Pioneer Foods (Pty) Ltd v Bekker (322/16) [2016] ZAECGHC 103 (20 September 2016)

IN THE HIGH COURT OF

SOUTH AFRICA

EASTERN CAPE

DIVISION

: GRAHAMSTOWN

CASE NO. 322/16

In the matter between:

PIONEER FOODS (PTY) Ltd

Appellant

and

JOHANNES MARTHINUS

BEKKER

Respondent

FULL

BENCH APPEAL JUDGMENT

GRIFFITHS, J.:

[1] On Sunday, 27 August 2011, and on subsequent days, Mr Johannes Marthinus Bekker (the respondent), a farmer from Aliwal North, lost a substantial number of livestock. He launched an action against Pioneer Foods (Proprietary) Limited (the appellant) for damages alleging that the animals died from ingesting screenings (a by-product of maize milling) supplied by the appellant which were contaminated with a respiratory metabolic nerve toxin called aluminium phosphide, alternatively with the botulism toxin.

[2] The matter came before Lowe J on the question of the appellant's liability, quantum to be decided only in the event of a positive finding in favour of the respondent. On conclusion of the trial, Lowe J made such a finding but upon application by the appellant granted it leave to appeal against his ruling.

The basis of respondent's claim

[3] The respondent founded his claim on an oral contract in terms of which he purchased 1.94 tons of screenings from the appellant. It became common cause between the parties that the claim was based on the actio ex empto, the appellant admitting that it was a manufacturer of the product in question who publicly professed to have the attributes of skill and expert knowledge in relation thereto and who, accordingly, became liable for consequential losses caused to the purchaser by reason of any latent defect found to exist in the product[1]. The following statement of the law by Corbett JA was also accepted:

"The legal foundation of respondent's claim is the principle that a merchant who sells goods of his own manufacture or goods in relation

to which he publicly professes to have attributes of skill and expert knowledge is liable to the purchaser for consequential damages caused to the latter by reason of any latent defect in the goods. Ignorance of the defect does not excuse the seller. Once it is established that he falls into one of the above mentioned categories, the law irrebuttably attaches this liability to him, unless he has expressly or impliedly contracted out of it."

[4] The appellant defended the matter mainly on the basis that its supplied product was not latently defective by contamination with either phosphine or botulism at the time of sale and delivery. Lowe J, in a detailed and well-reasoned judgment, found that the respondent had established the existence of such latent defect at the time of delivery and it is against this finding that the appellant has come on appeal.

The evidence

[5] The respondent called the evidence of five witnesses being himself, Dr Ferreira (an animal scientist), Mr Basson (a senior research technician at Onderstepoort), Dr Troskie (the attending veterinarian surgeon) and Professor Kaye (a retired professor of chemistry in which he holds a doctorate). Their evidence revealed that the respondent farmed on two separate farms which were some 120km apart. The first farm on which the respondent resides is named "Amazango" and the second, "Floukraal", on which his parents live. On these farms he farmed predominantly with sheep, cattle and goats.

[6] Commencing in 2008, he exercised his rights under a tender awarded to him by the appellant in terms of which the appellant supplied him both with screenings and a further by-product known as "hominychop". These products were mixed with two other products depending on whether or not he was feeding cattle, or goats and sheep. If he was feeding cattle, a product known as "beesvet 33" would be mixed in, and if sheep or goats were to be the recipients, a product known as "skaapvet" was placed in the mix. The mixing of the three products was performed in his shed at Amazango. Once mixed, the feed would be fed to the livestock on both Amazango and Floukraal.

[7] On Friday, 26 August 2011, he collected the feed from the appellant’s nearby mill. On doing so, he noted there to be some form of labour unrest as he was obliged to use a separate entrance and that, unlike previous occasions, there was a sharp smell associated with the feed he was collecting which burnt his nose and eyes. On Saturday, the feed was duly mixed by his labourers in the shed on Amazango under his supervision, after which it was fed to the animals on both farms.

