Guardian National Insurance Co Ltd v Springgold Investments (Pty) Ltd (526/08) [2009] ZASCA 112; [2010] 1 All SA 301 (SCA) (23 September 2009)

THE

SUPREME COURT OF APPEAL

REPUBLIC

OF SOUTH AFRICA

JUDGMENT

Case no: 526/08

No precedential significance

GUARDIAN

NATIONAL INSURANCE

COMPANY LIMITED Appellant

and

SPRINGGOLD INVESTMENTS (PTY)

LIMITED Respondent

_______________

Neutral citation: Guardian National Insurance Co Ltd v Springgold Investments (Pty) Ltd (526/08) [2009] ZASCA 112 (23 September 2009)

CORAM: HARMS, MTHIYANE, LEWIS, VAN HEERDEN and MAYA JJA

HEARD: 24 AUGUST 2009

DELIVERED: 23 SEPTEMBER 2009

CORRECTED:

SUMMARY: Insurance – contamination of palm olein oil by water leaking from steam pipes in storage tanks – policy excluding liability for loss caused directly and solely by contamination – insured alleging holes in steam pipes caused by sabotage – on factual evidence sabotage highly improbable – expert evidence not tipping probabilities in favour of insured.

ORDER

On appeal from: The High Court, Durban (Patel J sitting as court of first instance).

1 The appeal is upheld with costs.

2 The order of the court below is set aside and substituted as follows:

‘The plaintiff’s claim is dismissed with costs.’

VAN HEERDEN JA (HARMS, MTHIYANE, LEWIS and MAYA JJA concurring):

Introduction

[1] This is an appeal, with the leave of the court below, against a judgment in favour of an insured against an insurer for an indemnity claimed under a policy covering property damage.

[2] Edible palm olein oil, which had been stored in two large tanks on the insured’s property, had been contaminated by water leaking from steam heating coils inside the tanks. The oil was insured, but the insurance policy excluded liability for loss caused directly and solely by contamination. On that basis the insurer repudiated liability. The insured then instituted action against the insurer, alleging that the escape of steam from the heating coils into the oil had been caused by malicious damage to the coils and that its claim was thus not excluded under the policy. The Durban High Court (per Patel J) found that the holes in the pipes forming the coils had indeed been ‘deliberately and maliciously caused by persons whose identity is unknown’. It is against, inter alia, that conclusion that the present appeal is directed.

[3] The principal issue to be decided is whether the insured succeeded in establishing that the pipes forming the heating coils

had been sabotaged. The second issue in the court below, to which we do not get, was whether the benefit under the policy was forfeited as a result of fraudulent means or devices employed in making the claim.

Background

[4] During 1996, Feedmill Developments (Pty) Ltd (‘Feedmill’), trading as Capital Oils Mills, carried on the business of manufacture, distribution and storage of edible oil and related products in Pietermaritzburg. In terms of an ‘assets all risks policy’ the appellant, Guardian National Insurance Company Limited (‘Guardian’), insured Feedmill against loss of or damage to property owned by Feedmill or in its custody or under its control.

[5] Feedmill had three large tanks on its property which were designed for the storage of palm olein oil. This kind of oil is either solid or very viscous at room temperature and needs to be heated to about 50 or 60 degrees in order to be easily pumped. To that end, heating coils consisting of an arrangement of pipes were laid just above the floor of each tank, to carry steam to heat the oil in the tanks. The networks of pipes making up the heating coils were assembled outside the tanks and pressure tested there. The heating coils were then lifted by crane into the tanks. Once the coils were in position inside the tanks, the links between the steam inlet and outlet in each tank were welded. The coils were then pressure tested again. There are 154 welded joints in each of two tanks and 153 welded joints in the third.

[6] During July 1996, Feedmill established that a large quantity of palm olein oil in two of its storage tanks had been contaminated by water, evidently through steam leaking from the pipes comprising the heating coils. As a result of the contamination of the oil by water, it was no longer fit for use in the manufacture of margarine, as had been intended. In mitigation of its loss, Feedmill used the oil in the manufacture of soap.

