Rolfe v Katunga Lucerne Mill Pty Ltd

CATCHWORDS: PRODUCT LIABILITY - Sale of Goods - Merchantable quality - causation - no novus actus interveniens when plaintiff's horses poisoned by contaminated chaff from supplier in circumstances where plaintiff aware of rumours of possibility of contaminated product but not notified - Whether proper inference drawn from facts - Admissibility of evidence.

LEGISLATION CITED: Sale of Goods Act 1923 (NSW) ss19(1) and (2)

Alexander v Cambridge Credit Corporation Ltd (1987) 9 NSWLR 310 Annikin v Sierra (2004) 79 ALJR 452 Mahoney v J Kruschich (Demolitions) Pty Ltd (1985) 156 CLR 522 March v E & M H Stramare Pty Ltd (1991) 171 CLR 506 CASES CITED: Hadley v Baxendale [1854] 9 Ex 341 at 354; 156 ER 145 Lexmead (Basingstoke) Ltd v Lewis (1982) AC 225 State Rail Authority of New South Wales v Earthline Constructions Pty Ltd (in liq) (1999) 73 ALJR 306 Warren v Coombes (1979) 142 CLR 531

Mark ROLFE (Appellant) PARTIES: KATUNGA LUCERNE MILL PTY LIMITED (First Respondent) ROBANK FARM SUPPLIES PTY LIMITED (Second Respondent)

B J GROSS, QC/ T BOYD (Appellant) COUNSEL: S A CAMPBELL, SC/ A J BOWEN (Respondents)

Leitch Hasson Dent (Appellant) SOLICITORS: Henry Davis York (Respondents)

• 40 - IN THE SUPREME COURT OF NEW SOUTH WALES COURT OF APPEAL CA 40509/04 DC 1495/03

HODGSON JA SANTOW JA McCLELLAN AJA

28 JULY 2005 Mark ROLFE v KATUNGA LUCERNE MILL PTY LIMITED & 1 Ors Judgment

1 HODGSON JA: I agree with Santow JA. 2 SANTOW JA: OVERVIEW The appellant, a successful harness horse-trainer and driver, Mr Mark Rolfe, unsuccessfully sued Robank Farm Supplies Pty Limited ("Robank"), the second respondent, for damages as a consequence of being supplied contaminated chaff, itself sourced from a Victorian supplier, Katunga Lucerne Mill Pty Limited ("Katunga"), the first respondent. The latter played no role in this appeal. 3 Before the trial judge, Hughes DCJ, the claim was based both on tort and contract, the latter relying upon an implied warranty as to the quality or fitness for purpose of the chaff supplied, and, though not expressly pleaded, its merchantable quality; see ss19(1) and (2) of the Sale of Goods Act 1923 (NSW). 4 While written submissions and the grounds of appeal encompassed both alleged contract liability and tort liability, argument on appeal centred upon breach of the implied warranty of merchantable quality under s19(2) of the Sale of Goods Act. It was common ground that the pleadings could, to the extent necessary, be amended to make that basis of claim explicit. This obviated the need to consider further any grounds of appeal relating to fitness for purpose (s19(1) of the Sale of Goods Act). Having regard to the conclusions reached in this judgment, I have not found it necessary to deal with those grounds of appeal based on a claimed breach of a duty of care on the part of the respondents. 5 Robank, however, maintained that: (a) it had neither been established as a matter of fact, nor found by the trial judge, that such harm as befell the appellant's horses could be attributed to the contaminated chaff, and in particular (b) causation had not been made out given what the appellant knew or must be taken to have known concerning the contamination prior to his feeding his horses with the contaminated chaff; in short, that he either did so knowingly or otherwise behaved so unreasonably as to fail to show that any loss suffered resulted from any breach of warranty by the respondents. 6 I have expressed (b) as I have done, in order to embrace both elements of Robank's resistance to any liability for breach of the contract based on s19(2). Those elements comprise (i) supervening cause not attributable to any breach of warranty (feeding the horses knowing the chaff was contaminated or at least likely to be) and (ii) remoteness in relation to any losses which the appellant suffered. The reference in (a) above is to the causation in fact. It relies on making good the proposition that there was neither a finding of causation nor the basis for a finding in the evidence before the trial judge; namely, that the contaminated chaff materially contributed to the damage, quite apart from any unreasonable conduct on the part of Mr Rolfe. 7 Robank acknowledged that if the appellant were successful on each of (a) and (b) above, it must succeed on its contractual claim for breach of an implied warranty of merchantable quality under s19(2). It was therefore upon these issues that the appeal was ultimately argued.