[8] Before sunrise the following day, a bull and some cows on Amazango had died. Dr Troskie was called and performed a post-mortem on certain of the carcasses taking samples for analysis. He also took samples of the feed which were likewise sent for analysis. He was advised by Dr Troskie to re-inoculate against botulism, and to remix the feed. Despite this, the animals continued to die during the course of the next week and a half. Many of the stock at Floukraal also died. However, those animals which were not given this particular feed remained healthy. His evidence was largely supported by Dr Troskie who testified that he had diagnosed poisoning which emanated from the mixed feed. He suspected this to be phosphine. Certain symptoms of such poisoning were found to be present in the dead cattle, such as bulging eyes and protruding and frozen (i.e. stiff) tongues.

[9] Ferreira confirmed that neither the ‘beesvet’ nor the ‘skaapvet’ used in the feed mix contained any toxins, and in particular did not contain botulism. Mr Basson testified that he had performed a certain analysis on the feed, which he suspected of possibly containing phosphine[2] due to the sharp smell emanating from the feed. Such analysis proved positive in establishing the existence of phosphine in the feed, although in a relatively small amount. He, as supported by Prof. Kaye, testified that one of the properties of phosphine is that it has a relatively short "half-life" in that, especially when it is exposed to the air and to moisture, it will dissipate over a period of time. They both confirmed that the presence of a small amount of phosphine in the sample analyzed at Onderstepoort a few weeks after the mixing thereof, meant that at the time of the mixing and ingestion thereof by the livestock, phosphine would have been present in sufficient quantities to have caused death.

[10] Professor Kaye and one of the appellant's experts, Mr van Aswegen, produced a joint minute in which they indicated certain areas of their evidence upon which they were in agreement, and others which were in dispute. They were essentially in agreement that certain pellets containing aluminium phosphide used during the appellant's milling process produced phosphine gas which, in turn, dissipated throughout the maize whilst resting in silos. The essential purpose of this was to fumigate the maize which, in its rough form, contained insects and their larva or pupa. Over a period of time the phosphine gas would dissipate as it has a half-life[3] of approximately five hours when exposed to the atmosphere, whilst in the dark the half-life is approximately 28 hours. These experts

further agreed:

"In the use of phosphine as a fumigant the pertinent issues were the time needed for atmospheric dissipation of the phosphine to below hazardous levels; the absorption levels of phosphine in fumigated maize following ventilation (i.e. how long the phosphine remained in different grains); the possible presence in fumigated foodstuffs of unreacted aluminium phosphide (in this case pellets) residues capable of generating phosphine in contact with moisture or following ingestion.

Any phosphine present in the "product" would retain its normal properties, its effective toxicity would depend on the concentration of such phosphine in the quantity of the product consumed. Although the aeration of grain results in rapid desorption towards safe levels, minute traces of phosphine can remain in fumigated material for extended periods and phosphine has been detected in wheat 220 days after fumigation. There are no studies of the equivalent in maize. It appears that the rate of desorption of phosphine depends on the nature of the grain, wheat, barley and canola losing 80 to 90% of the absorbed phosphine after aeration for one week but only 40% desorption was observed for oats. There was agreement that aeration following fumigation is undoubtedly

essential.[4]"

[11] These two experts however differed on one essential point: Prof. Kaye contended that all the available evidence indicated that the animals died of phosphine poisoning and that the toxin could only have emanated from the appellant's milling plant whilst Mr Van Aswegen maintained that it was highly unlikely that: (a) any aluminium phosphide could have been retained in the feed given the extensive processes of aeration and agitation which occur post-fumigation in the silos and other factors, and (b) that in any event the evidence pointed to botulism poisoning which could have emanated from a number of possible sources such as dead rodents in the mixing room, or animal carcasses lying around in the veld.

[12] The appellant called a number of witnesses some of whom were experts. As I have indicated, Mr Van Aswegen testified that botulism was the cause of death and that there was no evidence to substantiate a probability that this toxin was present in the screenings or the hominychop supplied by the appellant. It appeared from his evidence that he, and the appellant's other expert, Professor Prozesky, were ad idem that their diagnosis of botulism was based in the main upon the fact that the animals died over a period of a number of days whereas, in their opinion, phosphine would cause an acute death, that is the death of the animals post ingestion would occur within 24 hours or less. This, in their opinion, led them to conclude that botulism was the more likely killer.