[7] Feedmill notified Guardian of its loss. After investigations had been conducted on its behalf by two loss adjusters, Mr Escott-Watson

and Dr Garforth, Guardian repudiated liability for the claim, relying on the exclusion in the policy of damage caused by ‘contamination

or pollution’. Feedmill then instituted action against Guardian in 1998. Feedmill was, however, subsequently liquidated and the

respondent, Springgold Investments (Pty) Ltd (‘Springgold’), as cessionary of Feedmill’s claim against Guardian, was substituted as plaintiff in the action.

[8] Springgold sought to get around the exclusion clause in respect of contamination by pleading that the introduction of steam into the oil was caused by ‘malicious damage to the said heating coils carried out by a person or persons, the identity of whom is to the plaintiff unknown’. The alternative basis of its claim (malfunction of the heating coil system) was subsequently abandoned.

[9] The hypothesis advanced by Springgold was that one or more persons gained entry to each of the relevant tanks when it was empty

and, using either a gas welding torch or a small diameter carbon arc rod, proceeded deliberately to make holes in the pipes so that steam would escape and contaminate the oil. In other words, sabotage, a cause falling within the ambit of the policy.

[10] Although initially disputed, during argument before the court below Guardian ultimately conceded that it would be difficult to resist the proposition that, if the holes in the steam pipes were caused by sabotage, those acts of sabotage were the proximate cause of the damage to the oil. So too, before this court, Guardian approached the appeal on the basis that, if Springgold had

established sabotage, then it would be ‘likely to succeed’ in the appeal. But, if sabotage had not been established, then contamination

was the proximate cause of the damage to the oil and Springgold must fail. The fate of the appeal thus depends on whether or not sabotage had been established.

Evidence

[11] Springgold accepted that it bore the onus of proving that the holes in the pipes had been caused by sabotage. In order to discharge this onus, Springgold led the evidence of two experts, Mr Nimmo and Mr Bodger, both mechanical engineers. Mr Nimmo was engaged by Springgold in early September 1997 to investigate the failure of the steam heating coils. He went into all three tanks and visually inspected every single weld. This visual inspection revealed that, in Tank 1, there were apparently four holes of about one millimetre in diameter, each in the heat affected zone adjacent to a weld, two of which were on straight sections of the pipes and two on bends. The two straight sections of pipe with the holes were cut out and sent to the Metallurgical Laboratory of the Department of Mechanical Engineering at the University of Natal, Durban, for microscopic analysis. According to Mr Nimmo, there was no evidence of any repair work having been done on the welds in Tank 1.

[12] By the time Mr Nimmo conducted his investigation, the ‘weld failures’ in Tanks 2 and 3 had already been repaired. Mr Nimmo

found 17 such weld repairs (randomly scattered) in Tank 2 and four in Tank 3 (these four all being on the first bend on the incoming steam line).

[13] In respect of Tank 1, the microscopic analysis of the two straight sections of pipe revealed one hole in each positioned similarly at the edge of the butt weld. The ‘hole’ in the one section had not, however, fully penetrated the wall of the pipe. The microscopic analysis was carried out by a Mr Bartholomew, but as he had passed away prior to the trial, Mr Bodger testified to Mr Bartholomew’s report and the photographs attached to it. Both the report and the photographs were also annexed to Mr Nimmo’s report and much of his evidence related to conclusions to be drawn from an examination of the photographs.

[14] According to Mr Nimmo, his examination of the holes in Tank 1 ‘from the outside’ did not lead him to suspect sabotage. The first time he had thought of sabotage was in his discussion with Mr Bartholomew after the latter ‘had ascertained that the line of the hole was away from the weakness, the line of weakness’. In his report, Mr Nimmo stated the following in this regard:

‘The weakest part of a weld join is the heat affected zone, ie the area alongside the welded join where the two different metals mix and fuse. Any natural attack, eg corrosion, erosion or cracking, in a weld join, would most likely occur along this zone line. It is clear from these photos [the photographs of the microscopic examination of the holes attached to Mr Bartholomew’s report] that the direction of the holes does not follow any natural line of weakness, but rather they move away from the weak points into the parent pipe material.’