SALIENT FACTS 8 Set out below are the salient facts, indicating areas of factual dispute. Where appropriate I have made reference to findings of the trial judge. 9 Mr Mark Rolfe was a harness horse-trainer, driver and owner. From 1998 he was conducting his training activities from a leased property at Ebenezer called "Cara Palida". Mr Rolfe leased part of the property from Ms Elaine Bronson who lived on the premises. He was sometimes contacted through her, as he did not have a fixed-line telephone (though he did have a mobile phone, the number of which changed from time to time before August 1999). In his business Mr Rolfe employed a Mr Halliday as a stable-hand. 10 Mr Rolfe did not live on the Ebenezer premises, which were horse stables and training facilities rather than domestic premises, although the trial judge stated "[I]t is not clear whether he lived on these premises." (Red, 14D). 11 As at August 1999, Mr Rolfe had in training at the property, eleven standard breed horses. 12 Mr Rolfe sourced his feed supplies from Robank, a farm produce merchant, which operated a store in Ebenezer. The directors of Robank were Mr Frank van den Nieuwboer and his wife. Also working in the business were his son John van den Nieuwboer, and daughter Mrs Marianne Roughley. Robank also employed a Mr Andrew Palmer as a storehand and deliveryman. Robank had cash and account customers. 13 On 5, 7, 10 and 16 August 1999, Mr Rolfe purchased from Robank on account a total of seven bags of chaff (at $17.50 per bag), part of a batch supplied to Robank by Katunga on Tuesday, 3 August 1999. This chaff was known variously as "Katunga Lucerne chaff" or "Victorian chaff", and was distinguished from NSW Lucerne chaff. Mr Rolfe had formed a preference for the Katunga Lucerne chaff, because he believed it was of better quality than the NSW chaff. 14 The bags of chaff bought by Mr Rolfe were contaminated with botulism toxin, as indeed was the batch out of which they came. 15 On Tuesday 17 August 1999 or Wednesday 18 August 1999 a new batch of Katunga Lucerne Chaff from Katunga arrived at Robank and was being sold to purchasers who asked for it from sometime on Thursday 19 August 1999 (Black, 600O-S). 16 On Tuesday 17 August 1999 Dr John Parbury, a Veterinary Surgeon, phoned Mr John van den Nieuwboer of Robank. According to the second respondent's diary for 17 August 1999 Dr Parbury told him "Sue Jackson's horse had died from (suspected) botulism which he thinks is linked to Robanks" (Blue, 590H-J, Black, 635F-H). This call was at 10 am (Black, 635B). Dr Parbury also told Mr John van den Nieuwboer that there was another client in the area that might also have the same problem and that he believed "also came from the feed" (Black, 635H-J). Dr Parbury did not tell him anything about what specific product he thought might be involved (Black, 635K). 17 Later on Tuesday 17 August 1999, at about lunchtime, Belinda Miller, another horse owner from a different property who had also bought Katunga Lucerne Chaff from Robank, told Mr John van den Nieuwboer by telephone that her horse had died on 16 August 1999. So by lunchtime on 17 August 1999, Mr John van den Nieuwboer suspected that this death was something to do with the chaff and that further investigations needed to be made (Red, 18Y-19C). 18 Later on Tuesday 17 August 1999, at about 6.45 pm, Belinda Miller rang Mr John van den Nieuwboer of Robank and told him she had sieved some Katunga Lucerne Chaff and found pieces of faeces, skin and fur in it. By this time Mr John van den Nieuwboer was suspicious that the product in question was Katunga Lucerne Chaff and he instructed his staff not to sell any Katunga Lucerne Chaff (Red, 19E-I). 19 On Wednesday 18 August 1999 a sign was placed on the counter at Robank warning people to return the chaff that had been delivered on 3 August 1999 (Red, 20M). 20 On Wednesday 18 August 1999 Robank had instituted a ban on the sale of Katunga Lucerne Chaff (Red, 20N). 21 By early morning of 19 August 1999, John van den Nieuwboer concluded that the concern definitely centred around Katunga chaff and instructed his staff to recall the product, and make a list of purchasers of it (Red, 19W). 22 Later on that same day, John van den Nieuwboer had his sister Marianne Roughley and staff telephone all people on the list including Mr Rolfe's landlady Ms Bronson (Red, 21) (who had bought Katunga Lucerne chaff for her own horses) and also Mr Rolfe, until they got an answer, working long into the night (Red, 19X-Y). He sent staff to customers' addresses to pick up "any bags left over", though not bags which were opened. Mr Andrew Palmer then attended on Mr Rolfe's premises pursuant to the recall (Red, 20B-C, 21P, 21W-X) for the purpose of picking up Katunga Lucerne Chaff from Mr Rolfe (Red, 21I-L). Mr Rolfe says that he was never informed by Robank of the problems with the Katunga chaff (nor was his landlady Ms Bronson). He says that the Nieuwboers never got through to Ms Bronson on the telephone and that neither he nor she ever encountered Mr Palmer. 23 As noted above, from Wednesday 18 August 1999, Mr John van den Nieuwboer prohibited the further sale of Katunga Lucerne Chaff and placed a notice in the store warning people of the recall (Red, 20C-D), this being, according to Mr Nieuwboer a "prominent notice" (Red, 21Q). There is dispute about the size of the notice and whether it would have been apparent to Mr Rolfe who was an account not cash customer. 24 Although Mr Rolfe denies he was contacted (Red, 20F), the trial judge found that Robank (as part of the recall) contacted Mr Rolfe but that if Mr Rolfe was not contacted, Robank contacted his landlady, Elaine Bronson, on her direct line and told her about Katunga Lucerne Chaff from the load that was delivered to Robank on 3 August 1999 (Red, 20F-I). That is a disputed matter, to which I will return. 25 After Wednesday 18 August 1999 Mr Rolfe purchased only NSW lucerne chaff (Red, 21Q-R). On 19 and 22 August, Mr Rolfe attended the Robank store to purchase more chaff, and bought a total of three bags of NSW chaff, the Katunga chaff being unavailable. (Red, 20P-Q, 21D). The trial judge inferred that Mr Rolfe "must have had some intelligence as to doubts surrounding the chaff since the purchase was that of his non-preferred NSW chaff" (Red, 20R-S). That is disputed, as is whether that fact was capable of grounding any such inference, when, on the facts, the Victorian chaff was no longer available so there was no choice. 26 Mr Rolfe says he never saw a sign at the store warning of risks with the Katunga chaff, nor was he told by Frank van den Nieuwboer who served him of the problems with the Katunga chaff. 27 On Friday 20 August 1999 Mr Rolfe "heard rumours" and a report from his landlady, Elaine Bronson, about Katunga Lucerne Chaff (Red, 20U-V, 21O-P). 28 On Friday 20 August 1999 Mr Rolfe went to the Robank store and spoke to Frank van den Nieuwboer (Red, 20V-21B), who, according to the trial judge's finding, did not reassure Mr Rolfe, as Mr Rolfe alleged, but "warned [Mr Rolfe] that there was a problem with Katunga Chaff" (Red, 21B-C). That is disputed, based on Frank van den Nieuwboer's denial that any conversation took place with Mr Rolfe on the occasion in question. 29 On Monday 23 August 1999 Mr Rolfe and Mr Halliday fed the horses with the Katunga chaff allegedly contaminated with botulism toxin, by tipping it into the feed barrel. The trial judge concluded that at that time Mr Rolfe had "knowledge" of "the doubts concerning Katunga Chaff" as the result of the following "pieces of acceptable evidence" (Red, 21M-R) and that he knew that to feed the Katunga Lucerne Chaff to his horses was a "very risky business" (Red, 21U) on the following bases: (a) he heard rumours from his landlady Ms Bronson (Red, 21V), these rumours being " warnings of the possible dangers of Katunga chaff " (Red, 22E-F); (b) after Wednesday 18 August 1999, he bought NSW chaff on two occasions and not his preferred Katunga chaff; (c) Andrew Palmer visited the premises pursuant to the recall; (d) on Friday 20 August 1999 Mr Rolfe spoke to Frank van den Nieuwboer when visiting the Robank store; and (e) there was a prominent notice in the store. The trial judge's overall conclusion was that Robank had warned Mr Rolfe of the "possible danger" of feeding his horses Katunga Lucerne Chaff "most probably via Elaine Bronson who shared the premises with the plaintiff" (Red, 22B). That conclusion and much of the evidence in support is disputed by Mr Rolfe, as are inferences sought to be drawn from that evidence. 30 The trial judge noted that Mr Rolfe at first in his direct evidence gave the distinct impression that "he did not know that the feed was possibly contaminated" though "he thought it was a different colour than usual" (Red, 17O-P). The trial judge observed however that "Indeed, he did not believe, nor could one believe that he would have fed contaminated chaff to the horses. In cross-examination he said he was reassured by the second defendant that the chaff was safe. There could appear to be a mistake in Mr Rolfe's understanding of the reliability of the chaff he fed to the horses on the evening of 23 August 1999" (Red, 17Q-T).

