeal fell into two main chapters. Firstly, they submitted that, on a proper construction of the contracts as a whole, Article 20 did not apply to the particular indemnity on which the actions were based because, firstly, Article 20 had no relevance to an indemnity granted in respect of loss arising from the reclaimers’ liability in delict towards claimants, the extent of which was properly measured by the law in regard to the assessment of damages in delict and, secondly, because Article 20 dealt with the causation of particular specified losses and not with remoteness of loss. Under the second chapter of their argument, the reclaimers, accepting that if Article 20 did apply to the indemnities the meaning of the words "indirect or consequential loss" could be equiparated to the second branch of the rule in Hadley v. Baxendale, submitted, firstly, that at the time of entering into the contracts the contracting parties, as reasonable men, would have had in contemplation that it was not unlikely that claims for damages at a Texas level arising from personal injury to or death of the contractors’ employees would be pursued; secondly, that the loss sustained by the reclaimers in settling claims at the mid-Atlantic level was a direct and natural loss and so recoverable; and thirdly, that the enhancements added to the Scottish level of damages, to raise it to a mid-Atlantic level, did not constitute a different type of loss so as to make them so remote as not to be recoverable. On the other hand, the respondents submitted, in regard to the first chapter, that Article 20 did apply to and limit the recovery which either party could make from the other under an indemnity clause. Under the second chapter, the respondents submitted that the clause excluded liability for losses which did not arise in the ordinary course of things, and in particular for losses attributable to unusual or special circumstances, even if those circumstances were known to the other party; that, in the ordinary course of things, injury and death claims arising from incidents in the North Sea were not pursued in Texas; and that, in any event, the losses in the present case arose as a result of unusual and special circumstances, namely the reclaimers’ arrangements for selling oil in Texas and the unique scale of the disaster. As part of that argument they submitted, under reference to a series of cases, that the expression "indirect or consequential loss" had acquired a standard meaning, which should be applied here.
I have attempted to summarise the argument schematically as it was presented to us. It does, however, appear to me that the argument does not really fall into two separate chapters at all. It seems to me clear that it is not possible to decide what the ambit of the clause exempting consequential loss is without first, or at the same time, deciding what is meant by the expression "indirect or consequential losses". In the discussion about the so-called first chapter argument, the respondents emphasised the very wide terms in which Article 20 is expressed, while the reclaimers stressed that there was no express link between Article 20 and the indemnity provisions and that there was no reason to connect Article 20 with losses sustained by a third party, which would be the subject matter of any claim under the indemnity. In resolving that difference it seems to me that it is necessary to determine what was the ambit of the losses which the parties had in mind: if they were looking only to losses measured by the contractual measure of remoteness of loss, that might well be a powerful reason for regarding Article 20 as directed towards claims for breach of contract and therefore claims arising between the parties themselves, and would therefore count strongly against the application of Article 20 to claims under the indemnity provisions. On the other hand, if the parties may be supposed to have had some other measure or type of loss in mind, that measure would require to be taken into account in determining the ambit which the clause is supposed to have.
That being so, it seems to me that the best place to start in attempting to resolve this part of the dispute is with the respondents’ argument that the phrase "indirect or consequential loss" has acquired a standard meaning which can appropriately be read into this provision. That contention was based on a series of four cases, in particular. The first is Millar’s Machinery Co. Ltd v. David Way & Son (1934) 40 C.C. 204. That was a case of a contract for the sale of a machine which contained a guarantee by the sellers to replace defective parts or make good defective workmanship and a provision "we do not give any other guarantee and we do not accept responsibility for consequential damages". On the facts, it was decided that the machine was not in accordance with the contract and that the defendants had been justified in rejecting it and, further, that on a true construction of the guarantee clause the plaintiffs, the sellers, were not relieved from the obligation to pay damages arising directly and naturally from their breach of contract. At first instance, Branson J. said (at 207-208):
"It is said for the defendants that the consequential damages referred to must be consequential damages arising out of the defective parts or defective workmanship referred to in the earlier part of the clause. In effect they say it means ‘We will replace defective parts for you, but if your machine has to stand idle while we do so, we shall not pay you damages on that account’. It is a question of construction, on which decisions on other contracts can give little assistance. It is fair to have regard to the results of one construction or another. I think that under our law parties are at liberty to agree that one party shall supply goods to the other at a certain price, but that if he fails to do so the other shall have no remedy against him in damages; but that would be a very curious agreement and would need to be made in the clearest and most unambiguous terms. The contention of the plaintiffs amounts to this, that they could contract to deliver goods but that from mere caprice they could at the last moment refuse to deliver and yet would not be liable to pay any damages. That does not seem to me to be a likely contract for business men to enter into, and I declined to put such a construction upon the documents in this case. I think this contract means that if it turns out that parts of the machine are defective and must be replaced the plaintiffs will not be responsible for loss caused by the machine having to stand idle while the parts are being replaced."
There was an appeal and the decision of the Court of Appeal is very briefly reported. Maugham L.J. said, (at 210) on the point which is relevant to this case,
"On the question of damages, the word ‘consequential’ had come to mean ‘not direct’, and the damages recovered by the defendants on the counterclaim arose directly from the plaintiffs’ breach of contract under section 51(2) of the Sale of Goods Act 1893. The plaintiffs’ machine would not do the work, and the defendants were entitled to get the best substitute they could for it."
Roche L.J. agreed and said that the damages recovered by the defenders were not merely consequential but resulted directly and naturally from the plaintiffs’ breach of contract.
The next case is Saint Line Limited v. Richardsons Westgarth & Company Limited  2 K.B. 99. This case arose on a reference by an arbitrator of certain questions for the opinion of the court. The dispute arose from a contract for the purchase of engines for a ship which included a provision that defective parts would be replaced or remedied by the engine builders but that their liability did not extend to any loss or damage caused by the failure of the defective part or by detention of the ship, but only to the cost of the actual replacement or remedy of the part, with a further provision in the following terms:
"Nor shall their liability ever, or in any case, either before, during or subsequent to the expiration of the said period of twelve months extend to any indirect or consequential damages or claims whatsoever."
The case came before Atkinson J., who pointed out that the damages were all claimed as general damages, no special circumstances being relied on, so that the damages recoverable would be the loss directly and naturally arising in the ordinary course of events from the breach or breaches of contract. He further pointed out that it would be for the arbitrator to determine whether and to what extent any sums claimed were recoverable within that rule. In considering the effect of the clause quoted above, he first of all indicated that the clause on its face recognised the distinction between direct damage and indirect damage and continued (at 103):
"But the relevant part of the clause is the third part, and one finds quite different language used, namely, that the liability is not to extend to any indirect or consequential damages. So one part of the clause recognises the difference between direct and indirect damages, a second part declines liability for any damage of any kind, and the third part excludes indirect or consequential damage.
What does one mean by ‘direct damage’? Direct damage is that which flows naturally from the breach without other intervening cause and independently of special circumstances, while indirect damage does not so flow. The breach certainly has brought it about, but only because of some supervening event or some special circumstances. The word ‘consequential’ is not very illuminating, as all damage is in a sense consequential, but there is a definition in the Oxford English Dictionary to which both sides have appealed: ‘Of the nature of a consequence, merely; not direct or immediate; eventual’. It cites the definition of consequential damages from Wharton as: ‘Losses or injuries which follow an act, but are not direct or immediate upon it’."
His Lordship then referred to Millar’s Machinery supra and continued:
"In my judgment, the words ‘indirect or consequential’ do not exclude liability for damages which are the direct and natural result of the breaches complained of. Mr Millar, in stressing the word ‘consequential’, sought to take advantage of the contrast with the word ‘immediate’ in the definition in the Oxford Dictionary - ‘not direct or immediate’ - and he argued that ‘immediate’ meant instantaneous, the first result in point of time, and that the only immediate damage here was the excess price which would have to be paid for the new engines. In Webster’s Dictionary the meaning of ‘immediate’ is, ‘not separated by anything intervening’, and it gives as a synonym the word ‘direct’. It merely means ‘flowing directly from without intervening cause’. I do not think ‘immediate’ adds anything to the word ‘direct’ and I do not think ‘consequential’ adds anything to the word ‘indirect’. What the clause does do is to protect the respondents from claims for special damages which would be recoverable only on proof of special circumstances and for damages contributed to by some supervening clause. I am satisfied that it does not protect them from the claims which are made in this case."
The third case is Croudace Construction Limited v. Cawoods Concrete Products Limited  2 Lloyds L.R. 55. In this case, the exception clause was in similar terms to those in the two preceding cases and the result of the case was the same, Megaw L.J. observing, in the Court of Appeal, that the decision in Millar’s Machinery Limited supra was binding on the court. Much the same can be said of the fourth case, British Sugar plc v. N.E.I. Power Projects Limited (1997) 87 B.L.R. 42. Again, the contract purported to limit the liability of a seller of electrical equipment in respect of consequential loss to the value of the contract. There was some discussion of a passage in McGregor on Damages which sought to explain the use of the term "consequential" and there was some suggestion of other possible meanings for the word besides that established in the previous cases. But the decision of the court was in line with the previous decisions, which were accepted as binding and Waller L.J. said, at page 50:
"Secondly, in any event, once a phrase has been authoritatively construed by a court in a very similar context to that which exists in the case in point, it seems to me that a reasonable businessman must more naturally be taken to be having the intention that the phrase should bear the same meaning as construed in the case in point. It would again take very clear words to allow a court to construe the phrase differently."
Waller L.J. also referred to the fact that given that there were two previous Court of Appeal decisions it would require a very clear difference in language to lead to a different result.
I have set out the authorities in detail because they constitute, as I understand the position, the essential plank in the platform on which the respondents take their stand. Having considered the detail of the cases, and of the clauses which were before the courts in them, it seems to me that what is most striking is the radical difference between the clause which we have to consider in this case and those which were before the court in the cases cited. In this case, the parties have included under the heading of "indirect or consequential losses", loss of use, loss of profits, loss of production or business interruption. As I understood the position, no one in the course of the debate seriously sought to dispute that loss of production and loss of use, at least following on a breach of contract by a contractor, could fall within the first branch of the rule in Hadley v. Baxendale and therefore could count as direct loss for the purposes of that rule and indeed for the purposes of the rule exemplified in Millar’s Machinery Co and the cases which have followed on it. There must also be circumstances in which loss of profits and business interruption would likewise fall within the first Hadley v. Baxendale rule. What is, in my opinion, quite clear is that the parties to the contracts before us in this case sought to exclude liability for loss of production and the other heads listed, whether or not those heads of loss were direct in the sense of the normal rule. What is more, there is every reason to see that as a sound bargain which commercial men might reach. The plain and obvious fact about work carried out by contractors on a North Sea oil platform is that a fault giving rise to an interruption of production could, in turn, give rise to enormous liabilities utterly disproportionate either to the value of the contract work or the resources of the contractor. In all the circumstances, therefore, it seems to me clear that whatever is meant by indirect or consequential loss for the purposes of Article 20, it is not the meaning defined in the cases on which the respondents relied, nor is it related to the rules in Hadley v. Baxendale.
That, however, leaves the question what the words do mean, and that is a question on which it is difficult to obtain much assistance from authority or dictionary definitions. A considerable time was spent in argument in analysing the decisions in Hadley v. Baxendale and the subsequent leading cases, including Victoria Laundry (Windsor) Limited v. Newman Industries Limited  2 K.B. 528 and Koufos v. C Czarnikow Limited  A.C. 351, but I have to say that it seemed to me that nothing useful emerged from those analyses. The refinements of the distinction between direct and indirect losses or the first and second branches of the rule in Hadley v. Baxendale in the context of recovery of damages for breach of contract, do not seem to me to assist in the construction of a clause in which, for the reasons I have already given, it appears to me that the parties departed from the ordinary distinction applied in such cases. Reference was made to Gloag - Contract 2nd Ed. p.696, where the author deals with "direct and consequential damages" but those words only appear as the head-note to a passage introductory to a discussion of the rule in Hadley v. Baxendale and it is, in my view, quite clear that for the reasons already advanced, the parties were not thinking along these lines. Reference was also made to Keating on Building Contracts but much the same comment can be made in regard to the discussion there. A number of dictionary definitions were mentioned in the context of the opinions in the cases discussed above, but again it is difficult to see that any of them are of particular assistance.
In the end, it seems to me, it is necessary to be guided by what the parties have themselves said. The clause itself gives four instances of what the parties regarded as indirect or consequential losses. Looking at the four together, it can, in my view, be said that they do display a common feature which is that they are concerned with circumstances in which there is an impact on the business and financial arrangements of the parties resulting from things which happen in the course of the contract. They are all examples of interference with the ability of the parties to earn profit. That is not, I appreciate, a precise statement, and it is not possible, I think, to be much more precise. Clearly the parties anticipated that there would be other types of loss similar to the four named heads which might also be covered. One must then, I think, ask whether the Texas enhancement element of the claims in respect of which indemnity is sought can, in any way, be seen as parallel to the sort of business interference with which the clause is primarily concerned. Like a number of other questions arising in this case, complex and important as it is, this may in the end have to be regarded as a question of impression and I have come to the conclusion that the Texas enhancement is a different sort of creature altogether from the type of consequential loss with which the clause was concerned. It is not an instance of interference with the ability of the reclaimers to earn profits: it arises from a liability incumbent on them.
Looking at the whole clause and all the arguments, in the light of the above approach, I have come to the conclusion that the proper view is that Article 20 does not exclude the reclaimers’ claim for indemnity for the amounts which they have paid in the five cases with which this part of the argument is concerned.
The relevant clause in the B.T. contract is clause 8.6 which provides:
"In any event in no circumstances shall B.T. be liable in contract, delict (including negligence or breach of statutory duty) or otherwise for loss (whether direct or indirect) of profits, business or anticipated savings or for any indirect or consequential loss or damage whatever."
While the clause is differently expressed, it seems to me that the differences are not material, given the approach which I have taken in the discussion of the clauses above, and accordingly that the result should be the same in the B.T. case.
6. Scottish levels of damages
(a) Effect of a Scottish decree in the settlement process
In the cross appeal, the respondents submitted two separate arguments both designed to lead to the conclusion that the reclaimers were not entitled to recover the full amount of the payments made to the various claimants but were only entitled to recover the Scottish element of those payments. Both arguments were put before the Lord Ordinary and rejected by him.
The first argument was founded on the fact that the settlement process involved the claimants in raising actions which were then settled but in which decree for the amount which the claimants were to recover was granted. It was submitted on behalf of the respondents that, in the light of that fact, the only event or peril arising under the indemnity clauses in respect of which the reclaimers could claim under the indemnities was a liability constituted by a Court of Session decree. The respondents were entitled to and did challenge the amount of the liability and in assessing whether the amount paid had been excessive the court could only have regard to Scottish levels of damages. As a corollary, even if historically there had been claims with reasonable prospects of success in Texas at Texas levels of damages, such claims had been extinguished, at the latest, when the Court of Session decrees were pronounced, and had no relevance to the present actions. So far as the reclaimers’ pleadings were concerned, the peril in respect of which the indemnities were invoked was not clearly specified and the pleadings were in places equivocal, but it was averred in particular that the payments were made consequent to the Court of Session decree. In their submissions to the Lord Ordinary, the reclaimers had focused on the word "claim" in the indemnity but the reclaimers could not now, or at any time after decree had been pronounced, found on claims since the payments were made consequent to decrees and any claim not previously extinguished was extinguished at that point. The reclaimers had submitted and the Lord Ordinary had accepted that the decrees were merely a mechanism, the purpose of which was to ensure that the reclaimers would not be exposed to further claims. However, whatever the explanation for the procedure adopted, the reclaimers could not now claim to be proceeding on the basis of a claim since a claim had been extinguished: see Comex Houlder Diving Ltd. v. Colme Fishing Co. Ltd. 1987 S.C. (H.L.) 85 case, per Lord Keith at 121 and Lockyer v. Ferryman (1877) 4 R. (H.L.) 32 per Lord Chancellor Selborne at 40. The reclaimers could not claim to proceed on the basis of payment because the indemnity against liability was an indemnity against loss occasioned by the liability, nor could they claim simply on the basis of there being a loss because there was the intermediate stage, that is to say a specific reference to liability. The indemnity clause contained a series of specific expressions and causes of action all of which had specific meanings. The reclaimers had insisted on a court decree for their own protection and that destroyed any argument that the decree was merely a matter of mechanics. It was necessary to identify in the clause what it was that the reclaimers sought indemnity for and the truth was that it was a liability, not a loss. Reference was also made to Ben Shipping Company (Pts) Ltd. v. An Bord Bainne  2 Lloyd’s L.R. 285. The way round the problem would have been for the persons sued by the original claimants to bring in the indemnifiers as third parties. Accordingly, both on the general law and in the particular circumstances of the case the respondents were entitled to put the recovery of the Texas enhancement in issue. If the relevant peril was liability under a decree, then the only question open to this court to consider was whether the decree constituted a reasonable estimate of the loss of the claimants. That would have been the question considered if the respondents had been brought in as third parties in the family actions and any averments about the Texas jurisdiction would have been irrelevant. A court could only properly have had regard to Scottish levels of quantification, a matter which was treated as procedural in Scottish private international law and therefore as a matter for the lex fori. By subjecting themselves to liability under a Scots decree, for their own reasons, the reclaimers had opted for one of the possible heads of the indemnity to the exclusion of the others with the consequence that the amount was open to be checked and the check could only be made by reference to Scottish levels of damages.
For the reclaimers, it was submitted that the obligation of the respondents was to indemnify the reclaimers against a loss arising from claims and liabilities and that that obligation was not extinguished by settlement of a claim, even if that took the form of the grant of a decree by consent. The issue to be resolved was whether the settlements agreed by the claimants and the reclaimers, including the grant of a decree of consent, were reasonable. In these circumstances, where the respondent did not challenge the Lord Ordinary’s finding that the settlement process actually pursued had been reasonable, they were not entitled to separate out the decree as an element by itself and claim that as a result of it payment should be restricted to Scottish levels. It was clear from the pleadings that payment made had been consequent upon a settlement of a loss resulting from the claims on liabilities and the Lord Ordinary had correctly concluded that the essential issue was whether the settlements were reasonable. The essence of the respondents’ argument was that the reclaimers had taken one head of the indemnity and let decree pass and that it was as a result of that that the Scottish levels of damages applied. The reclaimers have never sought to pinpoint a single head of the indemnity which covered a range of possible circumstances so that at any point in time during the process of settlement there was an obligation to indemnify. The respondents’ obligation to indemnify first arose when the claim was made two years before the decrees were passed and the respondents were in breach of their obligation from then on. Further, the respondents had encouraged settlement at levels above those which would be achieved in Scotland and had recognised that a premium should be paid to reflect Texan levels of damages. The respondents had been made aware of the settlement figures proposed and had recognised that those figures were reasonable for the reclaimers. The purpose of the decrees had been twofold, to protect the interests of children in the cases of family settlements and, on advice from the United States, to ensure that the disposal of the action amounted to "dismissal with prejudice" so as to prevent a later recourse to a claim in Texas. The procedure would also preserve any rights of relief against Score, including any subrogated rights.
In terms of the indemnity clauses, the contractors undertook to indemnify, hold harmless and defend the reclaimers against "any claim, demand, cause of action, loss, expense or liability" arising by reason of injury to or death of any of the contractors’ employees. It is, to my mind, clear that the clause was framed in the widest possible way so as to cover every form of loss which might fall on the reclaimers, irrespective of its precise form or the precise ground on which it might arise or the route by which it might be established. The essence of the respondents’ submission was that in some way the reclaimers had limited themselves to one particular form or type of liability and that that limitation carried forward so as to limit the reclaimers’ ability to recover from the respondents. I was, and remain, unable to see what basis in principle or in authority there was for this submission. I agree with the view of the Lord Ordinary that the process of settlement including the taking of a Scottish decree was entirely a matter of mechanics and that it has no effect upon the rights of the reclaimers against the respondents. Accordingly, as the reclaimers submitted, in the absence of any challenge to the finding that the sums paid were, overall, reasonable I see no ground on which this argument could be upheld.