[15] Counsel for Guardian submitted, correctly in my view, that from Mr Nimmo’s evidence and from a consideration of Mr Bartholomew’s

report, as testified to by Mr Bodger, it is clear that the path of the holes is central to the proposition that these holes were the result of sabotage. However, an examination of the photographs showing the penetration into both samples of pipe analysed does not support the conclusion that the ‘line of the hole was away from the weakness’. In respect of the fully penetrating hole, the penetration is in fact perpendicular, ie not away from the heat affected zone of the weld join, but equally not towards it. In respect of the second ‘hole’, which had not fully penetrated the pipe, the angle of penetration is actually towards the weld. Indeed, Mr Bodger conceded in cross-examination that the penetration was ‘more or less straight downwards’ and that, as regards the second hole, ‘the indentation here actually ate away some of the welding. It goes into the welding a little bit and eats some of the welding away.’

[16] From the expert evidence, it cannot be disputed that there was a second application of heat to the two weld joins which were examined by Mr Bartholomew and that this second application of heat

was deliberate. Both Mr Nimmo and Mr Bodger denied the possibility that this second application of heat might have occurred in the course of a botched attempt to repair or neaten the original welding work. However, contrary to Mr Nimmo’s report, it is clear from Mr Bartholomew’s photographs that there had been at least one attempt at repairing or neatening a weld join in Tank 1, by applying an extra weld run to an existing weld, thereby causing changes in the microstructure of the surrounding metal due to the applied heat. These changes are identical to those commented on by Mr Nimmo and Mr Bodger in respect of the metal surrounding the weld joins where the holes had been found. Furthermore, as regards Tank 2, one of the 17 weld repairs identified by Mr Nimmo was already present when Mr Escott-Watson examined the heating coils in Tank 2 in October 1996, but the pressure tests conducted by him on the coils in Tank 2 showed that water was not escaping through this weld join. This would also seem to indicate that repairs had been carried out in Tank 2 before the later repairs made to remedy the holes in Tank 2 through which steam escaped into the oil.

[17] Both Mr Nimmo and Mr Bodger initially maintained that the holes in the samples analysed by Mr Bartholomew had been made using a gas welding torch. In this regard, Mr Nimmo explained that –

‘. . . the reason is that a gas welding torch is more portable – I mean there is a lot of logistic reasoning, if you were going to go in there with an arc welding you would have to have electricity supply from a nearby power point, you would have to have a supply cable of 100/150 metres, you would have to have wheeled your welding machine down to the entrance of the tank and you would then go in to a completely steel vessel with your electric cable, all of those things tell me that somebody wishing to do this undetected would find it difficult, whereas if you get a portable welding torch you can get a gas cylinder about the size of a fire extinguisher and that is quite portable, you can walk in there with that thing and operate within the tank.’

[18] Mr Nimmo conceded, however, that ‘metal spatter’ would not occur with a gas welding torch being used to make a hole as there would be no deposit of metal around the hole. Metal would only be deposited if an arc welder was used. When his attention was drawn to the presence of metal spatter in the vicinity of both holes revealed by Mr Bartholomew’s report and photographs, he conceded the possibility that the holes were made using an (electric) arc welder, but stated that the diameter of the welding electrode used would have had to be very thin. Later on during his cross-examination, he denied that the holes might have been made in the course of a botched attempt to repair existing welds, giving as his reason for this denial the fact that there was no deposit of metal around the holes. At this stage, he appeared to have forgotten about or ignored the presence of metal spatter in the vicinity of both holes appearing from Mr Bartholomew’s report and photographs.

[19] Similarly, when Mr Bodger was asked in cross-examination whether he could exclude the possibility that the holes were the ‘unfortunate product of some rather inept work’ by one or more of the welders who had originally worked on the heating coils, his response was the following:

‘Yes, I would say so, because the welder would have been using an electric arc, and he would have been using a welding rod. He couldn’t make that hole with a welding rod, because a welding rod deposits metal. He would have had to change either to a carbon rod, which doesn’t deposit, or just use gas.’