31 However, the trial judge then concluded, with respect contradictorily, that when Mr Rolfe fed the horses on the evening of Monday 23 August 1999, he "knew that the Katunga Chaff was possibly dangerous" (Red, 22H). This is disputed by Mr Rolfe. 32 On Tuesday 24 August 1999, the horses had decreased appetite and became distressed after exercise. On 26 August 1999, Mr Rolfe rang his veterinary practice, and Dr Major attended the premises the following day (27 August 1999). 33 On 28 August 1999, one horse, Sharon's Jewel, was put down. The other ten horses were affected requiring them to be spelled from racing, and were seen again by another vet Dr Duckworth on 9 September 1999. When some of them ultimately returned to training in January 2000, only a few were successful. 34 Mr Rolfe claimed to have sustained loss of income as a trainer because his horses could not race and because his reputation and goodwill was affected by news of the botulism outbreak at his stables. Ultimately his horse-training business became unprofitable and he was obliged to give up the lease of his training establishment. 35 Mr Rolfe commenced proceedings against both Katunga and Robank in the District Court of New South Wales, alleging as against Katunga, negligence in the manufacture and supply to Robank of bags of Lucerne chaff which were unfit for use and consumption by horses because they contained botulism toxin. 36 Mr Rolfe pleaded as against Robank: (a) breach of a condition implied by s19(1) of the Sale of Goods Act 1923 (NSW) and/or s71(2) of the Trade Practices Act that "the lucerne chaff was not reasonably or at all fit for the purpose of consumption by the said horses but contained defects in that it was contaminated with botulism and therefore was not of merchantable quality." (Statement of Claim paras 6-7), or (b) alternatively, negligence in releasing the contaminated chaff to the plaintiff (Statement of Claim para 8) (broadening at trial into an allegation of negligence in failing in a timely manner to warn Mr Rolfe against using the contaminated chaff after becoming aware of it). 37 Robank cross-claimed against Katunga for contribution and/or indemnity. 38 The trial judge dismissed Mr Rolfe's claim, and entered judgment for the defendants. He made no finding on damages (against the possibility of a successful appeal). 39 The trial judge made the following credit findings: (a) Wherever the evidence of Mr Rolfe conflicts with any of the three Nieuwboers (Frank, John and Marianne Roughley), their evidence should be preferred (Red, 18B-E) as they appeared to be reliable witnesses, whereas Mr Rolfe did not (Red, 18E); (b) Wherever the evidence of Mr Palmer conflicts with that of Mr Rolfe and Mr Halliday, Mr Palmer's evidence should be preferred; he was a reliable witness (Red, 18F-G); and (c) Elaine Bronson is not accepted as a reliable witness - she admitted she had an interest in the outcome of proceedings since Mr Rolfe still owed her rent (Red, C-E). 40 The trial judge made the following key findings of fact: (a) Mr Rolfe preferred Katunga Lucerne chaff to NSW chaff because he believed it to be of better quality (Red, 14N-O); (b) The "bad batch" of chaff was sent by Katunga to Robank on 3 August 1999 and was in store on or about 4 August 1999 (Red, 18I-K). Another load of Katunga chaff arrived on 17 or 18 August 1999 (Red, 19K-P); (c) At the time of feeding the horses on 23 August 1999, Mr Rolfe knew that to feed the Katunga chaff to the horses was "a very risky business"; (d) It should be accepted from what the Nieuwboers said (i.e. that Mr Rolfe was contacted or if he was not that his landlady who had the direct phone line was contacted) that Mr Rolfe was informed about the Katunga chaff from the load that was delivered on 3 August 1999 (Red, 20F-I). He was told by Elaine Bronson after she was contacted by the Nieuwboers (Red, 21O-P); (e) Mr Rolfe heard rumours from Ms Bronson as to the possible danger of Katunga Lucerne chaff (Red, 22E-F, 21U); (f) It should be accepted from Mr Frank van den Nieuwboer's evidence that he warned Mr Rolfe when he came to the store that there was a problem with the Katunga chaff (Red, 20Y-21C), Red 21P-Q); (g) It should be accepted from Mr Palmer's evidence that it is more probable than not that he visited Mr Rolfe's training establishment on 19 August 1999 for the purpose of picking up Katunga chaff from Mr Rolfe (it was not put to him that this did not occur) (Red, 21I-L, P,W-X); (h) There was a prominent notice in the store (Red, 21Q); (i) Mr Rolfe only bought NSW chaff after Wednesday 18 August 1999. 41 The trial judge reached the following conclusions on liability: (a) Since Mr Rolfe knew that the Katunga chaff was possibly dangerous, he was either 100% contributorily negligent or the Katunga chaff was fed inadvertently to the horses (Red, 22G-I); (b) The recall was carried out in a reasonable and prudent manner by Robank, and was carried out thoroughly (Red, 22J-L); (c) Section 71 of the Trade Practices Act did not apply, since Mr Rolfe is not a consumer pursuant to s4B, and the price of the goods at $17.50 per bag does not exceed the prescribed amount (Red, 22N-P); (This is not in dispute on appeal.) (d) Section 19(1) of the Sale of Goods Act does not apply as there is no evidence at all that Mr Rolfe relied on the skill and judgment of Robank, and in fact he himself applied his own skill and judgment as a horse-trainer in insisting on the Katunga chaff (Red, 22U-23C); (e) Katunga knew nothing of Mr Rolfe until 27 August 1999 and had manufactured the chaff well before that. It follows then that the sale of the chaff did not cause his loss but was more likely than not due to inadvertence or indifference to risk by Mr Rolfe (Red, 23Q-T); (f) Even if there was a duty of care owed by Katunga and Robank to Mr Rolfe, his own action of feeding the suspect chaff was so unreasonable that any duty of care no longer applied (Red, 23V-Y); (g) Mr Rolfe has failed to prove his case on the balance of probabilities (Red, 24C).