(b) Construction of the indemnities in relation to levels of damages
The second submission made under this chapter of the argument on behalf of the respondents was that as a matter of general law and on a proper construction of the contract of indemnity the reclaimers were entitled only to recover Scottish levels of damages. The contracts all contained an identical governing law clause which clearly provided that the Scottish courts should have exclusive jurisdiction in regard to any question under the contract including any question under the clause of indemnity. The respondents could not be brought into any action as third parties other than an action in Scotland and that cut across any suggestion that the indemnifiers could be required to hold the reclaimers harmless in a foreign jurisdiction. A further consequence was that the quantification of the obligation to indemnify must always be a matter for the Scottish courts which would always apply the lex fori in a matter of quantification. The party to be indemnified could not come to the Scottish court seeking to discharge the onus of proving that what he had paid was reasonable and in doing so invoke foreign law. However the reclaimers had done just that, because they had asked the Lord Ordinary to hear expert evidence on foreign levels of quantification and make findings as to the values of claims in a foreign court. The Lord Ordinary had accepted that the compromise reached had been reached on a reasonable basis but that was on the basis of considering conflicting evidence as to quantification in Texas. Counsel for the respondents accepted that the authority in relation to quantification being a matter for the lex fori was limited to cases in which the Scottish courts were making a direct assessment of claims of damages but it could still be maintained that the contract treated as a whole envisaged only Scottish levels of liability.
In my view this argument can be quickly dismissed. It appears to me to be based on a non sequitur namely that a Scottish court in assessing the amount recoverable under the indemnity is restricted to considering the matter of the claim in the same way as it would if the question before it was one of quantification of a claim for damages in Scotland. I see no reason for that limitation and it seems obvious that it would involve a very substantial restriction upon the amount and value of the indemnity. In brief, in my view, the position is that a court considering a claim under the indemnity is not quantifying damages but determining what loss has been suffered by the claimants. In my view, the Lord Ordinary reached the correct conclusion on this aspect of the argument also.
7. Tax and Interest
With regard to the questions arising in relation to the impact of tax and interest on the reclaimers’ claim, I agree with the opinions of your Lordship in the chair and Lord Sutherland and have nothing to add.
8. General introduction to the factual issues
(a) Summary of platform lay-out and production process
Parties are agreed that the structure and layout of the platform, and the processes carried out on it, are accurately described in the Lord Ordinary’s opinion, subject to certain relatively minor corrections. Similarly, the Lord Ordinary describes in detail the organisation and management structures for the platform, the safety procedures in operation on it and a variety of technical matters relevant to the causation of the disaster. Many of these matters were also discussed in great detail in the argument before us. In my view, the issues which arise for decision in the appeal can be disposed of without entering into such great detail. It might be sufficient, in the circumstances, simply to refer to the Lord Ordinary’s opinion for a description of the platform and the processes and procedures operative on it. It does, however, seem to me that it is necessary to make some attempt to explain the position briefly, in order to identify those elements which are relevant to the decision of the appeal, while recognising that any such brief reference inevitably omits a great deal and, at least in that sense, cannot be completely accurate.
The part of the platform known as module C and the area described as the 68-foot level below that module are the centre of attention for the purposes of the appeal. At the production level of the platform (known as the 84-foot level from its height above the sea) there were four modules, separated from one another by firewalls. Each module was about 15 metres wide, 7.5 metres high and 46 metres long: and each stretched across the whole width of the platform from east to west at that level. Module C was the second module from the north. The control room and maintenance workshops and other structures were located in module D, at the north end of the platform. The well heads were in module A and the initial separation of gas, oil and water took place in module B, to the south of module C. At the east end of module C itself there were located three centrifugal compressors, occupying perhaps one quarter of the space in the module. To the west of the centrifugal compressors were the skids which accommodated the cooling and discharge scrubbers. To the west of those items there was a walkway running north to south. To the west of the walkway there were located two reciprocal compressors with associated cooling equipment and scrubbers. The west-most part of the module contained a water storage tank and various other items of equipment not immediately relevant to the issues in this part of the case. On the 68 foot level below module C there were located the JT flash drum, the condensate suction vessel, the booster pumps and the condensate injection pumps. The JT flash drum was towards the east end of the part of the 68 foot level below module C and the condensate injection pumps were towards the west. There was, however, a space between the westmost of the two condensate injection pumps and the west side of the platform. There was a stairway from the 68 foot level to the 108 foot level at the north-west corner of module C.
So far as the process is concerned, as it was in operation on the night of the accident, what came onto the platform from the undersea reservoirs was a mixture of oil and gas. The initial separation of oil and gas took place in module B. The oil separated in module B went, via booster pumps and a metering skid, to the main oil line pumps, from which it entered the main oil line and left the platform. Gas from module B was directed into module C, where it was divided into three streams, one of which went to each of the three centrifugal compressors. On the way to the compressors, the gas passed through suction scrubbers designed to remove any liquid: the amount of liquid obtained at this stage was relatively small. In the centrifugal compressors, the pressure and temperature of the gas were substantially increased. The gas then passed to the coolers on the centrifugal compressor skids. The cooling process generated quantities of condensate, so that a two-phase flow was formed, which passed to the discharge scrubbers. In the discharge scrubbers the condensate was separated from the gas by gravity and collected in a boot at the bottom of each of the scrubbers, from where it passed to the condensate suction vessel, located in the roof space of the 68 foot level. The gas from the discharge scrubbers went through a common manifold from which it was directed to two suction scrubbers, one associated with the first stage of each of the two reciprocating compressors. Any condensate which entered these suction scrubbers was routed back to module B, but little condensate was formed at that point. After the suction scrubbers, the gas passed through the first stage of the reciprocating compressors, which increased its pressure and temperature again. From the first stage of the reciprocating compressors, the gas went down to the 68 foot level where it passed through the JT valve, which reduced its pressure and temperature, creating a significant amount of condensate and therefore giving rise to a two phase flow which passed into the JT drum. Condensate coming from the condensate suction vessel was tied into the same line, i.e. the line from the JT valve to the JT drum, with the result that the JT drum collected not only condensate formed at the JT valve but also condensate coming from the discharge scrubbers of the centrifugal compressors. The greatest quantity of condensate formed in the process was formed at the JT valve. From the JT drum, condensate flowed first to the booster pumps and then to the condensate injection pumps, from which it passed, through a meter, into the main oil line. Gas from the JT drum was piped back up to module C where it went through the second stage of the reciprocating compressors. That gas was eventually routed into the gas lift system or to flare. All the stages of the process involved high temperatures and pressures. Accordingly, there were a variety of safety valves built into the system which would allow excess pressure to be relieved by releasing gas to the flare. Further, if any obstacle to the flow of gas and condensate through and out of the system should develop, (as, for example, if the condensate injection pump should fail to operate for any reason) it was possible to switch the compressors to what was described as a recycle mode, in which the compressors themselves would continue to run but instead of advancing hydrocarbons through the system would simply recycle the same material, excess pressure being released to the flare. It would appear that to recycle the reciprocating compressors was a relatively straightforward process. If the reciprocating compressors were recycled, condensate would nevertheless continue to be formed at the discharge scrubbers of the centrifugal compressors and to enter the JT flash drum, so that the level in that drum would rise, and eventually further action would be needed. To recycle the centrifugal compressors involved a much more serious interruption to production. If there were an obstruction to the flow of gas or condensate which could not be relieved, consideration would in the end have to be given to shutting down the entire production system. If the condensate injection pumps were to cease to operate, the production process could be kept in operation by recycling compressors and releasing pressure to flare for about one to one and a quarter hours, but after that time it would become necessary to shut down production.
The condensate injection pumps were reciprocating pumps. Each was fitted with gas operated valves (GOVs) both on the suction side and on the discharge side, the discharge side GOVs being also equipped with non-return valves. When the GOVs were closed the pumps were isolated from the remainder of the system. The control mechanisms were set in such a way that if a pump tripped, the GOVs would close: and an operator would require to reset them before restarting the pump. Similarly, the GOVs would close if an attempt to restart the pump was unsuccessful. Both the suction and discharge GOVs of the condensate injection pumps were pneumatically operated ball valves. There was a hole through the centre of the ball in each valve, through which condensate could pass when the ball was so positioned that the hole was in line with the pipework on either side. The ball could be rotated through 90 degrees, from the wholly open position to a wholly closed position: there might be some passage of condensate through the valve when it was in a partially open position. The ball was turned by a Scotch yoke mechanism operated by compressed air supplied from the instrument air system. The operation of the valves was controlled by push-pull buttons situated on stanchions placed at the south side of the pumps. If electrical power was available to the pumps, the position was that if the push-pull button was pulled into the open position, allowing the valve to open, it would remain open until the button was pushed to close it. If the pump was electrically isolated, the push-pull button could still be operated to open the valve, provided the instrument air system was not disconnected, but the valve would only remain open as long as the button was held in the open position, and if it was released the valve would close. There was a control panel from which the operation of the pumps was controlled when electrical power was available. That panel was situated on the east side of the pumps, between them and the JT drum. Condensate would enter the suction side of the injection pumps at a pressure of 670 psi. The pumps were designed to raise the pressure to about 1,100 psi. Pressure surges could occur, and there were pulsation dampeners built into the system to limit such surges. The operation of the pistons on the suction side of the pumps would force the condensate through, against a non-return valve, to the discharge side. On the discharge side of each pump there was a relief line leading to a pressure safety value (PSV). The PSVs themselves were situated in the roof space of module C, so that the pipework led up from the 68 foot level through module C to the roof space there. From the PSVs the relief line continued to the condensate suction vessel. The PSVs were situated in the vicinity of the walkway to the east of the reciprocating compressors. The PSVs were set to operate if the pressure should reach 1,750 psi: the pumps themselves were set to trip if the discharge pressure reached 1,700 psi. Adjacent to each PSV, and between it and the pump, there was a manual isolation valve which would be closed when the pump was isolated and depressurised and which would require to be reopened before bringing the pumps back into operation, to allow the PSV to perform its function.
(b) Safety procedures
A number of safety procedures were in operation on the platform. The Lord Ordinary deals with them in volume 1 of his opinion at p. 124 and following. The procedures which are most important for the present purpose are the "permit to work" system, which is described in production 12/405 and by the Lord Ordinary at p. 126 and following, and the hand-over system. The permit to work system applied to all work other than routine and uninterrupted work. The stimulus for carrying out particular work might arise from various sources, such as a planned maintenance programme or a decision to replace or upgrade equipment. Permits were of two types: hot work permits, on a pink form, were issued for work involving a risk of ignition: cold work permits, on a blue form, were issued for other work. Authority for carrying out the work would be sought by a person described, in the permit form, as the "requesting authority" who would complete the form with a description of the work to be done and of any safety procedures required, the time at which the work was to be done and any other necessary details including the name of the person responsible for carrying it out, the "performing authority". The form would then be submitted to senior management, the "approval authority" who would authorise the work. The actual issue of the permit to the performing authority was, however, under the control of the person in charge of the production process at the material time, the "designated authority". For the operations relevant for this case, the requesting authority would be the maintenance superintendent: the approval authority would be the offshore installation manager: and the designated authority would be the lead production operator. When a permit was issued, a copy would be kept in a box in the control room at the lead production operator’s desk. On completion of the work, the permit would be returned to and cancelled by the designated authority. At the end of the period for which a permit was valid, often the duration of a 12 hour shift, the permit would be returned to the control room but, if the work was not complete, might be extended or suspended. The object of the system was to ensure that the lead production operator, as the person directly in charge of the production process on any shift, should be aware of and able to control and take account of all the work being done during the shift. When equipment required to be electrically isolated to allow work to be done (as would be the case with maintenance work on a pump or recalibration of a PSV) red tags were completed: one portion was given to the electrician who carried out the work of isolation and attached to the relevant electrical equipment: the other was attached to the permit. On cancellation, the designated authority would sign that portion and give it to the electrician as authority to de-isolate the equipment.
The hand-over arrangements are described by the Lord Ordinary at p. 149 of his opinion. Generally, at a shift changeover (at 6 a.m. or 6 p.m.) the outgoing operators were required to tell the incoming operators anything which they "needed to know", for safety as well as other reasons. The handover was usually conducted with reference to notes or a log of what had been happening in the relevant area. Incoming operators, including the lead production operators, were also required to carry out an inspection, or "walk-round" of their area of the platform.
(c) Reclaimers’ theory of the accident
The reclaimers’ case was based on the following theory as to what took place on 6 July. On the evening of 6 July 1988, only one of the condensate injection pumps was in operation, namely pump B. Pump A had been taken out of service with a view to possible maintenance work and had been depressurised and electrically isolated. During the day-shift, the PSV for pump A had been removed from the system for recalibration by employees of Score. The PSVs were constructed with flanges on either side which were bolted to flanges on the pipework. Removal of a PSV involved undoing the bolts and lowering the PSV to a scaffold which had been erected for the purpose of the work. From there, the PSV was lowered to deck level and removed to the Score workshop (a cabin at a higher level on the platform). Blind flanges had been bolted in position by Sutton, one of the Score employees, to close the orifices in the pipework left when the PSV was removed. The proper practice was to secure the blind flanges by the use of combination spanners or a flogging spanner, so that the flanges would fit tightly enough to prevent an escape of hydrocarbons. Some time after 9.30 p.m. on 6 July, condensate injection pump B tripped and could not be restarted. It was necessary, because of the tripping of pump B, to recycle the reciprocating compressors. The lead production operator, Robert Vernon, decided to try to bring pump A back into operation. With this in view, he took steps to have the pump electrically de-isolated and himself went to the 68 foot level in order to re-pressurise the pump. In order to do so, he introduced condensate into the pump, by the process known as jagging, that is by operating the push-pull button for that pump. When pressure was introduced into the pump, however, condensate passed through the pump chest and into the discharge side and because, as it turned out, the blind flange fitted on the relief line was not sufficiently tight, condensate escaped into module C where it was ignited, by some means, and caused an initial explosion, which, in turn, led to further fires and explosions. At the appeal, the respondents did not dispute that, as the Lord Ordinary found, the initial explosion occurred in module C: at the proof, one of the respondents’ main contentions was that the initial explosion was in module B, but that contention was not reopened in the appeal. It was further not disputed that the fuel for the initial explosion was condensate and that a sufficient quantity of condensate could have been produced by the repressurisation of a pump, although it was maintained, as will be seen, that it would have been unusual to repressurise the pump in the way necessary to produce the requisite amount of fuel for the explosion.
(d) Respondents’ objections to the reclaimers’ theory
The defenders and respondents lodged a total of 49 grounds of appeal. Of these, grounds 1, 2, 3, 15, 16, 18, 21 and 22 can be said to relate to questions of law or construction of the contract documents or to their application to the facts. Grounds 4, 19, 20, 40, 44, 45, 46, 47 and 48 were not argued. Grounds 23, 24, 27 and 49 can be described as general grounds relating to the Lord Ordinary’s approach to the evidence and the correctness of his overall conclusion. The remaining grounds of appeal, which relate to the Lord Ordinary’s conclusion as to the facts and to the theory which the pursuers put forward can be grouped under six heads.
(1) An escape of condensate from the site of PSV 504 (the PSV for pump A) was not consistent with the evidence about events leading up to, and could not account for, the initial explosion. Under this heading may be grouped ground of appeal 34, to the effect that the expert evidence, in particular the evidence of Dr. Davies, did not support but was inconsistent with and fatal to the theory; 37, 38 and 39, to the effect that an escape of condensate at the site of PSV 504 could not have caused the pattern of gas alarms which was actually seen, in particular because of the presence of scaffolding near the site and because the expert evidence had wrongly relied on the behaviour of propane in modelling the behaviour of an escape and grounds 41, 42 and 43, to the effect that the presence of scaffolding would have influenced the release, that Dr. Davies had not taken account of that factor and that a gas release at the site would have produced alarms in section C2 of module C before those which were seen in section C3.
(2) It was not proved that a release of condensate at the site of PSV 504 would behave in the way which the pursuers’ theory required. Under this may be grouped grounds 35 and 36 to the effect that certain assumptions made by Dr. Davies were not proved and that therefore his evidence could not be relied on.
(3) The pattern of events envisaged in the pursuers’ theory did not fit with the evidence about the way in which the operation known as jagging, the introduction of condensate under pressure into an unpressurised pump, was actually performed. Under this head fall grounds 32, to the effect that the Lord Ordinary was wrong to hold that jagging had been done in a way consistent with the alarm pattern and the explosion, and 33, to the effect that he was wrong to hold that jagging by Vernon could have been performed in the time available, less than one minute.
(4) The jagging operation necessary to introduce the condensate into the pump and allow it to escape could not have occurred without the witness Grieve observing it. Under this head fall grounds 28 and 29 to the effect that the pursuers had never directly asked the witness Grieve, who was, as will be seen, present in the 68 foot level at a material time, whether Vernon had jagged the pump and therefore the pursuers could not claim that Vernon did so act; ground 30, that the Lord Ordinary was wrong to hold that the first stage of the jagging operation might have been performed before Grieve came on the scene and 31, which is generally to the effect that the Lord Ordinary was wrong to hold that Vernon jagged the pump.
(5) Vernon must have known that the PSV for pump A had been removed and not replaced. Under this head there fall a total of twelve of the grounds of appeal. Two of them, grounds 17 and 26, are in general terms to the effect that the Lord Ordinary was wrong to reject the argument that Vernon must have known, and two more, grounds 5 and 6, are to the effect that Vernon knew that the PSV had been taken out and replaced by a blind flange and was aware of its status. The remaining grounds make particular points about the evidence in this aspect of the case. Ground 7 is to the effect that Vernon had suspended the permit to work for PSV 504 at 6.00 p.m.; ground 9 is that the day-shift phase 1 production operator Grant would have told the night-shift phase 1 operator Richard that the PSV was out; ground 10 that Flook, the day-shift lead production operator, would have told Vernon; ground 11 that Vernon and Richard would have attended to a manual isolation valve adjacent to the site of the PSV before attempting to re-pressurise pump A and therefore would have known the status of the PSV; ground 12 that Vernon and Richard would have attended to a pressure control valve, PCV 1000, in the same vicinity and therefore would have been aware of the position; ground 13 that the Lord Ordinary failed to deal adequately with an admitted averment that Vernon had a duty to have the work site at PSV 504 checked; and grounds 8 and 14 which are generally to the effect that the Lord Ordinary’s conclusion was incorrect.
(6) The pursuers’ theory required that Sutton should have failed to tighten the blind flange adequately to prevent an escape of hydrocarbon, something for which there was no possible explanation. This head represents ground of appeal 25.
9. Novus actus interveniens
Having set out the competing positions on the main factual issues, it is convenient to deal with an argument, the effect of which, if successful, would be that the reclaimers could not succeed, even if the theory which they propose were held established. The reclaimers’ theory is that the disaster occurred, firstly, because Sutton failed to attach a blind flange to the pipework sufficiently tightly after the removal of PSV 504 and, secondly, because Vernon introduced system pressure to the pump leading to a leak of hydrocarbons at the insecure flange. The respondents submitted that, on the assumption that that was what had happened, Vernon’s action should be regarded as a novus actus interveniens, with the consequence that there was a break in the chain of causation between Sutton’s actions and the explosion which followed and that Sutton’s actions could not be regarded as, in law, a cause of the disaster.
The argument was presented with a wide-ranging reference to authorities concerned with questions of foreseeability and duty of care as well as with novus actus interveniens in a narrower sense. There are certainly situations in which the question of the existence of a duty of care and the question whether there has been a break in the chain of causation between a supposedly negligent act and a consequence shade into one another. On the facts which have to be assumed for the purposes of this argument, however, it appears to me that this is not a case which is concerned with the existence of a duty of care and that, if the question of a break in the chain of causation arises at all, it is a question to be considered strictly on the particular facts of the case. In these circumstances, the best approach to this part of the argument seems to me to be to set out the propositions which the respondents advanced and thereafter the factual situation against which the application of these propositions has to be judged. When that has been done, it appears to me that only relatively limited reference to authority is necessary.