Here too, the presence of metal spatter in the vicinity of both holes revealed by Mr Bartholomew’s report and photographs appears to have been overlooked.

Conclusions

[20] In my view, in light of the inconsistencies highlighted above, an overview of the evidence given by Mr Nimmo and Mr Bodger does not exclude the reasonable possibility that the holes in the two sections of pipe examined by Mr Bartholomew were caused by poor workmanship or by a botched attempt to repair or neaten the original welding. Mr Escott-Watson, the first person who had examined the heating coils in the tanks, testified that his impression was that it was poor welding that had caused the holes in the pipes and that –

‘It’s quite easy when welding to accidentally or on purpose, for that matter, if the welding voltage is too high, to not complete the weld properly and you will end up with a hole – if you stick the rod in too hard or too fast, yes, you could well end up with a hole.’

Mr Escott-Watson was, admittedly, not an expert in welding. However, Mr Nimmo also testified that there was evidence of welding with too high a current –

‘. . . those gouge marks on the outsides of the weld deposit material shows that he is welding with too high an electric current and a suitably qualified person would not have done that, they would have adjusted the current. So, it is a comment on the competence of the . . . welder.’

[21] In his report, the late Mr Bartholomew was highly critical of the quality of the welding. Thus, he considered the lack of root penetration on the samples analysed by him as being ‘totally unacceptable for welded pipework carrying steam’, as well as the poor fit-up of the pipes being ‘unacceptable in most codes covering welded pipe.’ In his conclusion, Mr Bartholomew stated that ‘the quality of the welding on the two samples of pipe examined would be considered unacceptable to any code of practice for fusion welded pipework carrying steam.’

[22] As pointed out by counsel for Guardian, although the expert, Mr Bodger, when speaking to Mr Bartholomew’s report, was a rather reluctant critic of the welding work, he ultimately had to concede that Mr Bartholomew’s analysis was correct.

[23] In Michael & another v Linksfield Park Clinic (Pty) Ltd & another 2001 (3) SA 1188 (SCA) at paragraph 40, this court stated the following:

‘Finally, it must be borne in mind that expert scientific witnesses do tend to assess likelihood in terms of scientific certainty . . . [The] essential difference between the scientific and the judicial measure of proof was aptly highlighted by the House of Lords in the Scottish case of Dingley v The Chief Constable, Strathclyde Police 200 SC (HL) 77 and the warning given at 89D-E that

“(o)ne cannot entirely discount the risk that by immersing himself in every detail and by looking deeply into the minds of the experts, a Judge may be seduced into a position where he applies to the expert evidence the standards which the expert himself will apply to the question whether a particular thesis had been proved or disproved – instead of assessing, as a Judge must do, where the balance of probabilities lies on a review of the whole of the evidence.” ’1

[24] On the factual evidence in this case, sabotage of the kind alleged by Springgold is, in my view, highly improbable. First, it would have been very difficult to obtain access to the tanks undetected. Each tank is about ten metres in diameter and ten metres

high. Access to a tank (when empty) is gained through a manhole in the wall of the tank. The manholes are secured to the side of the tank with very substantial 24 millimetre bolts. The shortest estimate of the time it would take to open one of the manholes appears to be that of Mr Nimmo, who put it at an hour and a half to two hours if two people were unbolting the hatch. Mr Escott-Watson estimated that it would probably take half a morning to remove one manhole cover, depending on how tight the bolts were, while Mr Essack, a director of Springgold who testified at length during the trial, expressed the view that it would take at least half a day to unbolt and remove a single manhole cover.

[25] The three storage tanks were situated near the north-western corner of the plant, the only entrance to which at that time was at the south-eastern corner, a considerable distance away from the tanks. Feedmill’s soap plant lay immediately to the north of the three tanks and its margarine facility ranged to the east/south-east of the tanks. According to Mr Essack, at the time of the discovery of the loss in 1996, the soap plant ran either five or seven days a week, 24 hours per day, while the margarine plant, depending on orders, sometimes worked a day shift and sometimes 24 hours per day. There was both internal and external security at the plant, comprising about six or seven security guards per shift. The plant was properly fenced, with brick walls, razor-wire fences and a gate. Because both plants ran night shifts at times, the whole area was well lit.