DISPOSITION 42 Section 19(2) of the Sale of Goods Act is in the following terms: "(2) Where goods are bought by description from a seller who deals in goods of that description (whether the seller be the manufacturer or not), there is an implied condition that the goods shall be of merchantable quality: Provided that if the buyer has examined the goods there shall be no implied condition as regards defects which such examination ought to have revealed." 43 As I have said earlier, the applicability of s19(2) if the matter of causation were resolved in favour of the appellant, is not itself in dispute. It is to that I now turn.

Causation 44 Causation has two aspects. The first can be simply dealt with. It concerns whether, as was put in question in argument on appeal, it can properly be inferred, in the absence of a finding by the trial judge, that the contaminated chaff actually caused the death of the appellant's horse, Sharon's Jewel. 45 The uncontested events are that on 23 August 1999 Mr Rolfe and Mr Halliday fed the horses at his stables with some lucerne chaff that is now admitted to have been Katunga Lucerne chaff (Red, 14Q). 46 The trial judge observed that on the following day, 24 August 1999, "the evidence from the plaintiff Mark Rolfe and his employee Peter Halliday was that the horses had left some feed in their feed bins. During the week there is some evidence that the horses were eating less, their appetite had decreased and they had shown signs of being distressed after they had been exercised.

On the Thursday night the plaintiff put through a call to the vet and the following morning, Friday 27 August, a vet from the practice that Mr Rolfe usually used, Dr Major, came to visit the training establishment. The horse that was thought to be ill was a two year old filly named Sharon's Jewel. Dr Major was not the usual veterinary surgeon who attended Mr Rolfe's horses but was a partner in the practice.