The respondents submitted, in the first place, in general terms that a person who is negligent is not liable in law for all the consequences which in fact flow from his actions and, in particular, that he is not liable for something which would not have happened but for his negligence if the harm is brought about by a supervening act or omission of a third party. In these circumstances, it was submitted, the supervening act or omission might be regarded as the sole cause of the harm, being a new event which broke the legal causal link between the original negligence and the harm. It was further submitted that an intervening human action broke the causal link between a prior act of negligence and a subsequent harm if the intervening action was not reasonably foreseeable or the kind of thing which was likely to occur or, in any event, if it had to be characterised as unreasonable, extrinsic or extraneous. It was then submitted that, on the evidence, it was not reasonably foreseeable or likely that pump A would be pressured up by human intervention while PSV 504 was missing or, in any event, that if Vernon pressured up the pump while the PSV was missing his conduct fell to be characterised as unreasonable, extrinsic or extraneous.
Merely setting out the bones of the argument in this way makes it clear, in my view, that there are really two distinct propositions involved. The first is that the chain of causation was broken by Vernon’s act of repressurising the pump because Sutton could not reasonably have foreseen that the pump might be pressured up while the PSV was missing. The second is that Vernon’s action or, I think, more properly, the combination of circumstances which included Vernon’s action, were so unusual or surprising as to constitute a new cause in law. The argument in support of the first of these propositions led to consideration of cases such as Hughes v. The Lord Advocate 1963 S.C. (H.L.) 31, McKew v. Holland & Hannen & Cubitts (Scotland) Ltd. 1970 S.C. (H.L.) 20 and Dorset Yacht Company v. The Home Office  A.C. 1004 which were concerned with the question how far it was necessary to be able to foresee the type of accident which might result from a negligent act or the type of conduct which might arise as a result of such an act if liability was to be established. It is this part of the argument which, in my view, does not properly arise on the facts which have to be assumed for the purposes of this particular case. It is clear on the evidence that it was well-known to everyone concerned, including the Score employees, that a blind flange had to be fitted, and fitted securely, in order, inter alia, to prevent any escape of hydrocarbon gas or fluid. There was clear evidence to that effect from the witnesses Todd (74/11080), Reid (107/15495) and Pirie (209/30337) and it is clear that it was well-known that the proper practice was to secure the flanges tightly, to prevent any escape. In addition, counsel for the reclaimers pointed out that there was evidence that it would not do, in accordance with proper practice, to fit any plate as a flange but that the flange must be appropriate for the type of pipeline: since the line was described as a 900 lb. line it was necessary to fit a 900 lb. flange, which would stand a pressure of 2,000 psi: and Sutton would have known that he had to select the correct flange. That makes it clear that what was to be anticipated was that the flange might have to stand pressure and in the face of that I am unable to see how it can be suggested that the introduction of pressure to the flange, whatever the precise mechanism which led to that happening, was outwith the scope of Sutton’s duty or unforeseeable by him. It is true, as counsel for the respondents submitted, that the witnesses who were asked about the purpose of fitting a flange spoke in terms of "seepage" of hydrocarbons as the sort of thing which might be anticipated and that they did not specifically envisage repressurisation of the pump while the flange was in position. For example, the witness Todd said, (11080), that the purpose of fitting a blind flange was "to prevent any seepage of fluids past the isolation valves escaping to atmosphere". A little later, however, he said that the flange must be installed correctly and that that meant that it should be secured, as it was meant to be secured, tightly, to prevent gas escaping or any fluids escaping from the open pipework and that it should be flogged up. Reid, (15495), said that the purpose of fitting a flange was "the same as any line, you don’t want an escape of the medium into the module or whatever". Pirie, a Wood Group technician (30337), said that the practice was to fit blind flanges securely by flogging or using combination spanners and that one of the reasons for that practice was to prevent leakage. Ferguson gave evidence (171/25056) of a situation in which pressure had unexpectedly been encountered in working on pipework which was supposed to have been depressurised. Rankin also gave evidence (176/11414) about the fitting of blind flanges. He concentrated more on the role of a flange in preventing any foreign substances from entering the pipework and seemed unable to separate consideration of the role of a flange in preventing an escape of contents from the pipework from the question of pressure testing of a flange or of pipework, but, as with other parts of his evidence, it is by no means easy to understand precisely what he was saying. Looking to all that evidence together, it seems to me to be clear that the known and established practice was that blind flanges should be fitted tightly to prevent any escape of hydrocarbons and that the occurrence of pressure within pipework to which a blind flange had been fitted was not beyond the scope of the foresight of those who were engaged in the operation of fitting a blind flange. If so, an accident caused by a combination of negligence on Sutton’s part and negligence on the part of some other person in allowing hydrocarbon under pressure to enter the pipeworks would, ex facie, in my opinion, fall to be treated as an ordinary case of joint fault on the part of two wrongdoers, giving rise to joint liability.
Accordingly, in my opinion, the question whether there was a novus actus interveniens in the present case comes to depend upon the respondents’ second submission, which was that the events which brought about the introduction of hydrocarbon to the pipe were so extraneous or unexpected as to constitute a new cause. In this part of the argument, reference was made to a number of authorities, including the rescue cases such as Haynes v. Harwood  K.B. 146. Again, however, it appears to me that those references are not helpful in the present circumstances. In Haynes and the other rescue cases the situation was that a danger had been created and a person, the rescuer, had acted, to his injury, in response to the danger. This case, however, is not one in which Vernon acted in response to any perception of the danger created by Sutton. On the assumptions which have to be made for the purposes of this part of the argument, Vernon acted in ignorance of the fact that the PSV had been removed at all. In these circumstances, the authorities which seem to me to be of assistance are those which deal with extraordinary or complicated series of events which arguably may interrupt the causal chain between an act of negligence and some consequences. The locus classicus in which the rule of law is stated is the dictum of Lord Wright M.R. in The Oropesa  P. 32 at 39. Lord Wright said:
"The question is not whether there was new negligence, but whether there was a new cause...To break the chain of causation it must be shown that there is something which I will call ultroneous, something unwarrantable, a new cause which disturbs the sequence of events, something which can be described as either unreasonable or extraneous or extrinsic. I doubt whether the law can be stated more precisely than that".
Earlier, at page 36, Lord Wright had emphasised that, as in other branches of the law of negligence, it is necessary to pay attention to the particular facts. I think that it is also helpful to consider what Stephenson L.J. said in Knightley v. Johns  1 W.L.R. 349, a case in which a motorist had, by an act of negligence, created a danger in a tunnel but in which injury to a police constable was caused as a result of a series of steps taken and instructions given, negligently, by a police inspector who was responsible for dealing with the incident. At page 366, Stephenson L.J. said:
"The question to be asked is accordingly whether that whole sequence of events is a natural and probable consequence of the first defendant’s negligence and a reasonably foreseeable result of it. In answering the question it is helpful but not decisive to consider which of these events were deliberate choices to do positive acts and which were mere omissions or failures to act; which acts and omissions were innocent mistakes or miscalculations and which were negligent having regard to the pressures and the gravity of the emergency and the need to act quickly. Negligent conduct is more likely to break the chain of causation than conduct which is not; positive acts will more easily constitute new causes than inaction. Mistakes and mischances are to be expected when human beings, however well trained, have to cope with a crisis; what exactly they will be cannot be predicted, but if those which occur are natural the wrongdoer cannot, I think, escape responsibility for them and their consequences simply by calling them improbable or unforeseeable. He must accept the risk of some unexpected mischances...But what mischances?
The answer to this difficult question must be dictated by common sense rather than logic on the facts and circumstances of each case".
It is true that in that passage Stephenson L.J. had in mind cases in which there was deliberate intervention, similar to that in the rescue cases, but in my view the emphasis on common sense, the particular facts and the way in which different facts may lead to different consequences is valuable.
Following that approach, it seems to me that the point on which attention should be focused is how it came about that Vernon introduced pressure to the pump. That question is explored in other sections of this opinion. In brief, the result is, in my opinion, that Vernon attempted to repressurise pump A in ignorance of the fact that PSV 504 had been removed. That involved negligence on Vernon’s part and also the failure of the systems of control which should have alerted Vernon to the absence of the PSV. It may well be that the information was not communicated to Vernon because the pump had been "handed over to maintenance" with a view to carrying out operations which would render the pump unavailable at least for several days, so that the significance of the absence of the PSV was overlooked. Whether that is so or not, the position remains that the event which occurred was precisely the event which had to be guarded against by the proper fitting of blind flanges; and it was an event which was only likely to occur if there was a failure of some sort elsewhere. It is characteristic of the arrangements on the platform that, because of the hazardous nature of the operations, there were double and triple safety systems in place. There may well have been, in my view, a regrettable, indeed deplorable, failure in communication but, bearing in mind that what it led to was the introduction of pressure into the pipework and that the very purpose of the steps which Sutton should have taken in securing the flange was to prevent the escape of any contents from the pipework, it does not seem to me that what occurred can be regarded as a novus actus interveniens. In these circumstances, in my view, the respondents’ argument that there was a break in the chain of causation between Sutton’s negligence and the occurrence of the disaster, should be rejected.
10. The available direct evidence
As a general rule, the presentation of the submissions on both sides in the appeal and also, so far as one can judge, to the Lord Ordinary, proceeded by setting out propositions and then marshalling evidence extracted from the available evidence in support of those propositions. Because so many of those on board Piper Alpha died in the disaster, the available direct evidence about what happened is limited. It seems to me that it may be helpful to try to assemble the direct evidence that is available before attempting to seek support for any particular proposition. I therefore propose to try to summarise the evidence of the witnesses who were on board the rig and who gave relevant evidence.
Alexander Davidson Rankin
Mr. Rankin was employed by Score UK Ltd. as a valve technician and had been so employed for about two years. He had previously visited Piper Alpha and had worked on other platforms. He had worked on board Piper Alpha about three months before the time of the disaster and had returned for another tour of duty towards the end of June 1988. On this tour he was acting as supervisor, a role which he had not previously fulfilled. The work done by Score was the recalibration of valves, including PSVs 504 and 505. The other Score employee on this tour was Sutton; Rankin’s evidence was that he himself was primarily dealing with re-certification while Sutton dealt with matters such as blind flanges (11416). Rankin claimed to be familiar with the procedure for taking out, suspending and cancelling permits to work, although he could not recall whether he had previously suspended a permit to work (11430). By 6 July all the valves which required recalibration except PSV 504 had been dealt with. On the morning of 6 July, Rankin was told by the maintenance supervisor Smith that at some point in the day PSV 504 would become available for removal. Smith gave this information early in the shift and mentioned that there was further work ongoing involving pump A which would allow removal of the valve (11438). Rankin obtained a permit to work which had already been written out in part with the valve number and location on it and which was in the Score container. Rankin took the form, which was a blue cold work form, to Curtis the offshore installation manager who approved it. He then approached the production supervisor who also signed the permit. The copy of the permit available indicated that Curtis signed the permit at 7.40 a.m. but Rankin could not remember when he took the permit to the control room: indeed he could not remember being in the control room (11452). The valve could not be removed until a scaffold had been erected, which was not completed until the afternoon. A production operator, Peter Grant, was present in the afternoon and isolated the valve and was present when the studs attaching the valve to the pipework were released. Rankin was not clear as to what precise operations Grant performed at that time (11456). Rankin himself did the initial work of opening the valve to the atmosphere and he and Sutton completed the job of removing the valve. This was done at some time after lunch, after about 2.00 p.m. A rigger was present and the valve was taken down onto a small trolley on which it was removed from the module to be lifted by the crane up to the Score container. Rankin could not remember giving instructions or making arrangements in connection with the availability of the crane. Rankin proceeded with refurbishment and recertification of the valve while Sutton was involved in fitting blind flanges (11462). After the flanges had been fitted, Sutton returned to the Score container and helped Rankin finish the valve. Rankin did not visit the site of PSV 504 to check what Sutton had done. The valve required to be checked by an Occidental quality controller who appeared before 6 o’clock and Rankin thought that it must have been shortly before 6 o’clock, half-past five or so, before the valve was finished (11476). The work on the valve must have taken about two hours. Rankin was asked his reaction to a suggestion that the valve might have been removed before midday and said that he would not remember but his recollection was that they removed the valve after lunch and that it was not finished before 5 o’clock. From his discussion with Smith, Rankin understood that there was other ongoing work on the pump involving the pump coupling but at the time of the discussion he hoped that the valve would be removed and replaced in the course of the same day. Rankin had to return the permit to the control room and he remembered going to the control room about 6 o’clock (with a certain degree of tolerance either way). He went to the control room to return the permit and to discuss the fact that a crane would be needed that evening, without which the job could not be completed. In the control room, he saw one person. He did not know that person but the person was beside the electrical panel at the permit desk. He spoke to that person, showed him the permit and told him that the valve was ready to be returned to the pipework and that a crane was necessary to complete the job. He was asked if he said anything else and said (11484),
"No, I just brought it to the attention, that the valve had been removed, was now ready to be replaced and a crane would be required in order to complete the work."
The person said that no crane would be available that night but gave no reason. Rankin could not remember any other people coming or going in the control room. His evidence then continued (11485),
"Now, what then happened after you were told there was to be no crane available that night? Well, I said that the permit would then have to be suspended.
Perhaps if you just carry on and tell us what then happened, Mr. Rankin? Well my understanding is that jointly we suspended the permit.
You say that is your understanding? Yes.
But what did you do on this particular occasion, can you remember? I don’t remember.
Did you do anything with the permit? I don’t remember.
Did you have the permit with you? Yes, I left the permit on the desk in the control room.
How long did this discussion you had with this person last for? Very brief."
Rankin did not know the person whom he saw in the control room and could not remember if he had met him previously. He said that if he had been able to continue with the job he would have extended the permit. After visiting the control room, Rankin returned to the Score container and discussed with Sutton the fact that no crane was available. Two other persons, technicians from Wood Group of Peterhead, were in the container at the time. Rankin then made his way to the accommodation module.
Rankin was cross-examined about the procedure for obtaining, suspending and extending permits. So far as the actual events were concerned, however, the only point raised with him was that after returning to the accommodation module he spoke to Smith, the maintenance supervisor. The conversation took place shortly after 6 p.m., immediately after he had entered the accommodation module and Sutton was present. Rankin told Smith that the valve was still off and that he had returned the permit. Smith only asked whether blind flanges had been fitted and Sutton and Rankin confirmed that they had.
Mr. Bollands was an experienced production operator. He had worked as a phase 1 and a phase 2 operator and as control room operator on previous occasions. On 6 July, he was working as the night-shift control room operator. He came on shift at about 5.15 to 5.30, replacing an operator named Price (9373). The phase 1 operator was Robert Richard and the phase 2 operator Erland Grieve and the lead production operator was Robert Vernon. The day-shift lead production operator had been Harry Flook and the day-shift phase 1 operator Peter Grant: he was not sure about the day-shift phase 2 operator. Mr. Bollands described in detail the layout of the control room with its desks and of the fire and gas alarm panels as well as of the procedure for accepting alarms. Bollands said that his practice after the hand-over was to go and have a cup of coffee at a table at the back of the control room. In that position he could not be seen by anyone who came into the control room and went straight to the lead operator’s desk. By that time the person he replaced would have left the control room. Bollands left the control room to make a telephone call and returned to it at about 9.40. At about ten to ten, the condensate pump which was online tripped giving an audible alarm and a flashing light. Either Bollands or Vernon accepted the alarm. Bollands picked up his radio and called Richard, the phase 1 operator, and told him that the condensate pump had tripped. Vernon took his helmet and left the control room himself. Bollands presumed that he would have gone to help Richard restart the pump (9447). After Vernon’s departure the next thing that Bollands was aware of was a JCP panel alarm. Having identified it, Bollands accepted it and again notified Richard and said that he remembered telling Richard that it would most likely be a high level in the JT drum and asking Richard if he had unloaded the reciprocal compressors. The alarm relating to the JCP panel was perhaps three or four minutes after the tripping alarm. Richard acknowledged Bollands’ message. On further questioning Bollands was not sure whether it was at that precise moment that he had asked Richard about unloading the reciprocal compressors but he distinctly remembered asking him about that at some stage (9461). He could not actually remember if Richard acknowledged that he had recycled the reciprocal compressors but he thought that Richard would be in the area of the condensate pump and the JT flash drum and would have been able to know the level of condensate in the drum. Not long after the JCP panel alarm, Vernon returned to the control room and made a comment about not being able to get the condensate injection pump started. Vernon commented that he could not get the thing going and he suspected that there was a problem with the lubrication oil system. He also said,
"He wanted to get the other pump going, he wanted to take it back off maintenance." (9468).
Vernon also said, in answer to a question from Bollands, that the reciprocal compressors had been unloaded (9474). At the time, Bollands had no knowledge of the PSV on the standby pump being missing. He was then asked (9480),
"Was there any mention made of PSV 504 by Mr. Vernon? No mention at all, no.
In your view, Mr. Bollands, had he known that PSV 504 was missing would he have mentioned it? Yes, I would have thought so. It would have been quite a drastic step, a decision on his part to have started that pump knowing the valve was missing. I don’t think he would have done it.
What then happened after you had this conversation you have mentioned to us with Mr. Vernon on his return to the control room? What happened then?
Yes? He asked me to shout for Alex Clark which I did on the tannoy system. He got the permit out of the box, the pink permit, the maintenance permit and Mr. Clark phoned the control room."
Very shortly thereafter, Clark telephoned the control room and spoke to Vernon. Vernon, when he came into the control room, took the pink permit out of the 68 foot level box, a box for permits that were in force (9483). Bollands himself had looked at the permit in that box earlier in the evening and said (9484),
"Yes I remember the permit was there for the maintenance instruments lads to do instrument PM on it and that’s what they had been working on that night."
By that he meant work on condensate pump A. Clark came to the control room within a minute or so of being contacted (9484) and either he or Bollands tannoyed for Savage, one of the electricians, to contact the control room. Vernon spoke to Clark and told him that he wanted the pump back and then again left the control room. Bollands was not sure whether Vernon had signed off the permit before he left but thought that there was a good chance he would have done so because he wanted the pump signed off. Clark was quite happy at the suggestion that the pump should be brought back online. Vernon left the control room somewhere about two to four minutes before the explosion and the permit, he thought, was on the desk with Clark when Vernon left. Savage contacted the control room and spoke to Bollands on the telephone. Bollands told him that Clark wanted him to reinstate the condensate injection pump but Savage said that he was in the accommodation area and had all his gear off. Clark then tannoyed for another electrician, one of the night-shift electricians.
After that, the next significant thing was a low gas alarm on C centrif (9493). Bollands accepted the alarm. This was about two to three minutes before the explosion. He contacted Richard by radio and when Richard acknowledged told him that there was low gas alarm on C centrif. The next thing that happened was that one of the centrifugal compressors, either A or C, tripped. Bollands accepted that alarm and conveyed the message to Richard. He did not attempt to check which of the gas detectors on C centrif might have produced the low alarm because there was a faulty gas head which used to go off prematurely so that there often was a low gas alarm on C centrif (9496). Bollands said that from that point on he had difficulty in remembering exactly what happened. He said (9497),
"So I didn’t just get them coming on once, I got A, B centrif low gas alarm coming up and as fast as I was trying to accept them they were coming up. I am beginning to think there is something going on here. At the same time I got a high gas alarm off one of the machines. I am trying to shout to Bob Richard, I am trying to put…"
He was then interrupted by the shorthand writer and it is obvious that his evidence was being given in an excited way. A little later he said,
"I got one gas alarm, which was the first one we have mentioned, and then the other gas alarms all came in within a space of seconds of each other prior to the explosion."
These alarms included low gas alarms for A and B centrifs and a low gas alarm for west module C and also one high gas alarm for one of the centrifs. Later again he said (9500),
"And the point I was making earlier was I wasn’t just getting one gas head alarm for these machines, they were just coming in one after the other, I just couldn’t stop them; I accepted an alarm, the lights stopped, came on, which was obviously the gas head that came in, and then they started again which meant another one was coming in and the alarm went again."