[13] Professor Prozesky, an expert in pathology, maintained that he could conclude conclusively from a set of photographs depicting the dead animals, that the respondent was not correct in his assertion that the animals at Floukraal had died over a short period beginning on the Wednesday, following the feed having been placed in certain self-feeding bins on that farm. He testified that, from the photographs, he could positively conclude that certain cattle depicted on those photographs had died within two to eight days prior to the taking of the photographs (some weeks after they had been fed the impugned feed) and that the remaining animals had died over an extended period of time and by no means on the same day. The purpose of this evidence was, in the main, to challenge the credibility of the respondent.

[14] The balance of the evidence tendered by the appellant concentrated on the processes at the milling plant and in particular the fumigation process and the means by which the phosphine gas was caused to dissipate once the maize had rested for sufficient time in the silos. This included various means, such as powerful extractor fans which extracted dust from the maize (which should have included any residual aluminium phosphide powder or pellets) while it was in transit on conveyor belts from the silos, agitation and the use of a sieving process which ought to have prevented any remaining pellets from being included in the screenings. Evidence was also tendered to show that it was unlikely that there had been any labour unrest at the relevant time when the respondent collected the feed and that it was unlikely that he would have perceived a sharp smell as he had testified to. It was also established that there had been only one previous occasion where the appellant's feed had killed livestock through phosphine poisoning. This had occurred when a farmer, one Le Grange, had been supplied with feed shortly after an Easter weekend during which the plant had been closed down and extensively fumigated with aluminium phosphide.

The findings of the trial court

[15] Lowe J found the following to be the inherent probabilities which emerged from the evidence:

"163.1 That there was poison in the mixed feed;

163.2 That this poison was a strong/potent poison;

163.3 That the animals on both farms died from eating the same poison source;

163.4 That the poison concerned was either phosphine or botulism;

163.5 That the poison concerned was introduced into the feed either by way of contaminated screenings alternatively during the feed being mixed, having particular regard to the fact that the animals died on both farms shortly after the animals had commenced to

eat the newly mixed feed containing screenings purchased on the relevant date;

163.6 The quantity of poison contained in the mixed feed must have been substantial, or put otherwise, sufficient to have contaminated the entire quantity of feed mixed on two separate consecutive occasions, the animals dying at the two separate farms being given separate mixes, the first mix on Amazango, the second mix at Floukraal, the first mix containing "beesvet" the second "skaapvet";

163.7 If the poison was botulism, the origin thereof was unlikely to have originated in defendant's premises having regard to extreme care taken in respect of cleanliness and contamination; the more probable origin thereof being plaintiff's barn or the substances therein (excluding the additives), this contaminating the mixed feed during two independent consecutive mixes;

163.8 The fact that the animals concerned had been inoculated against botulism, and accordingly had some protection against same, unless in substantial quantities, renders the cause of poisoning as botulism, improbable.

163.9 The fact that the same feed sources (mix) was consumed by sheep at Onderstepoort would inherently indicate that this was not contaminated by botulism as this, unlike phosphine, would have remained present in fatal quantities had this been present in such quantities originally.

163.10 If the poison was phosphine, the source thereof could only have originated in defendant's premises;

163.11 The large number of animals that died on both farms, independently, were poisoned by the same poison source, whether phosphine alternatively botulism;

163.12 The physical reaction from a veterinary point of view and symptoms of the poisoned animals were likely to indicate the probable cause of death and poison responsible."

[16] The trial court further proceeded on the premise, correctly in my view, that the credibility of the respondent was crucial to a determination of the disputed issues. In this regard, and after a thorough analysis of all the evidence, Lowe J found:

"I have already made positive findings relevant to Bekker’s demeanour, and impression made in the witness box, and I find that his candour and demeanour were beyond reproach. He exhibited, in my view, no bias, latent or patent, nor was defendant able to point, in my view, to any compelling internal contradictions in his evidence."

[17] As regards the external contradictions contended for by the appellant, which focused particularly on the period over which the animals died and other alleged contradictions, the trial court found:

"In my view these purported external contradictions are not such as to place his evidence in any real doubt, or to cause one to have doubts as to his evidence when weighed against the inherent probabilities, or the calibre and cogency of his performance compared to that of other witnesses testifying about the same incident or events."

[18] He also found Dr Troskie's evidence to be beyond reproach and was satisfied with the evidence of the respondent's remaining expert witnesses.