The following day the same vet returned and put down the young filly, Sharon's Jewel, and the body of the horse was taken to the University Veterinary Clinic in Camden, for a necropsy/autopsy." Red, 14S-15H. 47 The trial judge later observed, concerning the evidence from the usual veterinarian Dr Duckworth who had attended the horses on 9 September 1999, making a check on ten horses in the stables, that "it is noticeable neither then, nor in the witness box, when asked the direct question, did Dr Duckworth give a diagnosis of why the horses were in the condition he observed." (Red, 15M-O). 48 Finally, at Red, 17H-L, the trial judge states, "Mr Holmes submits that I should infer from the possible botulism outbreak in other horses in the surrounding area that an irresistible inference arises from the matrix evidence [sic] that any contamination involved in the death of Sharon's Jewel was botulism. However, I'm not sure that I need to go to that for reasons that will appear later." Essentially, those reasons concern the contention, which the trial judge accepted, that Mr Rolfe had knowledge of the possible contaminated chaff before he fed it to the horses and that led to the trial judge concluding against Mr Rolfe. I shall return to that second question shortly. 49 As it turned out, the pathological examination was not done, apparently for cost reasons. The appellant pointed out that Dr Duckworth was Dr Major's partner and it was Dr Major who, as I have said, first attended Sharon's Jewel. Naturally enough, Dr Duckworth having not so attended could form no opinion as to the correct diagnosis. 50 The respondent's contention was that botulism, caused by chemical toxins produced by bacteria of a similar name was highly toxic. The respondent contended that only one of Mr Rolfe's horses dying was indicative of another explanation than the death being caused by botulism, more especially as there had been no pathology examination definitively establishing that. On the basis that the onus remained with the plaintiff, it was put that the absence of a finding by the trial judge meant that a critical element of the plaintiff's case was never made out. 51 To this the appellant responded that it was quite understandable in certain circumstances that one could get a low mortality rate, that horses can be sick and not die in these circumstances, more particularly if they are elite animals stabled together who may be more resistant to the disease. The appellant also relies on the admission of Robank, supplier of the contaminated chaff, in a letter of 17 May 2000 addressed, as it happens, to Katunga. There the following was stated: "the recall action was taken by the company due to the death of six horses to date, caused by ingestion of Katunga Lucerne Chaff which was contaminated with Botulism. We believe that the recall action taken by our company prevented the spread of further illness and subsequent death to other horses owned by our customers who feed their horses on Katunga Lucerne Chaff." 52 No evidence was put forward by Robank contradicting that statement as to the number of horses which died following ingestion of the contaminated chaff, or pointing to those horses not including Sharon's Jewel, owned by the appellant. While the appellant still bore the ultimate onus of proving the essential elements of its case, the evidentiary onus had in my opinion passed to Robank to rebut the inference that the contaminated chaff supplied by Robank to the appellant caused the death of Mr Rolfe's horse Sharon's Jewel. That inference rests not only on the admission in the letter of 17 May 2000, which concededly depends on Robank's means of knowledge as the local supplier, but on the undoubted fact of Robank's supply of that contaminated chaff to the appellant, its likely ingestion by his horses, followed by the death of one of them Sharon's Jewel. Given the absence of any rebuttal, the only question that remains is that of Mr Rolfe's knowledge before he fed the horses and its implications for causation and liability, so far as any loss suffered was concerned. It is to that I now turn. 53 This question was correctly dealt with in argument on appeal as not being a matter going to whether the warranty of merchantable quality was to be implied. Under s19(2) there was no suggestion that, in terms of the proviso to that subsection, examination of the chaff ought to have revealed the botulism. Rather, the question concerned whether, accepting that the implied warranty was capable of application, such knowledge as Mr Rolfe had bearing on the likelihood of contamination by the botulism, would break the chain of causation between the defendant's breach and the plaintiff's loss or damage; cf Lexmead (Basingstoke) Ltd v Lewis (1982) AC 225. The kind of knowledge that would suffice to break the chain of causation would include both actual knowledge on the plaintiff's part of the contamination or a deliberate hiding of the eyes from that information when the plaintiff had every reason to believe it was contaminated. However, mere carelessness on the part of Mr Rolfe as customer could not be enough to break the chain of causation. Much of the reasoning of the trial judge in seizing upon carelessness on the part of Mr Rolfe was to that extent, with respect, in error. 54 What then is the evidence for whether Mr Rolfe as customer knew of the problem, that is of the possibility of botulism contamination, yet wittingly went ahead and fed the horses with that infected chaff, or deliberately hid his eyes from the information and fed his horses with the contaminated chaff? 55 To answer that question one needs to look more closely at the evidence which the trial judge relied upon to reach the conclusion adverse to Mr Rolfe that he did. That evidence centres around the steps taken to recall the contaminated chaff to see whether those steps necessarily brought to the attention of Mr Rolfe the contamination or at least its likely possibility. 56 In summary, the trial judge's findings that the plaintiff received the necessary warnings, information and knowledge concerning the risks from Katunga chaff were based on conclusions that: (a) On Thursday 19 August 1999 Robank (the Nieuwboers) contacted Elaine Bronson by telephone, and she then passed on the relevant information to Mr Rolfe; (b) On Thursday 19 August 1999 Mr Andrew Palmer of Robank visited the plaintiff's premises pursuant to the recall; (c) On Friday 20 August 1999 Mr Frank van den Nieuwboer warned the plaintiff that there was a problem with the Katunga chaff; (d) From Wednesday 18 August 1999 there was a sign on the counter at Robank warning customers of the recall, which the plaintiff could or did see on Friday 20 August 1999, when he visited the store.