Later, (9510) he said that the low gas alarms all came from the same area, the east end of C module, and that he had four lights in a row there, namely C2, C3, C4 and C5. The high gas alarm he said was one of the compressors but he could not say more. Between the very first low gas alarm and the alarms which came very quickly one upon another about a couple of minutes elapsed (9513). The high gas alarm came in immediately before the explosion. Bollands said,
"I was stood muting the alarm and accepting the alarm and talking on the radio when the explosion occurred, as I said, my hand was out here, I saw…"
Bollands was dazed by the explosion when it occurred. Clark was also in the control room and was injured and Bollands, with the help of Ian Ferguson, got him out of the control room. Bollands encountered Vernon in the area at the bottom of the stairs to the control room (i.e. at the west end of D module). Bollands was asked what Vernon said and replied,
"I conveyed … he was asking me what had happened and I conveyed about the gas alarms setting, the area we had had, and I conveyed the fact we needed to get the fire water pumps in operation because the fire water pumps were on manual because there had been divers in the water."
Bollands then saw Vernon leave with Carroll, wearing breathing apparatus, to try to start the pumps. Bollands himself succeeded in escaping from the platform. Bollands was asked whether he had seen any other event on the process mimic board and replied,
"We went through the gas alarms and we spoke about 1 centrif tripping, I believe 2 centrif tripped and maybe at the time all these gas alarms came in the third one went, because at some point at that time …"
Bollands went on to explain that it had occurred to him that if the centrifugal compressors tripped it would be necessary to put one of the John Brown turbines onto diesel fuel and repeated that he believed the third centrifugal tripped just before the explosion. There was no problem with any other equipment.
In cross-examination Mr. Bollands was asked questions about a number of matters, particularly in relation to the permits to work and hand-over procedure. He was not asked any questions about the events in the control room just before the explosion. One point which he did make (9565) was that he thought that if Vernon had decided to run the pump without the PSV in position he would have mentioned it to Bollands and he further said that he thought that Vernon would not have stuck his neck out and made the decision to run a pump without a PSV himself when there were senior production personnel and OIMs on the platform to whom he could have passed the buck.
Alexander Grossart Clark
Mr. Clark was maintenance lead hand on the Piper Alpha platform on the night-shift on 6 July 1988. At his hand-over, he was told by Mr. Smith that a decision had been taken to shut down pump A and carry out a repair to the Voith coupling. The application for a permit had been made and the isolation of the pump had been carried out and the isolation tags put into the "safety box" where they would stay until such time during the night as Clark needed to take the permit out to start work on it (9968). There was no reference to any work being done by Score (9971). Clark explained that he would not necessarily be aware if the pump had been isolated because Score had raised a permit for work of their own with the lead production operator (9973). On 6 July Clark had no knowledge of the status of PSV 504 and it had not been mentioned in the hand-over. He thought that if Mr. Smith had known that PSV 504 was missing at the hand-over time he would probably have told him. After the hand-over, Clark went down and raised permits for the work going on that night and then went and told the maintenance men where they would be working. He said that he knew he went to the control room some time after 6 but he could not be exact (9981). He saw Vernon there because Vernon had to sign the permits which Clark was extending. He said that it was busy, it was always busy been 6 o’clock and 7 o’clock, but he would pay no particular attention to who was there. Clark then visited most of the plant where work was going on. He did not pay any attention to any scaffolding in the vicinity of PSV 504 because there was so much scaffolding on the platform, "You never took any notice unless it was in your way or something like that" (9985). At about 9 o’clock he went up to the gas con module to speak to the men working there. He returned to his office about half-past nine and the next significant thing was a tannoy message for him (9988). He contacted the control room and spoke either to Vernon or Bollands, he thought probably Vernon. The message came at about 9.45 and he contacted the control room immediately. He was told that B condensate pump had tripped and could not be restarted and said,
"They wanted somebody down to look at it straight away and then they said, they asked if there had been any work started on the A pump to which I said, no, it was isolated, that was all. In the matter of the discussion we decided we would put A pump back on, deisolate it and put it back on and that would give us time to work on B pump."
Clark explained that he could not remember whether he spoke direct to Vernon or whether his communication with Vernon was through Geoff Bollands who would be in radio contact with him. In the course of the telephone call there was mention of the isolation stubs which would have to be signed off by Vernon and Clark so that an electrician could de-isolate the electrics (9992). Clark told White what was going on and then made a tannoy call for Savage to get in touch. He then went to the control room. He had difficulty putting a time on the interval between the first message to him and getting to the control room but it was something in the region of ten minutes. In the control room he saw Bollands. The phone rang and Clark answered: the call was from Savage who explained that he had gone off shift. Clark then decided to get the two other electricians to de-isolate the pump and he signed the de-isolations, namely the bottom part of the red tags. Vernon had already signed his part of the red tags. When he went into the control room, the isolations were already there and Vernon had already signed them: he did not see the permit which he did not need because it had never been taken out (9999). He saw only the tags, not the permit, and the tags were on the lead operator’s table. He did not remember attempting to contact Young. While Clark was in the control room a gas alarm came up for C module east. Bollands accepted it and radioed for Richard to go and check. He then said,
"I was just in the process of finishing signing the red isolation tags; then I lifted the tags and I had lifted my helmet, my hard hat, and I had turned to walk out when another gas alarm came up and I had gone to walk away just as Geoff stepped forward to accept it, and that’s when there was an explosion." (10002)
Only minutes elapsed between the two alarms. Clark was knocked to the floor with a lot of debris on top of him. Clark left the control room and went down the stairs to the north-west corner of the platform where he said he saw Vernon who said he was going to go and try and start one of the diesel pumps. Clark said that he thought the controls were away and that it was pointless but Vernon and Carroll took breathing apparatus and went into the smoke after which he did not see them again. Mr. Clark further explained that although he knew the pump had been shut down and electrically isolated he did not know whether or not it had been depressurised. He thought that the tags which would originally have been attached to the permit would have become detached when Vernon signed them because the only part the electrician needed was the tags. Mr. Clark denied seeing Vernon in the control room. He only saw him after the first explosion (10015).
Mr. Clark was extensively cross-examined. It was suggested to him that he would have wanted to know what work under the auspices of the maintenance department had been carried on during the day but he took the position that he did not need to know what jobs had been suspended but only those on which he would be working (10032). It was suggested to him that the maintenance department would have been told by the production department that the valve PSV 504 would be available for Score to work on but said that he did not think so and that Score had their list of valves and had to liaise with production about their availability. It was suggested to him, at some length, that that evidence was inconsistent with evidence which he had given at the Cullen Inquiry. He agreed that at the Inquiry he had said that normally the Score people would tell the lead maintenance hand what work they were doing and what the status of the valve would be and that normally that would be recorded and passed on by the lead maintenance hand. Mr. Clark confirmed that his visit to the control room was probably after 6 o’clock, that the permit for work on the pump coupling had not been taken out because he would have been the person to take it from the safety office if work was to commence. He was asked whether if on the evening of 6 July the permit was in the tray of the lead production operator’s desk it followed that it had been issued for the purpose of commencement of the work and said that it did. He insisted, however, that the permit had not been taken out but had only been written and that that was what he had been told at the hand-over. Mr. Clark denied that he had spoken to Vernon in the control room or that, as Bollands had said, there had been a permit on the desk with red tags attached to it. He agreed that if all that was on the desk was the two red tags, he would not have known which permit to work those tags had related to and therefore would not have known whether he was the performing authority in respect of that permit.
W. H. Young
Mr. Young was a technician working in the gas conservation module on the night-shift on 6 July. He worked until shortly before 10.00 p.m. He then left his place of work and went to the store and then to the instrument workshop where he arrived perhaps at ten minutes to ten. The technicians were engaged in writing their logs. Young heard Clark calling on the tannoy for Savage before he left the gas conservation module and again in the store. When he was in the workshop, the phone rang and one of the other technicians, Wakefield, answered it and then said that Young was wanted at the condensate pumps. The other technician there, Cassidy, asked if Young would like him to go but Young decided to go himself, picked up his tools and left. He went down the stairs to the 68 foot level but did not see anyone to remember on his journey and in particular did not pass Richard. At the 68 foot level he saw Vernon and Grieve near the condensate injection pumps. He started to walk towards them but they were coming away from the end of the pumps, which were very noisy, and therefore Young went slowly towards them. Before he could speak, there was a short burst of sound like the noise of an old steam engine ending in a "dull thud or boom boom". Young was facing south but the noise came from behind him. He started to look round and was then caught by a rush of hot air. He went backwards and lost his hat, ear defenders and spectacles. He thought the blast came from the direction of the stair behind him. Later he found that his hair had been singed. He went towards the stair. Two people rushed passed him from above and he found that the higher layer was full of smoke and returned to the 68 foot level. He was aware of an unusual mist appearance at the 68 foot level. He went across to the dive platform where he met Carey. He was covered in white dust, perhaps from lagging.
Neil Gallagher Cassidy
Mr. Cassidy was an instrument technician working on the night-shift. He confirmed that he was working in the workshop when a phone call came for Young to go to the condensate injection pumps. At about 10.00 p.m. there was an awful high pitched noise, like metal grinding, which he had never heard before. By this time Young had been away for about ten minutes. After the noise the whole workshop shook and he was knocked off his feet. When he left the workshop there was flames and black smoke coming from module C.
R. J. MacGregor
Mr. MacGregor was working in the mechanical workshop. He had been engaged in fixing a plate to a cooler which was surrounded by scaffolding in module C. He finished that work at 9 or 9.15 and got his permit signed off and went to have tea. While he was talking there was a noise like a banshee and after about thirty seconds an explosion. After the explosion he saw Vernon and Carroll at the north-west corner of the platform with breathing apparatus going to start the diesel pumps. They came back and said there was no chance of doing so. After trying to get through module D he went down to the 68 foot level by the north-west stairs and down a rope to the 20 foot level. He last saw Vernon and Carroll going down the steps at the north-west corner, they being behind him on the steps.
J. M. McDonald
Mr. McDonald was a rigger on the platform. During the day on 6 July, Sutton came down to get someone to take the valve out and McDonald sent Rutherford. Later he went down and helped to move the valve out. A chain block was left hanging after that had been done. At about 4.45 p.m. McDonald went down and found the chain block on the floor. He picked it up and put it in the rigging loft. He met Rankin or Sutton who said that the valve was ready but was not going in that day. This was about 5.30 p.m. He was in bed when the rig blew up. He had heard a man named Mochan say that he had heard Sutton and Rankin asking Smith if they could put the valve back in.
Mr. Grieve was working as a production operator on the night-shift on 6 July 1988. He was the phase 2 operator but because the gas conservation module was shut down he had little to do except watch the progress of the maintenance work. At about quarter to ten, Grieve heard the control room telling Richard that the condensate pump had tripped. The conversation was between Bollands and Richard. He was not aware of any response made by Richard. He could not recall the actual content of the message. He could not recall whether immediately after that conversation or whether after a further conversation he decided to go down and see if he could help with the condensate pumps. He did so perhaps five to ten minutes before the explosion. He went down the stairs at the east end of the 68 foot level and went passed the JT drum towards the condensate injection pumps. He would not have a view of the pumps themselves until he passed the JT drum. When he did so he was aware of Vernon and Richard at the west side of B condensate pump. It would be at the walkway at the corner of B pump when he was first aware of them, between the walkway on the west and the foot of the GOV area on the south side. He could not speak to actual distances. He was aware of coming past the control panel and being aware there was no pump running and eye-level was showing in the drum. He could not recall whether it was at that point or whether he approached the area where some sign language or shouting decided that they would go for a run on B pump. He was not sure whether he reset the GOVs on the way past or whether they had already been reset but the GOVs were reset for B pump and at that point the button was pressed for a start on the B pump. The impression he got was that Vernon and Richard were in the vicinity of B pump and a decision was made to have another go at starting it as the GOVs were re-latched at that point. The only thing he could add was that somewhere during that period Richard was called away. Grieve went to the local control panel of B pump and pressed the start button, watched the motor turn, took his hand away and the motor stopped. After that he made his way down to check the torque converter to make sure it was set to manual and the speed was wound off it. The call came for Richard shortly before or immediately after he had done this but he was not sure which but would say shortly before because all that happened after the attempted start was that he walked round the back and by that time Richard was going up to C module. He was aware of Richard answering his radio and going upstairs but he could not say what the radio conversation was. Richard did not hang about to say where he was going or why but he was not running up the stairs either. There was nothing that drew his attention to it. He was not aware of any great period of time, only a minute or so, between the call to Mr. Richard and the explosion. At the time of the explosion, Grieve was on the west side of the pump skid a few feet away from the speed controller. His recollection was that Vernon was in the vicinity of the GOVs on the south side of the pumps. He said,
"My last recollection of having seen him was between the GOVs on A and B pumps."
He could not be more specific and could not say whether Vernon was on his way back round to the local control panel or another reset to start the pump. He said,
"I’ve got no idea to what he was actually doing in the area."
He repeated that his last recollection was seeing Vernon between the two GOVs and in answer to a question as to where he himself was he said,
"I can’t recall. As I said that’s the last area that I was aware of him being in when I looked at him when I walked round the panel. I’m sorry I can’t help any further than that."
Mr. Grieve was asked in some detail about why he had gathered the impression that Vernon and Richard were trying to start the pump and it appeared that that came simply from the fact that they were present in the area and that he would have expected that they were trying to start the pumps. Mr. Grieve added that he recollected Willie Young arriving at the scene shortly before the explosion at the north-west stair of C module. Mr. Grieve felt the explosion was immediately above him and at the time he fell to his knees. He got to his feet again and started making his way to the west side of the module and as he turned back to look he noticed a ball of flame directly above the area of the injection pump which seemed to be hanging in the roof void space. This was just a large ball of flame which did not seem to have any force behind it and generally was a bluish sort of orange hue. After he had seen the flame he and Young went up the north-west stairs. At the 84 foot level he was aware of Vernon and Carroll putting on breathing apparatus. He was involved in an attempt to run out a hose. Later he found himself with Carey, Vernon and Carroll on the 68 foot level getting ready to go down a knotted rope but there was then a big explosion and they were engulfed by flames and went straight over the side. Grieve was not cross-examined on his account of the events which he described.
11. Vernon’s knowledge
The question whether Vernon knew of the absence of PSV 504 is critical in two respects. Firstly, if he did know, it may be argued that it is extremely unlikely that he would have repressurised pump A: there is clear evidence that it would be bad and dangerous practice to do so. Secondly, if he did know, it is not in dispute that he would be guilty of wilful default if he pressurised a condensate injection pump in the absence of its PSV and thus caused a gas leak, which led to the explosion.
The evidence bearing on this topic is, it must be said, fragmentary. The Lord Ordinary’s approach, broadly speaking, was, as I read his opinion, to refrain from making positive findings on particular matters of detail where such findings were not clearly justified by the evidence but to discuss and assess the effect of the evidence as a whole. On that basis, he reached the conclusion that an experienced operator such as Vernon would not have deliberately repressurised a pump knowing that the PSV had been removed and that there was nothing in the evidence, taken together, which compelled him to the conclusion that that was what Vernon had done. The Lord Ordinary’s approach was criticised, in a number of respects, by the respondents, but it became clear in the course of the argument that the position adopted by the respondents on a number of issues in the appeal was radically different from the position which they had adopted before the Lord Ordinary and that the manner in which the Lord Ordinary dealt with these issues resulted from the submissions of the respondents, even if the respondents now criticised the results. In my view, it would be a barren exercise to trace through all the passages in the Lord Ordinary’s opinion to which reference was made and to try to see what exactly the Lord Ordinary said and what argument he was dealing with at each stage. It is, in my view, a better approach to have regard to the evidence which can be identified as bearing on this issue and see, in the first instance at least, whether that evidence appears to point to a conclusion different from that at which the Lord Ordinary arrived.
Work on the platform on 6 July
For some time before 6 July, the gas conservation module, which was situated at a level above the 84-foot level had not been in operation and substantial maintenance work was in progress there. The gas conservation module contained equipment the purpose of which was to carry out further treatment of gas and extract further quantities of condensate. Operation of the platform without the gas conservation module was described as phase 1 operation: operation with it was described as phase 2. There was undisputed evidence from Lynch, the lead production operator at the start of the day shift on 6 July, that on the morning of that day Curtis, the offshore installation manager, telephoned him and asked him to switch over to condensate injection pump B and shut pump A down to make it available for planned maintenance. This must have occurred before Lynch left the control room at about 9.45 a.m. to prepare himself to leave the platform, as he did at about 11 a.m. Lynch accordingly instructed Grant, the phase 1 operator, to change the pumps over. There was undisputed, or at least relatively undisputed, evidence that the process of depressurising a pump would take in the region of five hours to complete: that period covers all stages from the starting of the replacement pump, the checking that the replacement pump was operating adequately, the switching off of the pump to be depressurised and the process of depressurising itself. By the time Lynch left the control room, pump B had been started and was running but pump A had not been shut down, isolated or depressurised. In the course of the shift, Smith, the lead maintenance hand on the day shift, brought to the control room an application for a permit for the planned maintenance on pump A: in accordance with normal practice, the requesting authority would have been White, the offshore maintenance manager. However, at about 5 to 6 p.m., Seddon, the onshore maintenance superintendent, spoke by telephone to White on the platform and advised him not to carry out the planned maintenance, which would have removed pump A from use for two to three weeks, but just to carry out certain work on a part of the pump called the Voith coupling. There was no evidence that any work was done on pump A, although the reclaimers maintained that it was a possibility, on the evidence, that some instrument work might have been done. The evidence of Clark, the night shift lead maintenance technician, was that he was told by Smith at hand over at the beginning of the night shift that pump A had been shut down and electrically isolated and the permit to work written out but not issued and that while he, Clark, intended to start work on the Voith coupling if circumstances allowed, in the event that was not done and the permit to work was never issued.
The evidence about the removal of PSV 504 depends to a substantial extent upon Rankin, the Score technician in charge of the valve refurbishments. There is controversy as to the reliability of his evidence in certain respects but as regards the actual steps which were taken there does not appear to be any particular challenge to what he said. According to Rankin, Smith told him on the morning of 6 July that PSV 504 would be available to work on at some time during the day, because pump A was to be shut down. Permits to work for the valve refurbishment programme had been written out in advance and kept in the Score container. Rankin obtained the application for the permit to work for PSV 504; it was a blue permit, a copy of which survives. The name of the person requesting the permit is illegible but it bears to have been approved by Curtis and signed by him at 7.40 a.m. The removal of PSV 504 could not proceed immediately because pump A still had to be depressurised and scaffolding had to be erected. Rankin’s evidence as to timing is disputed but for the moment it can be noted that Rankin and Sutton removed the valve with the assistance of a rigger, Rutherford, in the presence of the day shift phase 1 operator Grant. The valve was taken to the Score container by crane and work was done on it there. That work was completed by the end of the day shift and Rankin then went to see whether the valve could be replaced. What he did is a subject of controversy, to which reference will be made below.
Events on the night shift
As I have mentioned, there was no direct evidence of any work being done on pump A on the night shift. There was evidence from two instrument technicians, Cassidy and Young, who were on the night shift. Cassidy was working in the gas conservation module and he was in the instrument workshop at about 9.35 p.m. when four instrument technicians from the day shift who had been working overtime on the gas conservation module came in. Young who, according to his evidence, was working day shift plus overtime in the gas conservation module reached the instrument workshop about 10 to 10.
Apart from the instrument technicians, there was evidence from McGregor who carried out a small job of attaching a name plate to a compressor in C module. He finished that job at about 9 or 9.15 p.m. and had his permit signed off, by Vernon.