[19] As regards the evidence of the appellant's lay witnesses, Lowe J was once again satisfied as to their credibility. The dispute between the expert witnesses for the respondent, and those for the appellant, was thoroughly analysed by the trial court and I can do no better than to quote from its judgment as follows:

"Defendant argued that Prof Kaye took sides and went beyond what could be expected of a scientist. It was pointed out that he had no experience of the application of phosphine gas in practice and was not entirely unbiased. It was pointed out, correctly, that he had assumed that there had been continuous administration of pellets to the same mass of maize. He was criticized for failing to make the concession that having regard to the factory protocol it was unlikely that any unreacted aluminium phosphide would have remained to poison the animals, and that he had been unaware that the first screenings were not fumigated at all. He was finally criticized for having expressed what was referred to as a "judicial opinion" as to the cause of death – this while he was presented to the court as a scientist.

As I have already said, in my view there is no merit in the criticism that Prof Kaye was biased or partisan in any way, he as I saw it maintaining his independence as an expert chemist. That he had no practical experience in the application of phosphine gas did not detract from his expertise scientifically or his explanation that unreacted aluminium phosphide, covered with a waxy coating, could have remained in the maize in sufficient quantities, to cause the death of the animals. His evidence, and conclusions, seemed to me to accord with entirely logical reasoning. This accords with the probabilities, as also that phosphine was found in the screenings some six weeks later (only as a trace), and the improbability of the deaths having been caused by the alternative, botulism. This is improbable in turn, as pointed out above, as this would have had to be present in the mix in substantial quantities (for which there was no logical explanation), and the animals having been inoculated against botulism. Prof Kaye's careful explanation, as judged against the evidence seen as a whole, seems to me to be entirely logical.

The extent of disagreement between Prof Kaye and Mr Van Aswegen, does nothing, in my view, to disturb the conclusion set out above. These areas of disagreement are largely technical and do not, against the probabilities, impact on Prof Kaye's reasoning or conclusion. Whilst it is true that the grain was vigorously aspirated during its journey from silo to mill, there remained on the probabilities the likelihood that some aluminium phosphide, by some mechanism, which is unnecessary for plaintiff to articulate or explain, remained in the screenings at the second stage. At no time was defendant able to argue that what defendant had purchased was anything but at least a mixture of seconds and first screenings in separate bags. Finally, it seems to me that Prof Kaye's error as to the continuous application of pellets to the same mass of grain is not such as to disturb the findings I have made above." (paragraph numbers excluded)

And later:

"Insofar as Prof Prozesky is concerned, I have already commented above on his expertise and approach to the matter. Insofar as the first element of his evidence is concerned (the date of death of animals on Floukraal) he was placed in the unenviable position of having to attempt to determine periods of death from a small number of photographs (panoramic) of the animals taken in general context. He conceded that this was a difficult task and when vigorously challenged in cross examination simply insisted that he had taken all relevant factors into account without any clear indication in his report that this was in fact the case. This part of his evidence flatly contradicted that of Bekker, and must be seen in the light of the conclusion reached as to Bekker's credibility and reliability as a first hand witness thereof. In my view this head-on dispute must clearly be determined in plaintiff's favour, seen also in the light of the inherent probabilities, the absence of any reasonable explanation as to why the animals would have continued to die over a lengthy period, without the further involvement of Dr Troskie, and without any explanation, save the possibility that the existing carcasses carry the possibility of botulism contamination.

As to the opinion expressed on the probability of botulism poisoning, this flew in the face of the personal expert observations of a fully qualified veterinarian Dr Troskie, the animals he saw failing to display the classic sign of botulism poisoning, the hanging paralyzed tongues as opposed to those stiff protruding tongues as described by Dr Troskie. Whilst I am aware that the defendant's experts attempted to justify that this was not necessarily conclusive in respect of acute deaths from botulism, this did not accord with Dr Fourie's initial evidence that the hanging paralyzed tongue was the classic sign of botulism poisoning, which was absent on the evidence of Dr Troskie. In short, he was hamstrung by the Dr Troskie description and diagnosis, and further the lack of any other scientific basis for contesting that conclusion. Further, and in support of the above, on the probabilities, as already referred to above, relevant to the alternative of botulism poisoning, his evidence was clearly at odds with and contradictory to that of Dr Troskie.