Did Ms Bronson inform Mr Rolfe of the contamination or of its likelihood? 57 I take each of these in turn and look at the primary evidence said to underpin these conclusions. I start by noting that in relation to proposition (a) the trial judge did not accept Elaine Bronson as a reliable witness because she admitted she had an interest in the outcome of proceedings, since Mr Rolfe still owed her rent (Red, 15U and 22C-E). Given that the trial judge did not state any other basis for not accepting Ms Bronson as a reliable witness than that she had an interest in the outcome by reason of the plaintiff still owing rent, while according the trial judge the advantage of having observed the demeanour of the witness, the stated reason alone could not of itself be sufficient to reject her evidence. 58 The evidence bearing on this includes that Ms Bronson denied being contacted by the Nieuwboers or receiving any message left on her answering machine, which was in working order at the time and which she regularly checked (Black, 783U, 784S, 787K). So far, if that evidence were accepted, the communication was not with Mr Rolfe direct but through his landlady, in circumstances where Mr Rolfe did not live on the premises but used them for his training purposes. The trial judge accepted that any such communication was through Ms Bronson rather than direct. 59 However, there is a further significant piece of evidence not referred to at all by the trial judge. The recall list prepared by Robank's staff, being Exhibit AR (Blue, 711-725) tendered by Mr Rolfe as plaintiff (Black, 601Q) shows the method used for checking off customers from that list. This was done as they were contacted either by circling the names, putting a tick against their names, or putting a line through the names (Roughley Black, 612X). 60 At Blue, 711O, Ms Bronson's name appears as "Ms E Bronson (BronE), her address "Grono Farm Road, Ebenezer", with a sale on "3/8" and a reference to "1 bag", which can be taken to have been bought by her personally. There was a duplicate copy of the same page but not with any marking (Blue, 712O). Her name appears nowhere else in Exhibit AR. Ms Bronson's name, unlike numerous other names, is neither circled nor ticked and does not have any line through it. However, next to her name in handwriting appears "pick up current contact" (Blue, 711O). Other names have "left message" (Blue, 711R, 714S, 714U, 714V, 714X, 718D, 722K, 722M). Furthermore, there is no telephone number alongside Ms Bronson's name, even though phone numbers of numerous other persons are written in. 61 Mr Frank van den Nieuwboer who identified the list (Exhibit AR) said that he "wasn't there on the spot" (Black, 551H), that two girls went through the invoices, made a list of customers in order to contact them, and were telephoning those customers (Black, 551-2). 62 There was further evidence from Frank van den Nieuwboer's daughter, Marianne Roughley. She did all the accounts at Robank and said that her brother John on Thursday 19 August 1999 instructed her to go through every single invoice from the beginning of August 1999 to 19 August 1999. This was to work out who had bought that one product, Katunga Lucerne chaff, which he had "basically pinpointed" or "narrowed" down to being the problem (Black, 610B-S). She, her brother John and another female worker, worked through until about 10 pm or 11 pm on Thursday night 19 August 1999, going through every single invoice, and made a list, wrote down the telephone numbers of the customers, wrote down how many bags they had bought, and then rang them up (Black, 611B-J). 63 According to Marianne Roughley, "contact" of customers was either by telephoning them or, "if they weren't phoned, we had verbally spoken to them" (Black, 613H). They would "verbally speak to them without phoning them" by going to their place (Black, 613 I). Her brother, John, did the going out to their places (Black, 613K). At the time when they finished at 10 pm or 11 pm on 19 August 1999 she was satisfied that everyone on the list had been contacted (Black, 613F). 64 To the recollection of Marianne Roughley it was she whose hand put a blue line through the name of Mr Rolfe for each of the dates when his name appeared (Black, 614Q-615R). She said it was her brother, John, who did the circles. Other names (but not Elaine Bronson's name or the Mr Rolfe's name) have ticks against their names. Marianne Roughley later added that she put the single line through Mr Rolfe's name (Blue, 715T, 718C, 719 C). Ms Bronson's name only appears at Blue, 711O, a duplicate of which is at Blue, 712O. 65 Marianne Roughley said that the line through the information set out in Exhibit AR meant that they were contacted (Black, 614W) which means that "I rang them" (Black, 614Y). However, immediately following is this question and answer: "Q. It doesn't mean he spoke to anyone? A. No I rang them." (Black, 615B). 66 When asked to clarify this in examination-in-chief, Marianne Roughley, in answer to the question "At the time you rang, did you get any answer?", replied "I cannot remember that. I don't know. I can't - I don't remember." (Black, 615G). In the case of people whom she rang and got no response from, she wrote a list on the back of "the people we didn't get hold of and my brother went and seen them" (Black, 615 I). In cross-examination she added that the two pages at the back (Blue, 724-725) was the list of customers she wasn't able to contact, and that "obviously by looking at that list you can see that after that someone has been able to contact, being either my brother or the other staff member" (Black, 615W-616D). 67 However in cross-examination, Marianne Roughley stated that to the best of her recollection she did the stroke through and other ticks, but that she did not have "a recollection of contacting Mark Rolfe" (Black, 618D). She did not know that the telephone number written alongside Mark Rolfe's name (4579 9244) (Blue, 715T) was not his telephone number (Black, 618M). To the best of her recollection she had got that telephone number off the computer, under customer details (Black, 618G). When it was suggested to her that it was not Mr Rolfe's telephone number and that she did not make that telephone call to Mr Rolfe, Marianne Roughley asserted, "All I know is I made sure I compiled a list … and every person that was on that list we contacted" (Black, 618Q), but then added "All I can remember is that every name on there, whatever number was there, we rang. If we didn't ring, if we didn't get hold of them, then my brother or my father spoke to them in person" (Black, 618W-X). 68 Marianne Roughley then asked if she could add something and added "To the best of my recollection, he had a manager or a lady that, on many occasions we never could get hold of Mark and we would speak to a lady by the name of Elaine Bronson, so that (4579 9244) possibly could be her number" (Black, 619H-K). However she said that she did not have a recollection of contacting Elaine Bronson, and as it was 5 years ago she did not remember talking to any of them (Black, 619N). 69 Marianne Roughley later in cross-examination acknowledged that not just at this time but not at any time did she ever contact Mr Rolfe and speak to him personally (Black, 629F-N). She said that if she had to contact Mr Rolfe she "always in the past spoke to" (Black, 629 1), "had to speak to a lady called Elaine Bronson because Mark was a very difficult person to get hold of" (Black, 629Q). She recalled that in the past she would leave messages with Elaine Bronson in relation to Mr Rolfe's account. She did not know who Elaine Bronson was to Mr Rolfe, but she knew Elaine Bronson was a customer for a little while. Finally, in cross-examination she said that if she did anything to contact customers it was by way of a phone call, that she did not visit customers on that day (Thursday 19 August 1999) and when it was put to her that she did not contact Mr Rolfe on that day she said "That I don't remember" (Black, 630K-Q). 70 Mr Rolfe's phone number was not on the recall list, as the phone number appearing alongside his name in the recall list (Exhibit AR) was Elaine Bronson's phone number (Blue, 715T; Bronson, Black, 782C). 71 When Mr John van den Nieuwboer gave evidence, he did not give any evidence of any contact that he had made with Mr Rolfe either by telephone or in person. Further, John van den Nieuwboer did not give any evidence of any contact having been made by him with Elaine Bronson. However, he gave evidence that he made the note on the recall list alongside Elaine Bronson's name, "Pick up - can't contact" (Blue, 711O). 72 John van den Nieuwboer said that on Thursday 19 August 1999, after a couple of staff members prepared a list of account customers who had bought Katunga Chaff, he and other staff members commenced ringing until 10 pm. He personally made the circles on two pages of Exhibit AR (Blue, 711 and 714). He said that the writing or marks he put on the page were people he contacted or tried to contact on Thursday 19 August 1999 while other people were engaged in the same or similar work (Black, 662V-663B). On 19 August 1999 there were people, he said, that he could not contact by telephone when he tried to ring them, and what he did about those people was that he "instructed some employees to pick up - go to their premises, and pick up chaff, where possible" (Black, 664U-Y). He added that he instructed them to do that "when I couldn't get hold of them ... on the Thursday" (Black, 665B) and that when he said "Wherever possible" he meant "If they could get access into their place" (Black, 665B-E). 73 John van den Nieuwboer, when shown in cross-examination the first page of Exhibit AR (the recall list) said he had made the entries alongside Ms Bronson's name "Pick up" and "Can't contact" (Black, 699U-F; Blue, 711O). By "Pick up" he meant "well, to the best of my recollection, I would have sent someone around there to pick it up" (Black, 700G). He later added, in relation to another customer's name but describing what occurred on Friday 20 August 1999, "Well, if we couldn't contact them and we can't get access to their place, it would have been a call from one of the other guys if he could get access or not, one of the delivery drivers, well then we'd have to do it on the usual delivery run" (Black, 700P-Q). 74 In re-examination John van den Nieuwboer said that contacting the customers "one way or another" meant "by phone or verbally", that is "If I didn't speak to them by phone, I tried to speak to them verbally, or tried to organise somebody to talk to them verbally" and "If I had to send somebody to pick up the chaff, for instance, they'd see that person to let them know" (Black, 703U). Where possible, if Mr John van den Nieuwboer was there, contact "by telephone or verbally" included "somebody going out to the customer's premises and telling them about it, as for those who came into the shop for some purpose or another, being told then as well" (Black, 704D-G). He said that the pick up from Ms Bronson would have occurred on that day (presumably Friday 20 August 1999). He did not remember who was sent out to do that on that day, and that apart from instructions to pick up the product, instructions were given to give the customers a verbal message, namely "To let them know that there was a problem with the feed and we were just replacing it" (Black, 704S-T).