The control room operator on the night shift was Bollands. Again there is controversy about some parts of his evidence but other parts are apparently unchallenged. According to Bollands, he and Vernon were in the control room at about 9.45 p.m. when pump B tripped. Bollands reported the tripping to Richard, the night shift phase 1 operator, who was thought to be in module C at the time: he did this by radio. Vernon left the control room and went to the 68 foot level. Three to four minutes after the pump tripped, Bollands received a JCP alarm which he accepted and reported to Richard, saying that it was probably a high level alarm in the JT flash drum, that is, an alarm indicating a rise in the level of condensate in the drum. Bollands asked Richard whether he had unloaded the reciprocal compressors. There was evidence from the observers on the standby vessels that there was a substantial increase in flaring a few minutes before the explosion, which is consistent with unloading the reciprocal compressors. Vernon then came back to the control room and said that he could not get pump B started and indicated that there might be a lube oil problem and that he wanted to get pump A back into operation. Bollands asked whether the reciprocal compressors had been unloaded and Vernon said that they had. Bollands then called Clark on the tannoy. There is some controversy about the details of what took place thereafter but there was some communication either by phone or directly between Vernon and Clark after which Vernon again left the control room. Shortly thereafter, Bollands began to receive gas alarms. Grieve, the phase 2 operator on the night shift, overheard the message that the pump had tripped at the 68 foot level and went to see whether he could assist. He saw Vernon and Richard at the side of pump A furthest away from pump B and he was then aware that Richard had been called away by radio and last saw him heading up into C module. The explosion then occurred. The last pieces of evidence about Vernon’s movements come from Clark and McGregor. Clark was in the control room at the time of the explosion and left it and went to the bottom of the stairs, where he met Vernon who said that he was going to try to start one of the diesel pumps. Vernon then took breathing apparatus and, along with another man, a safety operator Carroll, disappeared into the smoke. He came back and saw McGregor and told him there was no chance of starting the pump: and was last seen descending the stair to the 68’ level.
Rankin’s evidence as to timing and dealing with his permit to work
As I have mentioned, Rankin obtained a blue cold work permit on the morning of 6 July 1988. He could not remove the PSV until the pump was depressurised and in the event the PSV was removed some time after lunch, at about 2 p.m. It would take about half an hour to remove the valve and take it to the Score container where the work occupied about two hours so that the work was not completed until some time near the end of the shift. Rankin then said that at about 6 p.m. he came back to the control room, with his permit, in order to return it at the end of the shift and see whether a crane could be made available to finish the job (which would involve extending the permit). Rankin said that he saw one person in the control room and that person was sitting at the lead production operator’s desk. Rankin did not know that person. He was told that there was no crane available and he, and the person at the desk, then suspended the permit. The precise terms of his evidence at this point are important and will be quoted later. Thereafter, Rankin went back to the Score container and spoke briefly to Sutton and then went to the accommodation module where he spoke to Smith. The only point which Smith raised was whether blind flanges had been fitted and Rankin said that they had.
The evidence of Bollands and Clark as to what precisely took place at about 9.45 p.m.
There is a difference between Bollands and Clark as to what precisely occurred at this point. As I have mentioned, Bollands called Clark on the tannoy. According to Bollands, Clark telephoned the control room and he and Vernon spoke on the phone. Clark then came to the control room. Vernon took a pink permit, with two isolation tags, from the rack of live permits and spoke to Clark about it. After that, Vernon left leaving Clark with the isolation tags, which he still had when the explosion occurred.
Clark, on the other hand, said that he was in the maintenance office with White when he heard the tannoy message at about 9.45 p.m. He contacted the control room and spoke on the telephone either to Vernon or Bollands. He was told that B pump had tripped and had not been restarted and that someone was wanted to look at it straight away. He was asked whether any work had been done on A pump and replied in the negative. It was then decided that A pump would be de-isolated and put back on, to give time for work on B pump. Clark then told White what had been decided and made a call over the tannoy for Savage, a day shift electrician who he thought would be the electrician most quickly available to carry out de-isolation. He then went to the control room, where he saw Bollands. Savage called from the accommodation module and after speaking to him, Clark decided that it would be quicker to get the other two electricians (who were working in the gas conservation module). Clark signed the electrical de-isolation tags which were on a table in the control room already signed by Vernon and not attached to a permit. Clark did not recall contacting or trying to contact Young. At this point, a gas alarm came up for C module east and Bollands accepted it and radioed to Richard. Clark had just finished signing the tags when a further gas alarm came up, followed by the explosion.
Conclusions on direct evidence
The direct evidence is fragmentary, but it does permit some reasonably clear conclusions. In the course of the day shift on 6 July, a permit for the planned maintenance work on pump A was prepared and approved and taken to the control room. Pump A was then electrically isolated and isolated and depressurised. There is, however, no evidence that any work was actually done on the pump. A permit for removal of the PSV was also applied for and approved and was issued and the work of removing and refurbishing the PSV was carried out but the PSV was not replaced. Vernon was in the control room at about 9.45 when pump B tripped. He went to the 68 foot level and attempted to restart pump B, with Richard, but was unable to do so. He then returned to the control room where he contacted Clark and a decision was taken to restart pump A. Vernon then returned to the 68 foot level while Clark prepared to arrange for electrical de-isolation of pump A. Vernon and Richard were together on the 68 foot level in the vicinity of pump B when there was a gas alarm received in the control room, following which a message was passed to Richard who proceeded to C module. There were then further gas alarms followed by the explosion. After the explosion, Vernon was seen at the level of C module where he intended to try to start diesel pumps.
Evidence of Bollands and Clark about events in the control room
There was a potentially important difference in the accounts given by Bollands and Clark as to what precisely took place in the control room after the B pump tripped. According to Bollands, there had been instrument technicians working on a condensate injection pump on the night shift. After the tripping of the pump, Vernon returned to the control room, as has been said above, and Bollands then called Clark on the tannoy. Bollands said that Vernon and Clark spoke by telephone, in the first instance, but that Clark then came to the control room. While he was there, Vernon took a pink permit from the rack of active permits and spoke to Clark about it. Vernon signed off the permit and the isolation tags and again left the control room, leaving the tags with Clark, who still had them when the explosion occurred.
Clark, as has also been said above, said that he visited all the modules after the handover and was in the maintenance office when he heard the tannoy. He contacted the control room and spoke to Vernon or Bollands and was told about the situation and was asked if any work had been done on a pump to which his reply was in the negative. It was decided that A pump would be de-isolated and put back on and Clark, having told White what had been decided, made a tannoy call for Savage. Clark said that he then went to the control room where he saw Bollands and where Savage spoke to him, following which he decided that it would be quicker to get other electricians. Clark said that he signed the de-isolation tags which were on the table in the control room already signed by Vernon and not attached to a permit. The first gas alarm came up and Bollands accepted it and radioed for Richard. Clark himself had just finished signing the tags when another gas alarm came up and then there was the explosion.
There is, therefore, no conflict of evidence between Bollands and Clark as to the fact that pump B tripped, Clark was contacted and Clark and Vernon decided to bring pump A back into circulation. There was also no dispute that steps were taken to deal with the electrical isolation of pump A, that Vernon signed isolation tags and that Clark was in the process of doing so when the explosion occurred. The significant differences between Bollands and Clark are as to whether Clark and Vernon actually met in the control room and as to whether there was a permit with isolation tags attached to it. The critical point about this piece of evidence is that if Bollands is right the planned maintenance permit, which was a hot work permit and therefore pink, must have been issued during the night shift. It is perhaps possible to say that Clark’s evidence does not directly contradict Bollands’ evidence on that point because Clark does not claim to have seen a permit at all and therefore there is nothing in his evidence which bears on whether there was a live permit and, if there was, what kind of permit it was. However Clark, at least, does not give any support to Bollands’ evidence on the point of the existence of a live pink permit. The position is complicated by the fact that Bollands was not cross-examined at all on the point, while Clark was extensively cross-examined. Not only that, it was strongly suggested to him that there was a permit which was a blue cold work permit which would, of course, have been more likely to be the permit for the PSV removal. If that were correct, of course, it would have a very material bearing on the question of whether Vernon knew about the PSV removal. In the circumstances as they are, however, it is by no means easy, and it was not easy for the Lord Ordinary, to make any definite judgment on the reliability of either Clark or Bollands on these points. The Lord Ordinary formed an impression that Clark was perhaps distancing himself, to some extent, from the events of the night, but that does not take matters very far. Looking at the evidence as a whole, and taking account of such scraps of evidence about the activities of instrument technicians as there are, it is, in my view, difficult to reach a conclusion from the direct evidence as to what happened in the control room with any degree of probability, beyond the conclusion that Vernon and Clark were concerned in signing off isolation tags for the pump, which is not surprising considering what had happened during the day shift. As the reclaimers argued, the possibility that there was a live planned maintenance permit and that instrument technicians were working under it in the early part of the night shift cannot be altogether excluded, although it does, in my view, seem rather on the improbable side.
Evidence of Rankin
The outline of Rankin’s evidence about dealing with the PSV has been given above. The critical questions which arise about his evidence are firstly in connection with the timing of the carrying out of the work on PSV 504 and secondly, as to what Rankin did with his permit to work at the end of the day shift. Rankin’s evidence was that he was told that the PSV would be available but that there was a delay because of depressurising the pump and getting scaffolding erected and that the PSV was removed by Rankin and Sutton, with Rutherford and Grant, some time after lunch, at about 2 p.m. If so, allowing half an hour to remove the valve and two hours to do the work, the valve would be ready for reinstallation somewhere about 4.30 to 5 p.m. Rankin’s evidence was that he came back to the control room with his permit at about 6 p.m. in order to return the permit and see if a crane could be made available to finish the job. He said that he saw one person in the control room and that person was sitting at the lead production operator’s desk. Rankin did not know that person. Rankin was told that no crane was available and he and the person at the desk then suspended the permit. Rankin then returned to the Score container and spoke briefly with Sutton and then went to the accommodation module where he met Smith and explained the situation to him: Smith only asked whether blind flanges had been fitted to the pipework.
There is some evidence which goes some way to support Rankin’s evidence about timing. Rankin had visited the control room in the morning when Lynch and Flook were at the lead production operator’s desk. Flook took over from Lynch, as stand-in lead production operator for the day shift when Lynch went to leave the platform. Lynch, according to his evidence, told Rankin to see Smith in order to get the paperwork for the valve refurbishment in motion. As has been mentioned a copy of the permit showing approval by Curtis survived. According to Bollands, Grant was involved in the process isolations and reported to Flook, who might even have checked the position himself, and Flook issued the permit to work to Rankin. In these circumstances, Flook must have been aware of the removal of the PSV and Rankin and Flook had met in that connection. There was also evidence from Clark that the removal took place after lunch. The reclaimers argued that, notwithstanding that evidence, there was room for the possibility that the timing was out and that the removal and completion of the work might have been carried out by, say, 4 p.m. but even if that argument has some substance, there can be little doubt that before the end of the shift Rankin was looking for a crane to try to get the valve replaced. It was, after all, the last job which he had to do and completion would have enabled him to leave the platform at the next opportunity.
There is also, in my view, a powerful argument that Rankin must have seen a lead production operator and suspended his permit at the conclusion of the day shift. The alternatives are either that he did not suspend the permit at all or that he saw someone other than a lead production operator. Both of these are, however, improbable. If Rankin had not suspended his permit, or done something with it, the permit would have remained outstanding at the end of the day shift and its existence would have been obvious to the incoming lead production operator, Vernon. As regards the other possibility, the evidence was, overall, clear that only a lead production operator would deal with permits, whether in relation to issuing, suspending or extending them.
On the other hand, there are aspects of Rankin’s evidence which do not fit with what might be expected. He said that he saw only one person in the control room, which would appear unlikely, because, if he was seeking to extend or suspend his permit at the end of the day shift, it is probable that there would have been other persons present dealing with other permits and, in any event, it is not likely, although not impossible, that he would have seen only the lead production operator in the control room. The control room operator would, as his job title implies, normally be in the control room. There is what might be described as a side issue in the evidence concerning Bollands’ statement that after coming into the control room to take up his shift he went and sat and had a cup of tea at a table behind some of the panels where his presence might not have been obvious to someone coming into the control room: but even if that is a possibility, the doubt remains as to whether there would not, in all probability, have been other persons seeking to have permits dealt with at around the time Rankin says that he went to the control room. Because of the work in the gas conservation module, there were a number of permits to be extended, suspended or cancelled.
It has to be said also that the Lord Ordinary was not particularly impressed by Rankin as a witness. The reasons for that view of his evidence are not spelt out in great detail. It is, however, clear even from the printed notes of evidence that Rankin found difficulty in taking in and responding to questions and that quite often his answers were not directed to the point of the question. When it comes to the particular matter of what he did with his permit, his answers are curious. When asked about what happened to the permit he said, "My understanding is that jointly we suspended the permit" (11485). He could not remember visiting the control room at all, according to his evidence, at that point. Of course, Rankin like the other witnesses had been through a terrible experience and was being asked to remember matters which probably did not seem significant or out of the way at the time when they occurred. Nevertheless the answers which Rankin actually gave in regard to this critical point of the proceedings are such as to indicate that the Lord Ordinary did have a basis for his reservations about Rankin as a witness and accordingly to leave a question mark over this vital chapter of his evidence.
Certainty is, of course, impossible but despite the criticisms of Rankin's evidence it can, I think, be said to be probable that he did go to the control room and suspend his permit, because, if he had not done so, the fact that the permit was outstanding would have become obvious. If so, the probability is also that he saw a lead production operator: but the evidence so far considered does not, I think, justify a conclusion that it must have been Vernon that he saw. It is at least a possibility that Rankin went to the control room once the work on the valve had been done, to see about reinstating the valve, and that that visit took place before Vernon came on shift. It is suggested that if Rankin had seen Flook, he would have known him, having seen him earlier in the day, and there is force in that suggestion, but, for the reasons previously mentioned, there are question marks over Rankin’s evidence, and I do not think that his evidence alone would justify the conclusion that he must have seen Vernon at that time. The question whether Vernon was aware of the PSV work has, in my view, to be considered in the light of all the circumstances.
Vernon’s means of knowledge
In the course of discussion of the evidence, a total of eight ways have been suggested by which Vernon would or might have come to know that PSV 504 had been removed. These are (1) that he would be aware because he himself suspended Rankin’s permit; (2) that he would be told by Flook at the handover; (3) that he would find out about the removal by examining suspended permits in the control room; (4) that he would be told by Richard at the time of carrying out the jagging operation, Richard having learnt about it at handover from Grant and Grant’s log; (5) that he would learn about it as a result of a necessary adjustment of PCVs 1000 A and B when the reciprocal compressors were recycled; (6) that he would learn about it when resetting the manual isolation valve adjacent to PSV 504 before attempting to repressurise pump A; (7) that he would know as a result of his walk around, and in particular because of the presence of scaffolding; and (8) that he would know as a result of suspending McGregor’s permit for the work of fixing a name plate to a compressor.
I have already dealt with the evidence relating to the suspension of Rankin’s permit. Of the remaining possibilities, the two most important are that Rankin would be told by Flook at the time of the handover and that he would be told by Richard at the time of jagging, Richard having been informed by Grant of what had happened during the day shift. There was a considerable amount of evidence about the system of handovers from operators on one shift to operators on the succeeding shift, but I do not think it necessary to look at that evidence in detail. In some of the evidence, there was reference to the use of a more or less formal log or other written notes. Whatever the precise method, there can be little doubt that, broadly speaking, the system of handover, whether conducted specifically by reference to a log or by reference to notes or otherwise, was intended to convey from the outgoing to the incoming operator all material information which the incoming operator ought to know in order to do his job properly and safely. Looking at the matter broadly, it would appear almost obvious that the absence of a PSV would be a matter which both the lead production operator and the phase 1 operator, in whose area the PSV was situated, would need to know. On the other hand, the reclaimers submitted that if pump A had, as they put it, been "handed over to maintenance" then it might not be surprising if particular emphasis was not put on the removal of the PSV. If the full planned maintenance had been undertaken, pump A would have been out of commission and unavailable to the production side for two to three weeks. Even if only the Voith coupling work had been done, the pump would still have been unavailable for two or three days. In these circumstances, it was submitted, it might well be that both Flook, in handing over to Vernon, and Grant, in handing over to Richard, might have thought that it was sufficient to point out that pump A had been handed over to maintenance. In connection with this argument, some attention was given to passages in the evidence suggesting that handing over to maintenance might be given a more or less definite meaning. Looking at the matter as a whole, it seems to me difficult to say that there was any clear evidence of such a specific meaning being given to the phrase. There are references to "handing over to maintenance" and "getting back from maintenance" (e.g. Bollands at 63/4469) but there is no clear evidence that there was a recognised status of "handed over to maintenance". The possibility remains, however, that the complication of the two pieces of work, the planned maintenance and the valve refurbishment, involving two permits, could lead to the removal of the PSV being overlooked or not emphasised at the time of handover. Some such explanation could account for both Vernon and Richards being unaware of the absence of the PSV. The reclaimers’ argument would, of course, become stronger if it were accepted that the planned maintenance permit was actually live, as Bollands evidence may suggest, but for the reasons given above it seems difficult to come to any definite conclusion on that point. As regards these two routes by which Vernon might have known, therefore, it seems difficult to go beyond saying that the probabilities would tend to favour reference having been made to the absence of the PSV but that the possibility cannot be excluded that it was not.
The other potential routes of information can, I think, be dealt with relatively quickly. The evidence does not, I think, support the proposition that there was any duty on the lead maintenance operator to examine suspended permits as a matter of routine, apart from special circumstances (see Henderson at 11624 and Lynch at 11180). The handover system should have ensured that he would receive such information about the status of the plant as was necessary but beyond that the evidence does not really go. The evidence about the adjustment of PCVs 1000 A and B suggests that an adjustment would be necessary if the reciprocating compressors were to be recycled but does not show that the carrying out of that adjustment would necessarily place anyone in a position to observe that PSV 504 had been removed. The evidence, similarly, does not suggest that the absence of the PSV, which was at some height above the deck in module C, would necessarily be observed in the course of a walk around if attention was not particularly drawn to it, and the presence of scaffolding would not necessarily be remarkable. As regards McGregor’s permit, it is true that, by the book, the lead production operator should have examined the site of the work; but given that the work consisted in attaching a name plate to a compressor it would perhaps not be surprising if the matter was not done precisely by the book.
Another, in my view minor, point concerns the electrical isolation tags. It was suggested that Clark’s evidence was that the maintenance permit, which would have been pink, was never issued and therefore that any red tags left for Clark to sign would have related to Rankin’s blue permit; and further that Vernon would have been aware of Rankin’s permit. I do not see how significant weight can be given to that suggestion, in view of the fact that Bollands’ uncross-examined evidence was that the tags which Clark signed related to a pink permit. There are, as have already been noted, difficulties over the evidence of Bollands and Clark, but this point is not, in my view, material in itself.
That leaves only the question of the setting of the manual isolation valve which was immediately adjacent to the site of PSV 504 and between the pump and the PSV. That valve would be closed as part of the procedure of shutting down the pump and would have to be reopened before the pump was restarted, to allow the PSV to perform its function. There was evidence about the process of bringing a pump back after isolation and depressurisation, in particular from the witness T.A. Henderson, a very experienced operator. There was some dispute as to how the evidence should be read, but looking to Henderson’s evidence as a whole it appears to me that he was envisaging a regular sequence of operations according to which the manual isolation valve would be opened, to make the relief line available, before repressurising the pump, not merely before commencing to operate it (e.g. at 79/11678). The counter argument to that is that, given the circumstances at the time, Vernon and Richard might very well have started the jagging operation at the 68 foot level before one or other of them went to reopen the manual isolation valve.
There is further, and very important, evidence which bears on the likelihood that Vernon might have repressurised the pump even though he knew that the PSV had been removed. There was substantial, and I think uncontradicted, evidence that it would have been completely contrary to all good practice to attempt to repressurise a pump in the absence of the PSV and that no experienced operator would have done so. Henderson, for example, said that "it just would not be done" and when asked whether Mr. Vernon would do it answered "No". He repeated that he would not act in that way even if he had been encountering substantial problems with the plant and that he would have shut the gas plant down (79/11817). Henderson affirmed that evidence very positively in cross-examination. There was evidence to the same effect from Bollands and from Clark and from another operator, J. Murray. Murray said that he would not run a pump at any time without the PSV and with a blind flange on the pipework and that if a pump was to be run at all without a PSV it would be in some dire circumstances and would be run under the approval of the offshore installation manager (105/15144). The Lord Ordinary made findings in accordance with that evidence, in his opinion at page 720. With regard to the suggestion that an operator might have decided to run a pump relying on the presence of a blind flange, when the PSV had been removed, it also has to be borne in mind that there was evidence, in particular from Henderson, that when equipment was being brought back into service after work which had involved breaking into the pressure envelope, an elaborate procedure of checking for gas leaks was followed.