In respect of Dr Fourie, whilst also clearly expert and unbiased, he was, on his own thesis, unable to substantiate the five pillars of a scientific diagnosis, which he set out himself, in respect of botulism poisoning. Just as he was unable, scientifically, to reach a conclusion of phosphine poisoning, so he was unable to substantiate more than his "gut feeling" that this was a case which had botulism written all over it as he said. Just as with Prof Prozesky as referred to above, his reasoning fails on the probabilities that present themselves in this matter, and on the necessary legal test on a balance of probabilities as opposed to the well-established scientific approach based on certainties. Again his evidence contradicted that of Dr Troskie notwithstanding his explanation of alternative symptoms exhibited by animals with acute early botulism poisoning. In my view, Dr Troskie's evidence (and conclusion/diagnosis) prevails." (paragraph numbers excluded)

The submissions

[20] Mr Kruger, who appeared for the appellant, has argued vigorously that the trial court was wrong in two categorical respects, these being (a) that the evidence of the respondent and that of Dr Troskie should not have been accepted given certain contradictions and improbabilities, and (b) given the fairly extensive mechanisms put in place by the appellant to obviate the presence of phosphine in the second screenings and/or the hominychop, and the likelihood of the animals having died from botulism or alternate poisoning, the trial court ought to have found that the respondent had not proved its case on a balance of probabilities and ought, therefore, to have rejected his claim. He also argued that the trial court incorrectly found that there were only two possible sources of contamination. Mr Cole, for the respondent, with equal vigour supported the findings of the trial court.

Law and evaluation

[21] This being a court of appeal, it is once again apposite to bear in mind the proper approach that such a court should take in its assessment and evaluation of an appeal. In this regard:

“The fundamental rule to be applied by a court of appeal is that, while the appellant is entitled to a rehearing, because otherwise the right of appeal becomes illusory, a court of appeal is not at liberty to depart from the trial court’s findings of fact and credibility, unless they are vitiated by irregularity, or unless an examination of the record of evidence reveals that those findings are patently wrong. The trial court’s findings of fact and credibility are presumed to be correct, because the trial court, and not the court of appeal, has had the advantage of seeing and hearing the witnesses, and is in the best position to determine where the truth lies. See the well known cases of R v Dhlumayo and Another 1948(2) SA 677 (A) at 705 and the passages which follow; S v Hadebe and others 1997 (2) SACR 641 (SCA) at 645; and S v Francis 1991 (1) SACR 198 (A) at 204c-f.”[5]

[22] An appeal court should not, however, regard itself as being hamstrung by the rules set out in Dhlumayo’s and subsequent cases, but should recognize a degree of flexibility depending on the type and nature of the evidence under review. In this regard:

"The principle that an appellate court will not ordinarily interfere with a factual finding by a trial court is not an inflexible rule. It is a recognition of the advantages that the trial court enjoys, which the appellate court does not. These advantages flow from observing and hearing witnesses, as opposed to reading 'the cold printed word'. The main advantage being the opportunity to observe the demeanour of the witnesses. But this rule of practice should not be used to 'tie the hands of appellate courts'. It should be used to assist, and not to hamper, an appellate court to do justice to the case before it. Thus, where there is a misdirection on the facts by the trial court, the appellate court is entitled to disregard the findings on facts, and come to its own conclusion on the facts as they appear on the record. Similarly, where the appellate court is convinced that the conclusion reached by the trial court is clearly wrong, it will reverse it."[6]

[23] The court a quo approached the matter on the basis that a number of issues between the parties flowed from mutually destructive versions, and fully analysed the approach which the courts are to take where such is the situation. In my view, however, this matter took a slightly different course to the classic situation of "mutually destructive versions". It is so that the only eyewitnesses to the relevant facts as to what happened in relation to the collection, mixing and feeding of the livestock, and the subsequent deaths of the livestock, were the respondent himself and Dr Troskie. The contended "destructive" versions on the disputed issues put up by the appellant relied almost solely on expert opinion and inferences to be made therefrom.