Did Mr Andrew Palmer of Robank visit the plaintiff's premises pursuant to the recall? 75 Mr Andrew Palmer gave evidence-in-chief that he was employed by Robank as the pre-mixer of feed and to do deliveries and provide service to the customers, that in August 1999 he heard an allegation in relation to stock feed supplies by Robank that they had trouble with the lucerne chaff being contaminated and making a few horses sick, and that he made a visit to Mark Rolfe's property after he had heard that Mark Rolfe's horses got sick and one had died (Black, 708O-709D). He made this visit probably while delivering feeds (Black, 709D). He gave some further evidence about one of the sick horses of Mr Rolfe that died, and taking some formula to Mr Rolfe for treating the sick horses after they fell sick (Black, 709E-710E). 76 In cross-examination, Mr Palmer said that on what he thinks was Thursday sometime prior to the problem (in re-examination he clarified this as being 19 August 1999: Black, 713F), he went to Mr Rolfe's property, as part of his work for Robank going out and replacing the contaminated feed (Black, 711 B-X.). He did the collection of bags of feed for a full day, and cannot remember exactly all the places he went to, because there were too many on that day (Black, 711U, 712X). 77 I agree with the appellant's submission that therefore there was nothing in the evidence of Mr Palmer about any visits by him to the property where he attests to any conversations or contact with Ms Bronson or Mr Rolfe on 19 August or 20 August 1999 or at any time before Mr Rolfe fed his horses the contaminated feed on Monday 23 August 1999. That is significant, given: (a) the only phone number on the recall list was Ms Bronson's not Mr Rolfe's; (b) the note on the recall list alongside Ms Bronson's name "pick up - can't contact"; (c) that Maureen Roughley conceded that she had "no recollection of contacting Mark Rolfe"; (d) she had no "recollection of contacting Elaine Bronson", though recognising this was five years ago. 78 I also agree that it was not proven that the written instruction "Pick up" followed by "Can't contact" in Exhibit AR (the recall list), was ever implemented by Mr Palmer or anybody else from Robank before Mr Rolfe fed his horses on Monday 23 August 1999. 79 I also agree with the appellant's submission that it is very important to note that "Pick up" means collection of full bags. Exhibit P, being the letter from Robank, under the hand of its van den Nieuwboer Managing Director Mr Frank van den Nieuwboer, to the General Manager of Katunga stated, inter alia: "Please find enclosed documentation (page 2) of the costs bourn by our company for the recall action on Katunga Lucerne Chaff. We have deducted the costs of the recall on Katunga Lucerne Chaff and the costs for the replacement of the contaminated Katunga Lucerne Chaff with NSW Lucerne Chaff to our customers. Please note that due to the high costs of labour, any partly used bags of contaminated Katunga Lucerne Chaff were not picked up from customers , which accounts for the difference in the amount of bags of Katunga Lucerne Chaff returned." (Blue, 537T-W) [emphasis added]. 80 It follows that the recall of bags of Katunga Lucerne Chaff from customers involved picking up full bags, but only if access could be obtained to those bags at the property; that is to say, if the occupant were in or full bags were in plain view and able to be replaced by the delivery driver. 81 Because Mr Rolfe and Ms Bronson both deny any visit to the premises or verbal contact with them from anyone from Robank, and Mr Palmer gave no evidence of any such contact with them, the appellant submitted that it was likely that Mr Palmer may have taken a full bag of Katunga Lucerne Chaff if he was able to see it and access it, but did not give any warning to either Ms Bronson or Mr Rolfe. The uncontradicted fact remains, that one of the bags of Katunga Lucerne Chaff which Mr Rolfe had purchased from Robank remained at the property and that on Monday 23 August 1999 Mr Rolfe opened this bag, emptied it into a feed bin and fed his horses contaminated feed. 82 Ms Bronson in evidence-in-chief (who was not accepted by the trial judge) said that her telephone number was 4579 9244 (Black, 780C) (the same number as someone wrote alongside Mr Rolfe's name on the second defendant's recall list (Exhibit AR)). Ms Bronson said (as I have earlier recounted) she was not contacted by phone and no message was left on her answering machine from Robank (Black, 783R-X). The trial judge inferred otherwise, but that inference does not square with the note against her name in the recall list "Pick-up-can't contact". 83 In cross-examination Ms Bronson said that whatever date Sharon's Jewel (one of the plaintiff's horses) died, "All we'd heard on the grapevine that they thought something was being caused by feed, but that nothing had been stated in actual blunt fact to me that it was Katunga Chaff's chaff." (Black, 734V-W). She could not remember or say when she heard that on the grapevine as "it's too long ago" (Black, 734X). She thought it was on the grapevine before Sharon's Jewel died because other horses in the area had already died and she had heard that "roughly in my estimate probably 3, two days, something like that before Sharon's Jewel died" (Black, 785F). Approximately two days before Sharon's Jewel died she first heard something about Katunga Chaff (Black, 785P-Q). When she heard it she mentioned it to Mr Rolfe who at that stage said he had a couple of horses not right and a few horses not eating up (Black, 785V-W). She did not know if she mentioned it to Mr Rolfe on Friday 20 August 1999 or thereabouts as she could not recollect the right date (Black, 785V-W). Later she conceded that on or before Friday 20 August 1999, "around that date, possibly", she "told Mr Rolfe that there was some problem with the feed supplied by Robanks", after describing this "problem" as a "suspected problem not with Katunga chaff in particular as I - nobody else at that stage knew what was causing these animals to die in the area" (Black, 786T-X). 84 Ms Bronson in cross-examination, denied the suggestion that she knew about the problem on or about Friday 20 August 1999 because she had heard of it from Robanks, stating "No, Sir, I've never heard of the problem from Robanks at all. It was in the general area." (Black, 787K-M). She said she had no recollection at all of somebody coming around from Robanks to replace a bag of Katunga chaff bought by her from Robanks on 3 August 1999, and "As far as I think, no they didn't." (Black, 787S-W). She added "I have never had anybody in to replace a bag of feed. As in actual fact, because nine times out of ten I go and pick my own feed up and I would, if I thought something was wrong with one, I would have taken it back myself" (Black, 788F-G). When shown an invoice from Robanks for one bag of lucerne chaff for $17.50 on 3 August 1999, she accepted that she had purchased it but added "Yes, but that could be Victorian chaff. It could be any other. He has numerous different chaffs and nobody ever came down and took a bag of chaff away." (Black, 788M). 85 Mr Rolfe's own evidence concerning contact with Ms Bronson, and hearing of rumours about chaff, was as follows: (a) He was not at the time living at the property at Ebenezer, as he was living with his parents and had his training establishment at the property at Ebenezer (Black, 166U-V). (That can, I consider, be accepted.) (b) He would be at the training establishment at Ebenezer for 14 hours a day plus going to the races and then returning, adding that races were on every day of the week and "It depended how many runners we had at any particular time." (Black, 166X-Y). (It is assumed that the 14 hours is a reference to the total time in the day, which would include travelling to and from racetracks.) (c) Before he saw Mr Frank van den Nieuwboer (this is "Mr Frank Robanks" as the plaintiff calls him - Black, 172X-173E) on Friday 20 August 1999 and before Monday 23 August 1999, he had heard "on the rumour mill around Ebenezer" that one horse had got sick from "a product supplied by Robanks" (Black, 172Q). He heard rumours of one horse in the locality only (which was not described as dead but just ill, he thought) supposedly through eating some product provided by Robanks (Black, 173S-X). He thought the kind of products supplied by Robanks might have been mentioned, but could not recall (Black, 173X-174C). (d) This rumour was a second-hand report which he heard from his landlady on Friday 20 August 1999 (Black, 174F-H). He was not aware it was actually chaff that was the problem, and did not know as at Friday 20 August 1999 what feed it was that was the problem; he had heard there was "some problem with some product" (Black, 184TX). 86 The cross-examination of Mr Rolfe as to what Robank maintained was any contact with Mr Rolfe as part of the recall process proceeded as follows: (a) Initially in cross-examination of Mr Rolfe, it was suggested to him that " Isn't it the case that Mr John van den Nieuwboer rang you on the Thursday before the 23rd (this would be Thursday 19 August 1999) to tell you that he thought there was a 'problem with the feed' ?" to which Mr Rolfe replied " Not that I can recall, no " (Black, 167DE). (b) Later in cross-examination Mr Rolfe was reminded that he had been asked "Yesterday about John van den Nieuwboer contacting you on the 19th of August and you told me that you didn't remember such a conversation" and then asked "Could it be that another person from Robanks may have contacted you on the 19th of August 1999 about these matters" to which Mr Rolfe replied "I can't recall any conversation from Robanks or their employees on that day" (Black, 179N-P). (c) Finally, in cross-examination, it was put to Mr Rolfe that "Somebody from Robanks told you on the 19th that there was a problem with the Katunga product ...?" to which Mr Rolfe replied "Untrue" (Black, 190D-G).