The Lord Ordinary was understandably reluctant to come to definite conclusions and even to exclude possibilities where the evidence was not absolutely certain. In one or two respects it is, I think, possible to go further than the Lord Ordinary did in reaching conclusions on particular points. The Lord Ordinary was prepared to keep open the possibility that Vernon did become aware earlier in the shift that PSV 504 had been removed but forgot or allowed the matter to slip his mind. In my view, that is not acceptable. I find it impossible to envisage that Vernon would not have remembered that the PSV had been removed when he came to consider bringing the pump back into service. Again, as I have already indicated, it is not, in my view, rally a serious possibility that Rankin’s permit was not suspended: if it had been suspended, it would have been very obvious. As regards the remainder of the evidence in this chapter of the case, I accept that there is evidence which points, with some considerable degree of probability to the conclusion that Vernon ought to have known of the removal of the PSV, and that, if he did not know, there must have been a serious failure in the operation of the system for transmitting important information about the status of equipment on the platform. On the other hand, there is weighty evidence pointing to the conclusion that Vernon would not have attempted to restart the pump if he had known that the PSV was missing. I do not think that it is possible to reach a conclusion one way or the other on this chapter of the evidence, considered by itself. What has to be done, in my view, is to consider the evidence in this chapter of the case along with the rest of the evidence bearing on the cause of the disaster.
12. Two particular issues
This section of the opinion is concerned with two matters about which separate arguments were presented and which are to some extent related to one another. The respondents submitted that it was possible to reach conclusions in respect of these matters, taken by themselves, which would destroy or at least have a very significant impact on the reclaimers’ circumstantial case.
The word "jagging" is used to describe the operation of slowly introducing hydrocarbon into a pump in order to raise the pump to its operating pressure. If the full operating pressure were introduced into a pump all at once and suddenly, there would be a risk of damage to the mechanism of the pump. The operation of jagging was designed to avoid the risk of such damage.
In the case of the condensate injection pumps, jagging was carried out by operating the suction side gas operated valve (GOV). That valve would have been closed as part of the process of isolating and depressurising a pump. The operating pressure on the upstream side of the suction side GOV was approximately 600 psi and the operation of jagging was designed to bring the pump pressure up to that level. It was also possible to make use of an equalisation line to repressurise the discharge side of a condensate injection pump when the other condensate injection pump was running.
As has been explained previously, suction and discharge GOVs of the condensate injection pumps were pneumatically operated ball valves. In the middle of the valve there was a ball through which there was a bore which would permit the passage of the flow of hydrocarbon condensate. When the bore was in line and flush with the upstream and downstream pipework, the valve was fully opened. When the valve was fully closed, the bore was at right angles with the line of the upstream and downstream pipework. Accordingly, the ball required to turn through 90° to travel from a fully open position to a fully closed position. Each valve was operated by a pneumatic actuator, the air supply for which was supplied from the instrument air system. The air caused a piston within the actuator to move and the movement of the piston compressed a spring on the opposite side of the actuator. The rotation of the ball was brought about by a Scotch yoke mechanism which converted the linear motion of the piston into the rotational motion necessary to turn the ball. If no air pressure was introduced into the actuator, the spring would hold the valve in the fully closed position. The introduction of air into the actuator would cause the piston to move and the ball to turn, thus opening the valve. According to the evidence of Mr. Wottge, the valve required about eight seconds to move from the closed to the open position; other evidence may perhaps suggest about ten seconds. The valve would remain in the open position so long as there was air pressure in the actuator to hold down the spring but when the air pressure was vented, the spring would close the valve. The evidence suggests that the closing of the valve was slightly quicker than the opening. The opening and closing of the valve was regulated by a push/pull button situated on a stanchion situated a few feet to the south of the condensate injection pump. Pulling the button would introduce air into the actuator and open the valve and pushing the button closed would cause air to be vented and close the valve. If the pump had not been electrically isolated, the button would remain in the out or open position after the valve had been opened until it was pushed to close it again. If, however, the pump had been electrically isolated and racked out, as was probably the case at the time of the disaster, the push/pull button could still be used to open the valve but the valve would only remain open so long as the button was held in the open position: as soon as it was released, the button would return to the closed position and the valve would close.
The particular point discussed in this chapter of the argument was the length of time required to carry out the jagging operation. This could be significant in relation to the evidence of what Vernon was doing in the period immediately before the disaster. Further, the reclaimers’ theory of the accident requires that there should be two releases of hydrocarbon separated by an interval, so as to fit in with the evidence about the occurrence of the gas alarms, discussed later in this opinion. It could therefore be significant if the evidence suggested that jagging was normally carried out as a single process. In the course of the argument on this point, the respondents also made reference to a contention that, as a result of the various experts’ calculations of the amount of hydrocarbon release required to permit the formation of a flammable mass sufficient to account for the explosion, it would be necessary for Vernon to have held the push/pull switch open for a period of 20 seconds or more, if the theory was to be acceptable.
It seems to me that the best way of approaching this part of the question is to set out the evidence given about the jagging operation. I have already mentioned the evidence of Mr. Wottge. In cross-examination, (47/7328) he said
"Well the pneumatic operator provides some control, just opens slowly to a certain extent. When I say it slowly opens in the period of I think about eight seconds, which is not a slam open. And in that process the valve goes from shut to open and it doesn’t go to a hundred percent open, it opens, you probably have one, one and a half, maybe a two second period where the valve throttles. In other words, it allows liquid into the pump at certainly a much lower rate than with the valve open.
Q. But it would be very difficult, if at all possible, to open the GOVs in a controlled way by pulling the button?
A. I understand the practice was they would open the button and again push the button to decrease the opening speed, or in that manner open it slowly".
It might perhaps be observed that three pages earlier Mr. Wottge had been asked to give his opinion as to the appropriate means of pressuring up the pump and said that he had given it some thought and would have expected guidance from the manufacturer but had been unable to find any and that shortly thereafter the questioner observed to him that he (the questioner) was not interested in how operators might have gone about things but was asking Mr. Wottge to approach the question of the appropriate means of pressurising the pump as an engineer. There may, therefore, be a question as to how far Mr. Wottge was in a position to, or was attempting to, give an account of actual practice.
The witness Erland Grieve, the Phase 2 process operator, was asked (61/9245) what steps he would take if one pump tripped and the other pump was available but not pressured up and said:
"To restart a pump that had no pressure on it generally would mean just going to the GOV and, as they call it, jagging the GOV by introducing an air signal on to it so that the valve opened slightly and allowed pressure in on the suction side of the pump, and then letting it go again and letting the valve close, and then doing that two or three times and slowly increasing the pressure on the pump".
At 9246, he said:
"You would just take the push button relating to the suction GOV and pull it out, allowing an air signal to go to the valve which would then begin to move to the open position.
Q. Now would you hear anything when that was done?
A. You would hear the air go into the valve and you would be able to see it on the valve itself, starting to lift off its seat, and you would hear the pressure going into the pump as the valve opened.
Q. Then what would you do?
A. As soon as you heard pressure going into the pumps you would let the push-pull button go and shut the valve again.
Q. And then what would you do?
A. You would go through that procedure again slowly introducing pressure into the pump".
He then said that the suction and discharge pressure gauges were available to be watched and that the operation would generally be done bringing the pressure up in stages "under the best control that you could given the fact that you were using the GOV to introduce the pressure" and he expressed the view that to pressurise a pump in that manner would probably take between one to three minutes, something like that.
The witness T.A. Henderson, was also asked to describe the process of jagging. Henderson approached the question of repressurisation of a pump initially on the assumption that the pump was being brought back into service after some operation such as maintenance which would have involved opening up the pressure envelope in some way. What he initially described, therefore, was a process of gradually introducing pressure into the pump using nitrogen or a substance referred to as "snoop" while checking the parts of the system which had been opened to ensure that there was no leak. He said (79/11766) that after that stage:
"So you would then start jagging the suction valve to allow hydrocarbons into the pump to slowly pressurise it up; at the same time the operators would also be sniffing the holes in the tape to see if there was any hydrocarbon leaks".
He went on, in the following pages, to describe the process of operating the push/pull button to introduce pressure gradually into the pump. At 11770, still under reference to the reintroduction of pressure after maintenance or some similar operation, he was asked about the time which it would take and he said:
"If all these joints had been disturbed, you would take your time in doing it in case you had effectively had any leaks; so I would say 15 minutes".
At 11771 he said, with reference to the jagging exercise,
"When you jagged you only moved the valve actually from its seat and then you closed it again, you never let the valve go fully open, to keep the control of introducing air pressure you gave it an open signal and closed it when you heard the pressure going through and watching your gauges. So you did not move the valve very far into the open position actually".
At 11780 he was asked about the operation of repressurising when there had not been any interference with the pressure envelope and said:
"You would bring the pressure up more quickly, you would leave the valve to open more fully, to bring the pressure up more quickly".
He then expressed the view that you could pressurise it in maybe four/five minutes.
The only other witness who gave detailed evidence about the jagging operation was Murray (105/15041). A considerable amount of time and, in my view, a great deal more heat than was justified, was expended by the respondents on a detailed examination of Murray’s evidence. The origin of this was, so far as I could gather, a suggestion that Murray’s evidence had originally been led or given before the Lord Ordinary in a misleading way and that it had been necessary for it to be corrected. In his evidence in chief, Murray began by saying:
"The suction GOV obviously was operated by compressed air and to pressure up the pump through suction, through discharge, up to the, possibly, discharge GOV and the PSV, the suction GOV would be operated just fractionally - that is, they would pull the latch button, the push/pull button, until the indicator on the GOV showed the valve was just open. This allowed some condensate into the pump, through the pump and through the discharge. They would then push the push/pull to snap the GOV back shut. This would take maybe three times but no more than four to pressure the pump up".
He was then asked about the timescale for this operation and the evidence was interrupted for some time by a wholly spurious objection by counsel for the respondents based on a suggestion that in some way the witness’s evidence sought to contradict earlier evidence which had, as had been put, been led without challenge. When the witness’s evidence was eventually resumed, he said (15059):
"It would only be a few seconds to do three or possibly four jags, I wouldn’t give it any more than 20 to 30 seconds".
It was this suggested timescale which obviously alarmed the respondents. Murray was cross-examined at some length by reference to evidence that he had given at the Cullen Inquiry. What emerged was that Murray’s own practice in repressurising a pump had not been to carry out the entire operation by using the suction GOV but to make use of the equalising line. The equalising line was a small bore pipe which made a connection between the discharge pipes of the two pumps. There were control valves on the equalising line, which were necessary in order to enable each of the pumps to be isolated. If one of the pumps was running, pressure could be introduced into the discharge side of the other by manipulating the valves on the equalising line. It appeared that Murray had been in the habit of using this method and accordingly that when he required to operate the GOV on the suction side of a pump, he only required to use it to repressurise the suction side. It emerged that Murray had never himself utilised the method of pressurising the whole pump from the GOV on the suction side and the final upshot of his evidence, at 15169, was that it would take about four minutes to pressurise the discharge side by the method he used and about one minute to pressurise the suction side, the discharge side having been repressurised already. There was nothing misleading about Murray’s evidence, although it was necessary to explore exactly what he was saying
There was also evidence in general terms from a witness Sneddon as to the need to carry out the operation slowly. The only other significant evidence on this part of the dispute came from Dr. Richardson, one of the expert witnesses, who said that, in theory, having regard to the capacity of the pump and associated piping and the pressure on the suction side, if the suction GOV were fully opened it would take something in the region of one or one and a half to two seconds to fully repressurise the pump.
The respondents’ submissions were directed to establishing four propositions namely:
(1) That the jagging operation was in practice carried out as a continuous
(2) That the jagging operation was carried out by operators in minutes (rather than
seconds) and that there was no suggestion that it could be carried out as quickly or as slowly as the operator thought fit:
(3) That there was nothing in the evidence to suggest that, after full
repressurisation of a pump, the operator might hold open the push/pull button for a period of 20 seconds or longer.
(4) That there was nothing in the evidence to indicate that the process of jagging
might be interrupted for a period of, for example, two minutes.
The reclaimers on the other hand submitted that the Lord Ordinary had been entitled to proceed on the basis that the range of times for repressurisation of a pump on which no maintenance had been carried out stretched from 20 seconds, for a partially filled pump, to five minutes and that while there was nothing in the evidence to indicate that the operation of repressurising might be interrupted there was no reason why it should not be if circumstances required it and no reason why it should necessarily take as long as the several minutes indicated by some of the operators in the evidence.
In my view, very little can really be made of the evidence on this topic, viewed as a separate topic. Looking at the nature of the operation, there would seem to be no reason why it should always be carried out at a uniform speed. The opening and closing times of the valve must have provided some constraint upon the intervals at which a series of jags could be performed, but we were not referred to any evidence of any attempt to calculate the effect of that factor. Subject to the constraint, however, of the nature of the valve and its operation it would seem fairly obvious that the operation might be performed somewhat more quickly or somewhat more slowly, as circumstances dictated. What can, I think, be said is that the respondents’ attempt to argue that the description of the operation given by the operators pointed to or required that the operation should have been carried out continuously and without interruption seeks to put much too much weight on the evidence that was given. It is, I think, fairly clear that in the ordinary case the operator would, having started the operation, continue it until it was completed but equally, should there have been any reason to interrupt it, there is nothing to indicate that such an interruption would not have been permissible. It is true that there is nothing in the evidence to indicate any reason why an operator should continue to introduce pressure by holding open a push/pull button after a pump had been fully repressurised. This, however, leads on to matters which, in my view, should properly be considered in the context of the whole evidence about the likely occurrences on the night of the disaster.
(b) The evidence of Grieve
In this part of the argument, the respondents sought to submit that the Lord Ordinary had erred in law in determining that he was entitled to hold that Vernon had jagged condensate injection pump A shortly before the explosion when, as they submitted, that critical question had not been directly addressed with Grieve. They also submitted that the Lord Ordinary erred in law in rejecting a submission by them that the onus was on the pursuers to address with Mr. Grieve the critical question whether or not Vernon jagged the condensate injection pump, or could have done so, while Grieve was present at the 68 foot level. Grieve, according to his evidence, was at the 68 foot level for two to three minutes before the explosion and it was submitted, and indeed accepted by the Lord Ordinary, that he must have been there when the second stage, at least, of the jagging which, on the pursuers’ theory, led to the explosion took place. The essence of the respondents’ submission was that the pursuers should have asked Grieve directly whether he saw Vernon jagging the pump and whether Vernon could have done so without him seeing it, and that their failure to ask these questions amounted to a failure to adduce evidence on a critical issue, with the result that that issue must be determined against them.
I have already set out the substance of Grieve’s evidence, in an earlier section of this opinion but, even at the cost of some repetition, I think that the best way to approach this is to set out how Grieve’s evidence was given. Grieve was an experienced operator and was phase 2 operator on the night of the disaster. Because phase 2 was not in operation, he had relatively little direct responsibility. He heard about the tripping of the pump and went to the 68 foot level to see if he could be of assistance. He entered the 68 foot level from the eastern side and ducked under the JT flash drum. He was then asked (61/9284):
"Q. Now, what then did you do?
A. Well, at that point I was in the area of the injection pumps where I would have walked past the control panel and up to B injection pump. At that point I was aware of Bob Richard and Bob Vernon being in the area on the west side of the B injection pump".
He was then asked some questions about the precise position of Vernon and Richard and said that they were on the walkway on the corner of B injection pump when he was first aware of them and, a little later said:
"I can’t say exact distances between them, all I can say is that they were within the area between the walkway on the west side and the foot of the GOV area on the south side".
Grieve said that he was aware of coming in past the control panel and being aware that it was showing a high level in the flash drum and continued:
"I can’t recall whether it was at that point or whether I approached this area where, not exactly a form of conversation, but say, sign language or shouting something at each other, we decided that we would go for a run on B pump and I took it it was then…the thing I am not sure of is whether I reset it…
Q. Well sorry, carry on, I interrupted you?
A. Whether I reset the GOVs on the way past or whether they had already been reset. But on that day, under the area beside the control panel on B pump we reset the GOVs on B condensate injection pump and at that point we pressed the start button for a start on the motor for B condensate injection pump".
He was then asked some further questions about what he observed at the control panel and at 9289 he was again asked about the position of Richard and Vernon and replied:
"Just basically a recollection of them being in this area on the west side of B pump".
And later again he said:
"Because they were in the vicinity of the GOVs and I took it with the pump being shut down that there would be an attempt made to start it. As I said, some form of communication was made that they wanted to start B pump".
Grieve further explained that an attempt was made to start B pump and that, as he took it, either Richard or Vernon must have reset the GOVs to make an attempt to start it possible. Grieve himself went to the local control panel and pressed the start button and said that he watched the motor turn, took his hand off the button and the motor stopped. He then said that after that he made his way down to check the torque converter to make sure it was to manual. The examination continued (9293):
"Q. What about Mr. Vernon and Mr. Richard did you see either of them at that stage?
A. As I said I was aware of Bob Richard being called away and going up into C module.
Q. In relation to the time that you pressed the start button for B pump when you say the call came for Mr. Richard?
A. Well it was shortly before, or immediately after, I can’t say but I would have said shortly before because basically all that happened after starting it was that I walked round the back and by that time Mr. Richard was going up to C module".
Grieve went on to explain again that he had been aware of Richard answering his radio and going up stairs but not what the call was, that he could not add any detail and that there was nothing that drew his attention particularly to Richard’s departure. He further said that he could not say that he was aware of any great period of time elapsing between the call for Richard and the explosion, only a minute or so. At 9296 he said that after he made his way round the motor to check the speed controller:
"Well that’s when the explosion went off. I never had a chance to check it".
There followed some further questions about the route taken to reach the speed controller and there is then, at 9299, a critical passage in the evidence, which is as follows:
"Q. What about Mr. Vernon. Where was he?
A. The recollection I have got of Mr. Vernon was he was in the vicinity of the GOVs here.
Q. You indicated by reference to the pointer on the schematic - can you do that again - where your recollection is with regard to the location of Mr. Vernon?
A. On the south side of the pumps. My last recollection of having seen him was between the GOVs on A and B pumps.
Q. When you say "between", can you be a bit more specific?
A. No not really. I was aware of him in that area. I can’t say whether he was on his way back round here to the local control panel or another reset to start B pump. I’ve got no idea to what he was actually doing in the area.
Q. Can I just ask you in this way then, what then is your last recollection of where you last saw Mr. Vernon before the explosion?
A. In that area I told you. The last recollection I have got seeing him is between the two GOVs and the south side of the room.
Q. When you talk about the two GOVs what do you mean?
A. The two sets of GOVs in between A and B pump.
Q. Where were you when you saw Mr. Vernon in that area.
A. I can’t recall. As I said that’s the last area that I was aware of him being in, when I looked at him when I walked round the panel. I’m sorry, I can’t help any further than that.
Q. With regard to the other condensate injection pump, pump A, what was your state of knowledge regarding that pump at that time when you were on the 68 foot level?
A. As far as I was aware on that pump that it was on standby available to be run".
On the following day, Grieve was taken over some of the same ground again but added little or nothing to the passages above quoted except that on a number of occasions he repeated that he had difficulty in remembering details of what had happened. At 9323, he explained that he went down to the west side of the pump to check out the speed controller, as has been quoted above, and that as he arrived on the west side of the pump:
"There was an explosion which I felt was directly above where we were and at the time I fell to my knees, the lights went out and there was dust falling in the area".
The respondents submitted that the Lord Ordinary, in particular at pages 822 to 841 and 881, had correctly identified a critical issue as being whether Vernon could have repressurised pump A without Grieve seeing him. They submitted however that the Lord Ordinary had been wrong to regard the fact that no direct questions had been asked of Grieve on this issue as neutral. It was submitted that in view of the time intervals spoken to, the jagging operation must have been continuing at the time Grieve was there. Accordingly, the reclaimers’ failure to question Grieve about Vernon’s actions amounted to a failure to adduce evidence on a critical issue and consequently that issue should be determined against them, with the result that it should be held that Vernon could not have jagged the pump as the theory demanded. In support of this submission, reliance was placed on passages from Alison: Practice 505, Dickson on Evidence, Walkers on Evidence and the Stair Memorial Encyclopaedia vol. 10.