[24] In dealing with the approach to expert evidence the trial court had this to say:

"A comprehensive summary of an expert witness's duties, made by Creswell J in an English case, (National Justice Companza Naviera SA v Credential Assurance Co. Ltd) also reflects the South African desiderata and has been cited with approval by Davis J in Schneider NO & others v Aspeling and another 2010 JOL 24953 (WCC) : (South African law of Evidence at 330). His evidence must be, and must be seen to be "uninfluenced by the exigencies of litigation" in its form and content; he must not assume the role of an advocate but must give an unbiased opinion on matters that are within his expertise; he should state the facts or assumptions upon which his opinion is based; he should not omit to consider matter that would detract from his opinion; he should make clear when a particular question or issue falls outside his expertise; if he has not fully researched his opinion, he must say that it is provisional; and if something stated in his opinion requires quantification, he must say so: Lexis Nexis: Essential evidence: Zeffert and Others; Chapter 10 para 4 (b). (Zeffert)”

[25] In Michael and Another v Linksfield Park Clinic (Pty) Ltd and Another [7]the approach of a court in assessing the evidence of an expert was set forth as follows:

" [36] That being so, what is required in the evaluation of such evidence is to determine whether and to what extent their opinions advanced are founded on logical reasoning. That is the thrust of the decision of the House of Lords in the medical negligence case of Bolitho v City and Hackney Health Authority [1997] UKHL 46; [1998] AC 232 (HL (E)). With the relevant dicta in the speech of Lord Browne-Wilkinson we respectfully agree. Summarised, they are to the following effect.

[37] The Court is not bound to absolve a defendant from liability for allegedly negligent medical treatment or diagnosis just because evidence of expert opinion, albeit genuinely held, is that the treatment or diagnosis in issue accorded with sound medical practice. The Court must be satisfied that such opinion has a logical basis, in other words that the expert has considered comparative risks and benefits and has reached 'a defensible conclusion' (at 241G - 242B).

[38] If a body of professional opinion overlooks an obvious risk which could have been guarded against it will not be reasonable, even if almost universally held (at 242H).

[39] A defendant can properly be held liable, despite the support of a body of professional opinion sanctioning the conduct in issue, if that body of opinion is not capable of withstanding logical analysis and is therefore not reasonable. However, it will very seldom be right to conclude that views genuinely held by a competent expert are unreasonable. The assessment of medical risks and benefits is a matter of clinical judgment which the court would not normally be able to make without expert evidence and it would be wrong to decide a case by simple preference where there are conflicting views on either side, both capable of logical support. Only where expert opinion cannot be logically supported at all will it fail to provide 'the benchmark by reference to which the defendant's conduct falls to be assessed' (at 243A - E).

[40] Finally, it must be borne in mind that expert scientific witnesses do tend to assess likelihood in terms of scientific certainty. Some of the witnesses in this case had to be diverted from doing so and were invited to express the prospects of an event's occurrence, as far as they possibly could, in terms of more practical assistance to the forensic assessment of probability, for example, as a greater or lesser than fifty per cent chance and so on. This 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 the position where he applies to the expert evidence the standards which the expert himself will apply to the question whether a particular thesis has 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'. "

[26] It was held in Motor Vehicle Assurance Fund v Kenny[8], a case involving a motor vehicle collision, that the credible evidence of an eyewitness should be preferred to that of an expert in accident reconstruction. Kenny's case was quoted with approval by Plasket AJA in Roux v Hatting[9]. Roux’s case involved a serious injury arising from a game of rugby where the court preferred the evidence of an eyewitness to that of experts. In similar vein to the present matter, the experts in Roux’s case weren't present at the rugby game but made their findings from their observations of video clips, together with photographs distilled from these[10].

[27] Addelson J in Putzier v Union and South West Africa Insurance Company Limited[11] dealt with the manner in which a court should approach the evidence where the testimony of an eyewitness is challenged by expert evidence as follows:

“It seems to me however that unless the opinion of the experts is either uncontroverted or uncontrovertable, one should look first at the evidence of the eye witness, if any. If such eye witnesses are unacceptable then naturally the Court is bound to decide, if possible, which of the opinions of the various experts is preferable and to found its judgement on such opinion. On the other hand, where a choice can be made on a balance of probabilities and on accepted principles between two sets of eye witnesses, the Court should first make a provisional assessment of which of the versions of the eye witnesses is acceptable. Having provisionally accepted one or other version, the Court should then consider the expert evidence and decide whether that evidence displaces the provisional findings made on the evidence of the eye witnesses. In this regard, where the onus is on the plaintiff and there is a dispute between the experts, it is my view that, if the eye witnesses favour the plaintiff, the evidence of the defendant must be shown to displace that of the plaintiff’s eye witnesses; but, if the eye witnesses favour the defendant, the plaintiff must show that the evidence of his experts must be accepted in preference

to the experts and the eye witnesses for the defendant. If, at best, the Court is left in doubt as to whether the experts for the plaintiff have advanced opinions preferable to those of the defendant, then it seems to me that the plaintiff has failed to displace the findings made in respect of the eye witnesses and has consequently failed to discharge the onus on him.” (my emphasis)