[26] I agree with the submission by counsel for Guardian that, given these facts, the prospect that saboteurs could gain entrance to the plant, traverse the entire length of it to get to the tanks, remove the manhole covers in full view of the soap and margarine plants and the employees in that area, make holes in the pipes inside the tanks and then re-seal the manhole covers, again in full view of others, is extremely remote. It is also highly improbable that saboteurs would go to the trouble of making about 21 tiny holes scattered in a random manner through the three tanks when, on the evidence, there were considerably simpler and more effective means of sabotaging the tanks so as to contaminate the oil contained in them.

[27] In explaining his conclusion that the holes in the pipes had been maliciously caused by persons unknown, Patel J stated the

following:

‘In reaching this conclusion I am alive to the fact that the plant was securely guarded. Further there may have been easier methods than the one used to sabotage the plant. Although it may have been difficult and time-consuming to obtain access to the tanks, it is not improbable that the damage may have been caused shortly after the tanks were pressure tested by the very persons who did the welding and subsequent repairs. Considering the evidence of Mr Haroun Essack and especially his demeanour in court, and without wanting to be uncharitable to him, he would in my view not have been an easy task master to work with. Just like necessity, perversity is also a mother of invention.’

[28] That hypothesis is simply not supported by the evidence. As counsel for Guardian pointed out, if there was any suspicion that the welders might have been disposed to do what Patel J speculated, Mr Essack, who was in charge of the whole plant at all relevant times, would certainly have said so. On the contrary, Mr Essack testified that the welders were local people, well known to Mr Essack and his staff –

‘Local people we knew. They always come around looking for work. If there’s any work for fabrication, if there’s any work for construction, they were the guys who would move around . . . people who had worked for me since I started Capital Oil, in and out, Whenever I wanted them, they were available.’

Moreover, Mr Essack was adamant that the welders were properly qualified – ‘they were coded welders and they were qualified

boilermakers’ – and he described at length the careful and meticulous process employed by them, stating that the end product,

after the final pressure testing, was ‘perfect’.

[29] No motive having been attributed to the welders in the evidence, or any finger of suspicion pointed at them, Mr Essack was not asked whether, as one would have expected, the manhole covers were immediately secured after the internal work in the tanks had been completed. His evidence was certainly not to the effect that the tanks were left open for any length of time, giving the welders the opportunity to return undetected and surreptitiously make random holes in the pipes after the welding work had been completed, or giving someone else an opportunity to gain access to the tanks and inflict the damage without having to remove and replace the manhole covers.

[30] Springgold thus failed to discharge the onus of proving that, on the balance of probabilities, the holes in the pipes through which steam leaked into and contaminated the palm olein oil were the result of sabotage. After all, the expert evidence was the high water mark of Springgold’s case in support of its asserted hypothesis of sabotage. As I have shown, that evidence, which admits of other reasonable inferences, falls short of tipping the probabilities in its favour. This conclusion renders it unnecessary for me to deal with the numerous inconsistencies and contradictions in Mr Essack’s evidence, or to address the issue of whether Guardian managed to establish that Springgold’s benefits under the policy stood to be forfeited as a result of the employment by Springgold of fraudulent devices in making the claim.

Order

[31] The following order is made:

_____

B J VAN HEERDEN

JUDGE OF APPEAL

APPEARANCES:

For Appellant: P J Olsen SC

Instructed by: Deneys Reitz

Durban

Webbers

Bloemfontein

For Respondent: S R Mullins SC

Instructed by: Shepstone & Wylie

Matsepe Inc

1 See further in this regard, D T Zeffertt, A P Paizes & A St Q Skeen The South African Law of Evidence (formally Hoffmann and Zeffertt) (2003) at 305-306.