In my view, this argument on behalf of the respondents is misconceived. Looking to the examination of Grieve which has been quoted, the position is, in my view, that Grieve was correctly invited to say what he could, in answer to non-leading questions, about what happened while he was on the 68 foot level. He explained what he could and indicated repeatedly that he could say no more. It is true that further questions might have been asked but everything in the evidence which was led, in my view, clearly points to the conclusion that it was unlikely that any additional information would have been forthcoming. In such circumstances, there is some discretion for counsel to exercise in deciding how far it is necessary to ask additional questions. If, however, the respondents felt that there was any obscurity in the evidence or that some additional material might have been obtained, it was open to them to ask the appropriate questions. Whatever may be the situation in which the questioning does not cover the full area about which a witness might be able to speak, in the present situation it seems to me that the Lord Ordinary was perfectly correct in holding that the fact that Grieve was not directly asked whether or not he saw Vernon jagging was neutral.
In any event, I do not think that the authorities quoted by the respondents support the proposition which they sought to draw from them. What the authorities support, in my view, is that if a party fails to bring forward a witness who can speak to a particular matter, there may be room for an adverse inference to be drawn. None of the authorities, however, do more than say that that is a possibility, but one which must depend on the whole evidential situation in the case. It has never been the law that a party has to lead every conceivable witness on a matter of fact at the risk of having adverse inferences drawn against him if he does not. The respondents founded on the fact that the reclaimers’ case was a circumstantial one, but in my view that does not advance the argument at all. All depends, in my view, on the strength of the case, not on its nature. In any event, the authorities seem to me to be saying that an adverse inference may be drawn where a witness is not adduced and by that to mean where the witness is not led. If a party has led a potentially damaging witness, and the other side do not take advantage of the witness’s presence to ask the damaging questions, perhaps for fear of getting a wrong answer, which could be damaging to them, then again it seems to me that the matter is entirely neutral and none of the authorities suggest, to my mind, the contrary. In these circumstances, this whole argument, as a separate argument, should be rejected.
There remains, of course, the question as to how Grieve’s evidence should be taken into account and what weight it should have when the whole circumstantial case supporting the reclaimers’ theory falls to be considered. The substance of his evidence has been set out already, and I do not propose to repeat it. It can, I think, fairly be said that it could be expected that Grieve would have had an opportunity to see Vernon jag pump A, if he did so. I do not, however, read Grieve’s evidence as purporting to give a full or perfect account of everything that happened at the 68 foot level at the material time. Grieve himself was participating in the attempt to restart pump B. He was aware, from time to time, of what Vernon and Richard were doing, but it does not seem to me that his evidence is so precise as to require the conclusion that Vernon cannot have jagged pump B because Grieve does not report seeing him do so.
13. Gas Dispersion
In holding that the pursuers had established their case, the Lord Ordinary accepted that it had been proved that releases of condensate from the site of PSV 504 could have caused the pattern of alarms which was reported by Bollands and given rise to the initial explosion. That conclusion was challenged by the respondents on grounds which can be summarised under four heads, namely,
(1) An escape of hydrocarbon from the site of PSV 504 could not account for the initial explosion because such an escape could not have caused the alarm pattern which was seen; the expert evidence did not support, but was inconsistent with and fatal to, the theory; the necessary flammable mass could only be produced by holding open the push-pull button at the condensate injection pump for more than six seconds; and account had not been taken of scaffolding in the vicinity of the site of PSV 504.
(2) It was not established that a release of hydrocarbon at PSV 504 would behave as the pursuers’ theory requires because the assumptions on which the expert evidence on behalf of the reclaimer was based had never been established and could not be relied on.
(3) The pattern of events did not fit with the way in which the operation of jagging a pump was normally performed and the time that would be required for such an operation.
(4) The necessary jagging could not have occurred without the knowledge of Grieve who was present in the 68 foot level at the material time.
Heads (3) and (4) of these objections have been discussed in the preceding part of this opinion. Heads (1) and (2) are closely related and essentially resolve themselves into two points. The first is that the expert evidence led in support of the pursuers’ theory treated a hypothetical escape of hydrocarbons as having a molecular weight in the region of 42, more or less equivalent to that of propane, but that it had not been proved that any such release would have such a molecular weight and, indeed, the evidence suggested that such a release would, if anything, be neutrally buoyant. The second proposition is that the effect of scaffolding in the vicinity of the site would be to produce a dispersed release which would, on the evidence, bring up an alarm in zone C2 before one in zone C3 and therefore that the evidence was not, and could not be made to be, compatible with the evidence as to the occurrence of the alarms shortly before the explosion. However before coming to deal with these major issues it is, in my view, necessary to deal with two preliminary matters. The first is the question of the precise location of PSV 504. The second is the question of what Bollands’ evidence about the occurrence of gas alarms amounted to and, in particular, whether he should be taken as describing gas alarms occurring in a particular order during what may be called the second phase of the occurrence of alarms.
The Lord Ordinary found that PSV 504 was located close to the second stage suction scrubber of the reciprocating compressor and was approximately 15 to 20 feet above deck level (Opinion I.49). The position of PSV 504 was spoken to by Mr. Wottge (36/5676). He identified the valve as seen in a photograph of the model of the platform, production 12/4A, according to which it appeared to be positioned adjacent to the south-east corner of the eastern of the two reciprocal compressors in module C, that is, adjacent to the walkway which ran north and south just to the east of the reciprocal compressors. Mr. Wottge said that the valve was elevated at about 15 feet above the floor. There was, however, another model photograph, 12/4B, which showed PSV 504 positioned somewhat further to the north although still above the region of the walkway. A difficulty arose because it appeared that changes had been made to the model and no one was in a position to speak to when the changes had been made or why. The fact that changes had been made only emerged in the evidence of Dr. Davies, the reclaimers’ expert witness who made investigations into the behaviour of gas clouds released at various points in module C, in Volume 129. The defenders therefore suggested that there was an unnecessary degree of doubt about the location of the PSV, that only Mr. Wottge actually spoke to its position and that he had spoken to the position shown in 12/4A. There was a further confusion in that there were photographs of an original version of the model taken before the condensate injection pumps, which were not an original part of the platform equipment, had been installed or shown in the model. The reclaimers, however, pointed out that Mr. Wottge had given evidence not only by reference to the photograph 12/4A but also by reference to two drawings, 12/58 and 12/55 and further referred to the evidence of Mr. Bagnall at 10509, again with reference to a schematic 14/39 and to a photograph included in appendix 7/4 at item 21 which showed the PSV and its relationship to the adjacent ball valve. In my view it is reasonably clear that that photograph, and Mr. Bagnall’s evidence, are consistent with the position shown in 12/4B rather than that in 12/4A. Further, the Lord Ordinary summarises the evidence of other witnesses about the position of the PSV in his Opinion and, without going through it in detail, it seems to me that that evidence also supports the reclaimers’ proposition about the position of the PSV. In my view, therefore, the Lord Ordinary’s finding about the position of the PSV should be accepted. In passing, it may be worth noting that it is clear that the PSV was so placed that the junction between the flange of the PSV and the flanges on the adjoining pipework was oriented east-west.
The next matter concerns the evidence of Mr. Bollands. I have set out Mr. Bollands’ evidence in some detail elsewhere in this opinion. It is quite clear that Mr. Bollands is describing the occurrence of one low gas alarm, relating to zone C3, followed by a distinct interval, perhaps of two minutes, and then a number of gas alarms close together. The question is whether what Mr. Bollands says about the second group of alarms can be interpreted as giving any particular order for their occurrence. In my view, it is clear from a straightforward reading of Mr. Bollands’ evidence that he does not give such an order. Indeed it is quite clear that Mr. Bollands was excited when giving evidence, that he was describing events which had occurred very quickly and in circumstances of some stress and that he was not able to give any particular order for the second group of alarms. That, as I understand it, is the view taken by the Lord Ordinary and there seems to me to be no reason to differ from it.
The defenders, however, referred to information which Dr. Davies had obtained in July 1989 and which had been derived from Bollands. This information is recorded in a note at pages 458 and 459 of 65/11 of process. Initially, the information provided (not through direct contact with Bollands) was that there was a low gas alarm at C3 followed by a gap of two minutes and then low gas at C4 and C5, low gas at C2 and high gas at one of C3, 4 or 5, all described as close together. Dr. Davies sought further clarification and was provided with information to the effect that after initial alarm at C3 there was a gap of two minutes: there were then low level alarms at C4 and C5 followed by a ten second gap: a low level alarm at C2 followed by a five second gap: and then the high level alarms. The defenders referred to a passage in Mr. Bollands' evidence at 9510E in which he described four lights appearing on the matrix panel for zones C2, C3, C4 and C5 and said that that evidence represented what he had always said. That may well be the case, but the precise timings given in Dr. Davies’ note are not derived directly from Mr. Bollands but from an interpretation by someone of information supplied by him. In my view, that cannot be relied on as against the evidence which Mr. Bollands actually gave in court, particularly having regard to the way in which the evidence was expressed. In my opinion, the matters to which the defenders referred derived from the notes supplied to Dr. Davies do not justify any view of Bollands’ evidence different from that taken by the Lord Ordinary and do not justify the view that Bollands was speaking to a particular sequence of alarms.
It is, however, appropriate to note in this context that one point which was clear in Bollands’ evidence and which is supported by the other evidence about the making of a radio call to Richard and Richard’s response, is that there was a distinct gap between the first low gas alarm and the second group of alarms. That is significant because Dr. Davies’ undisputed evidence was that that pattern could not be accounted for by a single gas release and therefore that there must have been a first release of gas, followed by an interval and then a second release of gas.
Against that background, I proceed to deal with the first of the two main points made by the defenders, that related to the molecular weight of any hydrocarbon gas released from the site of PSV 504 and the question of the correctness of the assumptions on which Dr. Davies proceeded. In his report 12/362A of process Dr. Davies set out his assumptions as follows:
"2.4 The composition of gas streams in Modules "B" and "C" were such as to have a molecular weight around 22 to 25. Allowing for reduced temperature at the release point, such a release may be considered as neutrally buoyant (equivalent molecular weight 29). The lower explosive limit (LEL) for such streams was typically 3.5%.
2.5 The composition of condensate streams in "C" Module was such as to have a molecular weight of 42 and an LEL of 2.2%."
Any release from the site of PSV 504 would be a release of condensate, rather than a release of a gas stream. There was evidence in the shape of a chart showing the composition of the various streams throughout the process on the platform from which it could be derived that the average molecular weight for the condensate stream entering the condensate injection pumps would be in the region of 42. In his experiments, Dr. Davies had sought to model the behaviour of a gas stream with a molecular weight of 42 (in effect equivalent to that of propane) which would be heavier than air, as well as that of a neutrally buoyant gas. All his evidence about the behaviour of a gas cloud in the module was based on that modelling and it would therefore be significant, if it cannot be taken to be established that a leak of condensate from the site of PSV 504 would have converted into a gas which would have behaved as a gas with a molecular weight of 42.
In his report, and in his evidence, Dr. Davies explained that what he had set out to do was to model the behaviour of gas clouds originating at different points within module C, including the vicinity of PSV 504. As contained in the system under pressure, condensate would be in a liquid phase and would only be converted, wholly or partly, to gas on release from the pressurised system. Dr. Davies’ position, as repeatedly explained in cross-examination, was that he was not concerned, indeed did not have the expertise to deal with, the question of precisely what would happen at the point of release of the condensate from the pressurised system or whether the whole or some fraction of the condensate would be converted to gas at that point or how long such a process would take. What he was concerned to do was to investigate the behaviour of gas once released and for his purposes it did not matter whether conversion to gas was instantaneous, through flashing at the orifice from which the condensate was released, or whether it occurred in the near vicinity of the orifice by a process of evaporation or flashing and evaporation. A substantial amount of time was spent during the cross-examination of Dr. Davies in attempting to suggest difficulties in this essentially simple evidence. It is, however, in my opinion clear from a number of passages in Dr. Davies’ evidence that the position which he adopted was as I have endeavoured to explain it; for example, it is so explained in Volume 136 at pages 19397 to 19403.
The point which the respondents sought to make, however, was that it was wrong, or at least not proved to be correct, to assume that the gas cloud formed in the vicinity of a leak from the site of PSV 504 would have a molecular weight equivalent to that of the condensate stream at that location. This was because the condensate stream is not a simple or uniform substance but a mixture of hydrocarbons of widely varying molecular weights and therefore densities. The contents of the stream extend from the extremely light, methane, to some very heavy hydrocarbons with very high molecular weights and densities. In simplest terms, the respondents’ proposition was that the lighter ends of the hydrocarbon mixture would flash-off, that is to say be converted to vapour, preferentially, leaving the heavier ends behind, in the form of liquid streams, liquid droplets, or an aerosol, but not in the form of a gas cloud. That could be important because hydrocarbon in a liquid, droplet or aerosol form would not activate a gas detector of the type installed in module C. There was, indeed, evidence that the process of instantaneous conversion of liquid to gas on release of pressure, referred to as flashing, would be preferential towards the lighter ends of the mixture. There is a difficulty about dealing with the evidence on this subject in detail, however, in that while the question of differential flashing or evaporation was addressed in the cross-examination of various experts during the proof, it was addressed primarily in the context of what would happen when liquid condensate was released through an orifice into an enclosed space. In particular, there was questioning about what would happen when liquid condensate was released by the operation of a ball valve into a pump chest (as must have happened if Vernon jagged the condensate injection pump). As a result, the respondents were not in a position to point to any direct expert evidence that condensate released into the atmosphere from an orifice, such as a gap at the flange at the site of PSV 504, would be converted into a gas cloud with a lower molecular weight than the average molecular weight of the condensate stream. They could only point to some general explanations of the flashing process and of the fact that it would tend to be preferential towards the lighter ends. That does not mean that the point made by the respondents does not have to be considered and dealt with. The general evidence is quite sufficient to show that there would be a flashing process on release of condensate from the site of PSV 504 into atmosphere and that there would be a tendency for the lighter ends to flash more readily than the heavier ends. It can be taken, therefore, that, in the first few microseconds after the release of pressure, the lighter ends would convert into gas while the heavier ends would remain in a liquid or at least aerosol form. The question, however, is not what would happen instantaneously or within microseconds after the release: the question is whether what would happen on the release would permit the formation of a gas cloud of an average density equivalent to that of the condensate stream. Put another way, the question is whether the parts of the condensate stream released as droplets, aerosol or vapour would continue in that state or would be converted into gas sufficiently quickly to allow, for practical purposes, the formation of a gas cloud which would meet Dr. Davies’ assumptions.
On this point, in my opinion, the evidence is, practically considered, clear. The process of flashing is described by Dr. Drysdale in his report 12/361 as a process which involves the near instantaneous evaporation of a substantial proportion of the liquid when the pressure is released, thus creating a large cloud of vapour and liquid droplets. The report continues,
"The heat necessary to convert some of the liquid to vapour is gained at the expense of cooling the remainder of the liquid to its atmospheric boiling point. This process is known as flashing: if the idealised, theoretical adiabatic flashing fraction (TAFF) is greater than about 35%, then following a sudden release, most of the residual liquid will become airborne as an aerosol rather than falling to the ground to form a liquid pool."
A little later, Dr. Drysdale states,
"The condensate contained a high proportion of low molecular weight hydrocarbons, methane, ethane and propane, and the flashing fraction would have been of the order of 40%."
In that part of his report Dr. Drysdale was discussing the discharge of condensate from the fractured condensate line causing the fireball which was observed during the disaster but the observations on the principle of flashing are relevant. Dr. Drysdale explained the position further in his evidence (100/14265) as follows.
"The flashing process involves evaporation of a significant proportion of the liquid. The heat for that process comes from the remainder of the liquid, so that once the flashing process is over, the liquid that remains will be at the normal boiling point of the liquid. What happens in fact is that the liquid, instead of remaining where it was, is expelled with the vapour usually in the form of a spray, and the individual droplets will gain heat from the surrounding air and evaporate, so you may end up with a very substantial quantity of what was liquid propane in the form of a vapour."
A little later (14267) he said,
"The experience … I believe the experience in the chemical industry is that a relatively small proportion of the flashing liquid will actually fall to the floor. Most of it is produced in the form of a spray, and the smaller droplets will actually evaporate before they reach the floor. I think we have to identify here that we are talking about flashing fractions of maybe 30% or more for this to be operational."
In another passage (14310) Dr. Drysdale said that the greatest proportion of the condensate stream was made up of methane, ethane and propane and said,
"If that pressure is released the methane will come out of the, potentially out of solution and a significant proportion of the ethane, perhaps all of the ethane would turn to vapour, and a very substantial proportion of the propane, perhaps 50% would flash. So we are talking about a high flash fraction, just by looking at these data."
In passing, it should be observed that there was a good deal of discussion in the evidence of the effect of the drop in temperature in the remainder of the liquid resulting from the use of energy in the flashing process and it was indicated that such a drop would, in the case of hydrocarbons in a confined space or chamber, act as a brake upon the vaporisation process. As was mentioned in one of the extracts quoted above, however, where the stream is released to atmosphere, heat can be drawn from the atmosphere surrounding the escaping stream and, as a result, much of the detailed evidence about the effect of temperature drop does not, in my opinion, assist in the present context. That is, in effect, supported by evidence from Dr. Richardson (16542) where he draws a distinction between what might occur at room temperature and what might occur in a confined space where there was a temperature drop.
The picture derived from Dr. Drysdale’s evidence as above quoted is, in general terms, supported by Mr. Wottge who said (42/6622),
"If we had a condensate leak as we talked, it is made up mostly of the heavier components of the gas, and the gas in that case that would flash off would be expected to be heavier than air and it would tend to be at the lower elevations."
It is also supported by Dr. Mitchison (83/12285) who observed the gases associated with condensate were denser than air and would tend to fall to the ground, a proposition which he repeated in cross-examination (85/12571). Again Dr. Balfour, in calculations of the effect of gas clouds on gas detectors recognised that a small proportion, but only a small proportion, of the pentane would be in a liquid form. In my view, therefore, the position is that there is direct evidence both theoretical from an expert and practical from an experienced engineer in the oil industry that the condensate released from the stream at PSV 504 would form a heavy gas cloud and that evidence is consistent with indications found in the evidence of a number of other expert witnesses. On the other hand, as I have already remarked, the respondents produced no direct evidence that the cloud would not form in such a fashion or that it would be lighter than air or even neutrally buoyant. The argument that the cloud would be neutrally buoyant was derived by counsel entirely from passages of evidence not directly related to the situation at PSV 504 and not directly spoken to by any expert witness. In these circumstances, in my opinion, Dr. Davies’ assumptions were sufficiently proved to have been correct.
The second major point in this chapter of the case concerns in particular the first gas alarm observed by Mr. Bollands. That alarm related to the C3 zone and since, on the reclaimers’ theory, the gas which gave rise to that alarm emanated from PSV 504, the problem for the reclaimer is how a discharge of gas from that site could have reached a C3 alarm without first passing and setting off at least one of the zone C2 alarms. This is a problem in relation to the first stage of the theory. On the view which I take of the evidence of Mr. Bollands, namely that what happened in the second stage was that a flurry of gas alarms occurred in a number of locations almost simultaneously, it is not a problem in regard to the second stage. I should perhaps add that there was a good deal of evidence about the functioning of the gas alarms, but, again on the view I take of Mr. Bolland’s evidence, I do not think that there is anything to be gained by going into the detail, which is fully set out by the Lord Ordinary in section 5.4 of his opinion.