[28] It is of importance, in evaluating the evidence, to understand that the evidence of the respondent together with the factual input given by Dr Troskie were of great importance, if not crucial, to the success of the respondent's case. The learned judge in the court a quo was acutely aware of this and spent much time in his judgment examining and evaluating the factual evidence provided by these witnesses. As I have indicated, the appellant has set much store by the apparent "external contradictions" which, so it was submitted, ought in the learned judge's mind to have established that these witnesses were not truthful in many vital respects, such as the period of time over which the animals died and the cleanliness of the respondent's barn.

[29] In my view there is little merit in the appellant's argument that the court a quo was wrong in its assessment of the credibility of the respondent. Quite apart from the positive demeanour findings with respect to both the respondent and Dr Troskie, such apparent contradictions to which the appellant refers were fully dealt with by the court a quo and found to be more apparent than real, or of relatively insignificant importance.

[30] Having concluded that the respondent was in his view an honest witness, Lowe J further analysed the evidence of Prof. Prozesky and Mr Van Aswegen with a view to establishing whether their evidence as to the length of time over which the animals died would be such as to dislodge his findings on the credibility of the respondent. Having read the evidence of these witnesses closely, I have to agree that Prof. Prozesky in particular came across as being somewhat dogmatic in his views and was impatient whilst under cross examination being "sometimes anxious to answer…". Even he, however, had to concede that the task set for him was a difficult one, viz. to make a diagnosis (from the photographs) as to the relative time periods over which the animals had died prior to the dates on which the photographs depicting the animals had been taken.

[31] Apart from this, however, in my view the most compelling argument against the appellant's submissions in this regard is that, if Prof. Prozesky were to be correct as to the deaths of these animals, this would mean that the respondent had simply falsified his evidence in this regard. This would mean further that, commencing from a base where he had experienced no unusual deaths amongst his stock prior to the administration of this particular feed, he must have suffered a series of deaths subsequent thereto over a period of some weeks, even months. Precisely why this might have occurred is not at all clear, but at some stage he would have had to have made a conscious decision to falsely truncate the time period over which the animals died, solely for the purposes of bolstering his claim against the appellant. In my view this is extremely unlikely, particularly in the light of the credibility findings made by the trial court. Had this witness been so inclined to lie to this extent one would have expected him to have been tripped up under cross examination, which he was not. Furthermore, it is also most unlikely that Dr Troskie would not at some stage have become aware of this deceit and I have little doubt that he would not have been prepared to place his professional reputation on the line in this regard. The trial court correctly summed this aspect up by concluding: "In my view this head–on dispute must clearly be determined in plaintiff's favour, seen also in the light of the inherent probabilities, the absence of any reasonable explanation as to why the animals would have continued to die over a lengthy period,

without the further involvement of Dr Troskie, and without any explanation, save the possibility that the existing carcasses carry

the possibility of botulism contamination."

[32] Once this is accepted, then in my view the evidence of Prof. Prozesky and Mr Van Aswegen relating to the period of time over which the animals died cannot be regarded to have dislodged the positive finding made by the trial court as to the respondent's credibility.

[33] With regard to the question raised by the appellant as to whether or not botulism was excluded as a probable cause of death, the trial court relied extensively on the evidence of Dr Troskie and, again, made positive findings as to his demeanour. This was likewise crucial to the respondent's case as Dr Troskie's examinations of the dead cattle during the days after they had eaten the impugned feed was of vital importance. The appellant has made much of the fact that Dr Troskie referred in his report to the fact that the tongues of the cattle were "hanging" out. Dr Troskie, in evidence, made it clear however that what he had meant in his report was that the tongues were sticking out and were stiff or "frozen". The trial court found him to have been truthful in this regard noting that, not only did he come across well in the witness box, but that he had absolutely no reason to lie about this fact, Lowe J stating that his "candour and demeanour seemed to me to be beyond reproach nor did he exhibit bias either latent or patent." Lowe J further stated, in dealing with the submission that Dr Troskie had contradicted himself as his report referred to suspected cyanide poisoning and not phosphine poisoning, that: "Not only did he deal satisfactorily with this challenge but in his evidence seen in its totality, he was disclosed as a reliable honest witness, who is most unlikely to adjust or tailor his evidence to dishonestly support plaintiff's case."