In order to appreciate the point, it is necessary to consider the position of the various alarms in relation to PSV 504. PSV 504 was placed, as I have said, adjacent to the second stage suction scrubber of the eastern reciprocal compressors, in the vicinity of the walkway which ran from north to south to the east of the reciprocal compressors and towards the middle of the module at a height of 15 or perhaps, according to some of the evidence, up to about 20 feet above the deck. The zone C2 alarms were situated in the area between the walkway and the centrifugal compressors. Alarm G101/1 was situated in the roof space about 20 feet above the deck and east and south of reciprocal compressor A, the eastern one of the two. G101/2 was situated about two or three feet above the deck in the area of the centrifugal skids near the boot of the discharge scrubbers. G101/3 was situated above the central centrifugal compressor at about 15 feet. If the positions of these detectors are considered on a plan, it is immediately obvious that a gas cloud dispersing when the general flow of air in the module is from west to east, as was the case on the night of the disaster, from the vicinity of PSV 504 must be likely to come near to G101/1. The most likely hypothesis on which such a cloud would not be detected by G101/1 is that it was a heavy cloud which descended rapidly below the level of G101/1.
As regards the remaining alarms, there were two in zone C3, which related to centrifugal compressor C, namely G102/1, situated in the air intake on the south side of centrifugal compressor C, and G103/1, situated in the fuel valve on the north side of centrifugal compressor C. A further alarm, G102/2, was situated between the C and B centrifugal compressors and, if activated, would show an alarm both in zone C3 and zone C4. Alarm G103/2 was situated in the fuel valve on the north side of centrifugal compressor B and would show an alarm in zone C4. G102/3 was situated between the B and A centrifugal compressors and would show an alarm in zone C4 and zone C5. Finally, G103/3 was situated in the fuel valve on the north side of centrifugal compressor A and would show an alarm in zone C5. There were three other detectors, G28, G31 and G34 but these were situated inside the turbine compartments of the centrifugal compressors and could not detect gas in the module itself. Of the detectors which could detect gas in the module, G102/1 (in zone C3), G102/2 (zones C3 and C4) and G102/3 (zones C4 and C5) would trip the centrifugal compressors on detecting gas. It will be recollected that Mr. Bollands said that at least two and perhaps three of the centrifugal compressors had tripped immediately before the explosion, which is consistent with a gas cloud affecting the centrifugal compressor area.
If the gas formed in the vicinity of PSV 504 was a gas of an average molecular weight similar to the average of the composition of the condensate stream, that gas would have been substantially heavier than air which has an average molecular weight of 29 (as opposed to 42). Accordingly gas released from the site would have tended to sink and a straightforward gas release might therefore have gone below the level of G101/1. If the release contrived to miss G101/1, it might well have been able to sink lower in the module and pass along to the eastern end where it could activate a C3 alarm. It is at this point, however, that the presence of scaffolding under the site of PSV 504 becomes potentially significant. As has been explained, scaffolding had been erected to enable PSV 504 to be removed, and it had not been taken down. There was no evidence as to the form or size of the scaffolding. One of the experiments conducted by Dr. Davies consisted of the observation of the results of a jet discharge vertically downwards from the site of PSV 504 directed onto a plate designed to mimic the presence of a scaffold under the jet. In that test, the first alarm to detect the presence of gas was G101/1. There were a series of other tests carried out by Dr. Davies in relation to fully and partly circumferential releases of gas, that is to say releases designed to mimic the effect of gas escaping from the whole circumference or from a considerable part of the circumference of the junction between a pipe and a blind flange. Those releases gave rise to what Dr. Davies described as a diffuse gas release, and they too characteristically set off alarm G101/1 first, or among the first. Dr. Davies therefore regarded the result of his test mimicking the presence of scaffolding as equivalent to, or producing results equivalent to, those of a diffuse gas discharge and took the view that such a discharge could not account for the pattern of alarms, that is to say it could not account for C3 showing before C2. Having reached that view, Dr. Davies did not test the matter further either in his original experiments, which were conducted in order to prepare to give evidence at the Cullen Enquiry, or in later work in 1993.
The respondents’ submission was that these results effectively disproved the reclaimers’ case that a first alarm at C3 could be produced by a release of gas at PSV 504 in the presence of scaffolding.
The reclaimers’ reply to this line of argument fell into two parts. Firstly, it was suggested that there was a high degree of variability in wind conditions, that the actual conditions on the platform might not precisely correspond to the conditions modelled by Dr. Davies, that the precise geometry of the scaffolding was unknown and that for these and a variety of similar reasons Dr. Davies’ results mimicking the presence of scaffolding could not be taken as excluding the possibility of a gas release which, even in the presence of scaffolding, would activate C3 before C2. It may be important to add that Dr. Davies conducted his tests on the assumption that G101/1 was set at a height of 12 feet above the deck, whereas on the evidence, as the Lord Ordinary held, it was set at 15 to 20 feet above the deck (Opinion vol. 2 p. 236). The second line of argument was that, notwithstanding the results in the one series of tests in which scaffolding was mimicked, there were at least two types of release envisaged by Dr. Davies which could give rise to the phenomenon of a C3 alarm before a C2 alarm, namely a slow continuous release and what was described as a puff release of condensate. In particular, the reclaimers referred to one of Dr. Davies’ series of tests, series 34, which simulated a fully circumferential leak at a rate of 12 kg/min and to evidence that such a leak could be considered to be in the same class as a downward jet or partly circumferential leak in that, in each case a dispersed source of gas would be produced. In series 34, the first alarm to operate was a C3 alarm, not a C2 one. To that line of argument the respondents riposted that there was nothing to show that Dr. Davies had taken into account the presence of a scaffolding in assessing the possibility of either a slow continuous release or a puff release leading to a C3 alarm before a C2 alarm.
One of the difficulties in assessing the full effect of Dr. Davies’ evidence at the appeal stage is that a considerable amount of it was concerned with the question whether a release of any given kind could have created the flammable mass necessary to fuel the first explosion. In the end, however, the respondents did not submit an argument at the appeal that such a flammable mass could not have been created. To some extent, therefore, looking at the evidence about releases which might or might not have triggered alarms in various points is looking at the evidence from one point of view only, while the actual evidence was directed towards two separate matters. It seems to me, however, that, in the state of play in the appeal, the only possible course is to look at the evidence as potentially indicating whether or not there was any form of release which would not activate C2 before C3, especially as, for the reasons earlier indicated, the issue before us concerns the initial alarm, not the later flurry of alarms. It is quite true that there is repeated reference to a low continuous release and to a puff release in Dr. Davies’ evidence. What I have found it difficult to do, however, even with the assistance of the reclaimers’ submissions both oral and written, is to identify any point at which Dr. Davies can be found clearly expressing the opinion that, notwithstanding the presence of scaffolding, either of these forms of release could account for the pattern of a C3 alarm in the absence of any alarm in C2. In the result, therefore, it seems to me that the reclaimers are forced back to the position that the uncertainties of the modelling process and the uncertainties of the actual conditions on the platform at the material time are such that there must be room for some result which might otherwise be unexpected, on the basis of the tests.
On the other hand, it does not appear to me that the respondents can say that the results of the single test that was done with a scaffolding in position are such as to make the reclaimers’ theory impossible as, I think at least at one stage of their submissions, they suggested. The results on which they founded were results for general diffuse releases of various kinds and for one jet release at a 180 degree angle to the vertical directed at a plate representing scaffolding. The situations considered are limited and it would be in my view building too much upon them to take the further step which the respondents suggest could be taken.
It is, I think, also relevant to bear in mind that Dr. Davies’ general approach, in accordance with his original instructions in 1989, had been to attempt to find out if there were circumstances which could account for the pattern of alarms and the explosion if there was a release at the site of PSV 504. It is also relevant that the overall result of all the tests was to show, as the reclaimers repeatedly insisted, that in the conditions prevailing at the platform on the night of the disaster there would have been a preferential flow of gas towards the south-east corner of the module: and that Dr. Davies modelled PSV 504 at a height of 12 feet as opposed to 15 to 20 feet above the deck. Bearing these factors in mind, it seems to me that it can be said that Dr. Davies’ evidence, taken with the other expert and lay evidence, clearly establishes that the pattern of alarms can only have been accounted for by a two-stage release process and that a cloud of gas of the type and size necessary to create the explosion could have been released from PSV 504 and might have behaved in a manner which fits with the available evidence. The fact that the first stage release did not, apparently, trigger an alarm in zone C2 remains a difficulty for the reclaimers, which has to be weighed along with the other considerations in the case at the end of the day.
(a) Approach to the factual issues
As a preliminary to a final consideration of the issues in the factual part of the case, it is necessary to refer to submissions made by the respondents in relation to the question of proof and in particular proof by circumstantial evidence. The respondents referred to various authorities including, in particular, Benmax v. Austin Motor Co.  A.C. 370 and submitted that, once primary facts had been established, this court was in as good a position as the Lord Ordinary to deal with questions of inference, particularly as little turned on questions of credibility. It is true that in this case very much depends on inference, and for that reason, I agree with the respondents that it is necessary to look critically at the Lord Ordinary’s reasoning. It is also true that in some particular respects, such as in relation to the question whether Vernon might have known of the removal of PSV 504 but allowed it to slip his mind, I do not find it possible to agree with the Lord Ordinary, so that it is necessary to consider whether any such disagreements might detract from his conclusions. On the other hand, it is also necessary to bear in mind that the Lord Ordinary heard a great deal of evidence, much of it highly technical, over a long period. A case like this could never be disposed of by a crude assessment of credibility or reliability, but the fact that the Lord Ordinary had the opportunity to assess the witnesses and their evidence over such a long period entitles his judgment to respect, particularly as, in my view, it is clear that he examined all the evidence carefully and critically.
The respondents pointed out, correctly, that the reclaimers’ case depended entirely on circumstantial evidence, since there was no direct evidence that hydrocarbon escaped from the site of PSV 504, or that Vernon had jagged or attempted to jag condensate injection pump A or that Sutton had failed to fit the blind flange correctly. Reference was made to a number of authorities in relation to proof by circumstantial evidence, two of which require to be particularly noted. Firstly, the respondents relied on a dictum of Viscount Simon in Woods v Duncan  A.C. 401 at 422. Viscount Simon said,
"If it be thought remarkable that after so long and elaborate an investigation no one should have been found liable for this loss of life, the answer is that in order to recover damages it is necessary to prove liability against one or other of the parties sued. The key to what is uncertain may have been lost among the 99 who perished in the disaster."
The second authority is Rhesa Shipping S.A. v Edmonds  1 W.L.R. 948. That was a case in which a cargo ship sank in calm weather in the Mediterranean Sea in deep water and the assured plaintiffs sought to recover, under two policies of marine insurance, in respect of the loss of the vessel, claiming that the loss was occasioned by a peril of the sea. A number of explanations for the sinking were advanced and discarded during the proceedings. Finally, at the trial, the plaintiffs advanced as an explanation a collision with an unidentified, moving, submerged submarine which was never seen or detected, and never surfaced. The judge at first instance, Bingham J. as he then was, expressly said that he regarded the plaintiffs’ case as inherently "extremely improbable" but held, nevertheless, that the alternative hypotheses had to be ruled out and therefore that the plaintiffs’ hypothesis had to be accepted as, on the balance of probabilities, the explanation for the loss. That decision was upheld by the Court of Appeal but reversed by the House of Lords, the decision essentially being that the only inference which could properly be drawn from the proved facts was that the true version of the ship’s loss was in doubt and that there was no justification for drawing the inference that it was due to a peril of the sea. The only extended opinion was given by Lord Brandon of Oakwood who began by pointing out that the onus of proof lay on the plaintiffs. He then surveyed the findings-in-fact and (at page 954) observed that the course taken in the proceedings had led the trial judge into approaching the decision as being a choice between the cause of loss relied on by the plaintiffs and an alternative cause put forward by the underwriters and that he had not paid attention to a third possible solution, namely to say that he was left in doubt and that the onus had not been discharged. Lord Brandon referred to the proposition attributed to Sherlock Holmes that when the impossible has been eliminated what remains, however improbable, must be the truth and went on to say that that was an inappropriate approach in a fact-finding situation such as that which the judge was faced with. That was so for three reasons. The first was that the judge was not bound to make a finding one way or another but had open the third alternative of saying that the party on whom the burden of proof lay had failed to discharge it. The second reason was stated as follows,
"The second reason is that the dictum can only apply when all relevant facts are known, so that all possible explanations, except a single extremely improbable one, can properly be eliminated. That state of affairs does not exist in the present case: to take but one example, the ship sank in such deep water that a diver’s examination of the nature of the aperture, which might well have thrown light on its cause, could not be carried out."
The third reason was that the legal concept of proof on the balance of probabilities should be applied with common sense and should not be applied in such a way that a judge would find himself forced to find established a proposition which he regarded as extremely improbable.
As the respondents stressed, there is a considerable similarity between the circumstances to which Viscount Simon and Lord Brandon referred and those in the present case, in that, in the present case, the majority of the persons directly involved in the events preceding the explosion, and in particular those most directly concerned with the alleged cause of it, died in the disaster and the structure of the platform itself was so damaged that no useful information could be obtained from it. They therefore submitted that, even taking the reclaimers’ case at its highest, it was not possible to hold that their theory of the disaster had been established and that the only proper view for the court to take was that the issue was left in doubt. There is clearly force in that argument. On the other hand, the respondents did, as in some other aspects of their argument, tend to push their propositions beyond a reasonable limit. I do not think that either Viscount Simon or Lord Brandon intended to say that the fact that some material evidence had been lost or that it was impossible to make some investigations which would clearly have been relevant is per se an obstacle to the making of findings. There are a great many cases in which a material witness, often a pursuer, has been killed and evidence has been lost. Further, the situation in the Rhesa Shipping case was an extreme one and, I think, much the same could be said of Woods v. Duncan, although I have not sought to analyse the facts in as much detail. It is certainly true that in a case like the present the paucity of evidence available is a very material consideration which must be borne in mind. I would, however, prefer to formulate the task facing the reclaimers as being that they should show circumstances so coherent and convincing as to establish, on a balance of probabilities, the hypothesis which they advance and to enable the court to hold that it should ignore the possibility that, had matters been otherwise, evidence pointing to a different conclusion might have been found. I would add that it is clear from the Lord Ordinary’s opinion (at p. 214-5) that he did have Woods v. Duncan in mind.
I should observe, also, that the reclaimers criticised the Lord Ordinary’s approach to circumstantial evidence on the basis that he had erred by selecting particular passages of evidence and then considered and rejected other passages of evidence as not fitting in with those which he had initially selected, whereas the proper approach would have been to survey all the evidence and see whether a conclusion could be reached from it, taking it all together. In my view that was a criticism which was most unfair to the Lord Ordinary. The Lord Ordinary, in my view, was at great pains to survey all the issues arising in the case, including many issues which have not been the subject of argument in the appeal. He also took great care to refrain from reaching positive conclusions on particular parts or sections of the case in isolation but to leave open all possibilities which must reasonably be taken into account. In my view, the Lord Ordinary’s approach was certainly not subject to the criticism which the reclaimers sought to make.
(b) Balancing factors
Having considered the test which the reclaimers required to meet, the next step is to review the known and established facts. In my view, the critical known facts can be summarised as follows:-
(1) The initial explosion took place in Module C, toward the eastern end.
(2) The fuel for the initial explosion was condensate.
(3) PSV 504 had been removed from its position and replaced with a blind flange.
(4) Immediately before the events which gave rise to the explosion, Vernon was involved in an attempt to start pump A. He asked Clark whether any work had been done on the pump, which suggests that it was the pump which was his concern. That fits with the evidence that he had, in the control room, a pink permit, which would be the permit for maintenance work on the pump.
(5) There were two releases of hydrocarbon gas separated by about two minutes, the first being a small release triggering one low alarm and the second a substantial release triggering a multiplicity of alarms immediately before the explosion. These alarms were concentrated in the eastern end of Module C, down-wind of the position of PSV 504.
Starting from those facts, there seemed to me to be three principal considerations favouring the reclaimers’ theory:-
(1) The gas release pattern cannot be explained unless there was some intervention in the process of production. A release from the site of the PSV could have accounted for the explosion and the pattern of alarms seen before it. There is nothing to suggest that there might have been any intervention other than human action.
(2) No other likely source of condensate or gas to fuel the initial explosion has been suggested. Of course a release of condensate could have come from any part of the system in which there was condensate. However there is nothing to suggest that there might have been any interference with any other valve or section of piping or any other event to permit an escape apart from the removal of PSV 504.
(3) The actions of Vernon, and of Clark and also of Richard strongly indicate that they were unaware that the PSV had been removed: had they been aware of its removal, they would not have acted as they did.
The main considerations in favour of the respondents’ argument seem to me to be as follows:-
(1) It is difficult to imagine how Vernon could not have known, given the whole circumstances, that the PSV had been removed. Such ignorance could only be explained by a failure of not one but several systems of operation designed to ensure that a lead production operator would be aware of any fact as significant as the removal of a PSV If Vernon had known, he would not have acted as the reclaimers’ theory requires.
(2) The operation of re-pressurising pump A required to meet the reclaimers’ theory would have had to be done in an unusual way involving, in particular, Vernon in holding the push/pull button open for a significant period of time in order to produce the flammable mass. It would have been expected that the operation of jagging pump A would have been observed by the witness Grieve, had it actually been done at the material time. Grieve’s evidence, so far as it goes, suggests that attention was directed to restarting pump B while he was at the 68 foot level rather than bringing pump A into operation; and, if anything, places Vernon nearer pump B than the controls of pump A just before the explosion.
(3) It is improbable that the initial first stage release of gas from the site of PSV 504 or its vicinity could have missed triggering the C2 alarm at G101/1 before any C3 alarm.
(4) The reclaimers’ theory requires acts of negligence on the part of Sutton for which there is no evidence and there is no sufficient reason why Sutton, who was an experienced workman, should not have carried out the fitting of the blind flange in a proper manner.
It may be added that to accept the reclaimers’ theory involves putting the blame for the disaster on two unfortunate men who died in it and cannot defend themselves.
In accordance with the approach which I have tried to explain, it seems to me that what is required is a balancing of all the factors comprised in these various arguments. Of the factors relied on by the respondents, I find the point in relation to Sutton’s negligence powerful. As against that, it can be said that when the blind flanges were fitted it was understood that pump A would be out of commission for a substantial time, while it was hoped that the valve would be replaced that day or, at worst, the next. There was some suggestion also that working at the site of PSV 504 was not easy. These factors might possibly help to account for some slackness in doing the work, although I would hesitate to place too much weight on them. The other factors relied on by the respondents seem to me to be capable of some explanation, one way or another, which would fit in with the reclaimers’ theory. While I accept that, for the reasons discussed earlier, Vernon should have known of the removal of PSV 504, the maintenance arrangements make it comprehensible that he might have remained in ignorance. Grieve’s evidence of his last recollection of Vernon puts him in a position between the GOVs of the two pumps and I do not think that any of the direct evidence about what took place at the 68 foot level is clear or specific enough to exclude the possibility that Vernon jagged pump A. It has to be accepted that Vernon must have held the push-pull button open for a rather surprisingly long period to create the flammable mass but again the evidence is not so clear as to exclude the possibility that he might have done so. Again, the evidence about gas dispersion and the timing and length of the necessary jagging operations leave difficulties for the reclaimers’ theory, but no more.
Balancing all the considerations together it seems to me extremely unlikely that Vernon would have acted as he did had he known that the PSV was missing, as we know that it was. It seems to me also to be legitimate to bear in mind that, given that elaborate safety systems were in place, there must have been some serious failure in at least one, and very possibly more than one, of those systems to allow the disaster to happen. It seems to me further that there is a sense in which the whole of the direct evidence and the other circumstances do fit together and do point to an explosion fuelled by condensate originating in the vicinity of PSV 504. In the end, I do not think one can resolve the fundamental issue by trying to attribute weight to particular considerations in a precise or quasi-mathematical way. I have attempted to take full account of the respondents’ argument that the reclaimers’ hypothesis "heaps one improbability on another" but it does seem to me that that argument goes too far. I appreciate that the Lord Ordinary’s decision involves the conclusion that both Vernon and Sutton acted negligently and I recognise that that is a conclusion which should only be reached with reluctance and after careful thought. Nevertheless, I have arrived at the view that the Lord Ordinary reached the correct conclusion on the facts and that the reclaimers’ factual case was established.
For all these reasons, I agree that these reclaiming motions should be dealt with as your Lordship proposes. I would like to associate myself with your Lordship’s observations on the manner in which the case was managed, prepared and presented to the court.