[34] In this regard, it was urged on us that the "sheep experiment" at Onderstepoort took the matter nowhere as, after the sheep had been fed the contaminated feed for a period of time, the experiment was abandoned. In my view, this confuses the scientific approach as to certainty with the question of legal probability. The fact that these sheep did not die from eating the feed, must surely be a further aspect to be placed in the "mix" of evidence. The abandonment of the experiment cannot simply negate the fact that the sheep did not die from eating the feed. This fact, together with the other factors such as the inoculations against botulism, gives added force to the probabilities in favour of botulism as not having been the culprit. This is all the more so as the experts were agreed that botulism was a potent poison and that it had to have been present in the mix in substantial quantities, for which there was no logical probable explanation. In my view, accordingly, the trial court did not err in finding that phosphine was the far more probable cause of death.

[35] With regard to the argument that the trial court erred in finding that there were only two possible sources of contamination, the court found that, in all the circumstances, the overall probabilities excluded the possibility of any other contaminant given the cleanliness of the mixing room and other factors. On a conspectus of all the evidence, I can find no fault with this finding. In addition, as submitted by Mr Cole, all these other possible causes of death pale into insignificance when one considers that animals died on two separate farms some 120 km apart, the only common denominator having been the feed mixed on Amazango, part of which was transported all the way to Floukraal.

[36] Despite the vigorous attack on the findings by the trial court in favour of the respondent, I remain unpersuaded that it was incorrect in its various conclusions. In the final analysis and excluding, as one must on all this evidence, the probability that these animals died of botulism, one has the facts that the appellant's process required the use of aluminium phosphide as a fumigating agent, that these animals died after having eaten the feed mix which contained by-products emanating from the appellant's plant, that the animals exhibited symptoms of phosphine poisoning, and that samples of the feed mix were found to have had traces of phosphine in them some weeks later. These factors, amongst others, persuaded the court a quo that the respondent had established its case on a balance of probabilities and I cannot fault that conclusion.

[37] In the result;

The appeal is dismissed with costs.

R E GRIFFITHS

JUDGE OF THE HIGH COURT

BLOEM, J. :

I agree

MBENENGE, J. : I agree

COUNSEL FOR APPELLANT : Mr Kruger

INSTUCTED BY

: Nettletons Attorneys

COUNSEL FOR RESPONDENT : Mr Cole

INSTRUCTED BY

: Wheeldon Rushmere & Cole

HEARD ON : 15 AUGUST 2016

DELIVERED ON : 20 SEPTEMBER 2016

[1] Kroonstad Westelike Boere Vereniging Bpk v Botha 1964 (3) SA 561 (A); D and H Piping Systems (Pty) Limited v Trans Hex Group Ltd [2006] ZASCA 29; 2006 (3) SA 593 (SCA); Holmdene Brickworks (Pty) Ltd v Roberts Construction Co Ltd 1977 (3) SA 670 (A)

[2] He also tested for ammonia, arsenic and nitrates all of which proved to be negative.

[3] The "half-life" of phosphine is reflected in its reactivity which is the time taken for 50 percent of a given quantity thereof to decay.

[4] Judgment of the court a quo

[5] S v Leve 2011 (1) SACR 87 (ECG) at paragraph [8]

[6] Bernert v ABSA 2011 (3) SA 92 (CC) at para 106

[7] 2001 (3) SA 1188 (A);[2002] 1 All SA 384 (A)

[8] 1984 (4) SA 432 (E)

[9] 2012 (6) SA 428 (SCA) at para 20

[10] Roux v Hatting (supra) at para 19

[11] 1973 ECD - (unreported - quoted in Abdo NO v Senator Insurance Company Limited & Another 1983 (4) SA 721 at 725F - 726A)