Conclusion 87 Recognising the advantage the trial judge had of assessing the witnesses, and noting that Ms Bronson was not accepted as a reliable witness by the trial judge, nonetheless, in what is a matter of inference, on the evidence so far recounted, including the more objective information of the note in Exhibit AR ("Pick-up-can't contact"), I would not accept that: (a) Ms Bronson was informed on the night in question (19 August 1999) of the possibility that the chaff was contaminated; (b) She must be taken to have passed on that information to Mr Rolfe; (c) Anyone from Robank spoke directly to Mr Rolfe with that warning including Andrew Palmer when visiting Mr Rolfe's premises (subject to consideration below of the sign in the shop and the alleged conversation with Mr Frank van den Nieuwboer on Friday 19 August 1999); 88 I do, however, accept that Mr Rolfe would more probably than not have known by Friday 20 August 1999 and before Monday 23 August 1999 "on the rumour mill around Ebenezer" that one horse had got sick from "a product supplied by Robanks".

Did Mr Frank van den Nieuwboer warn the plaintiff that there was a problem with the Katunga chaff? 89 As to (c) above, Mr Rolfe denied that Mr Frank van den Nieuwboer gave him any such warning. Mr Frank van den Nieuwboer himself denied and had no recollection of any conversation with Mr Rolfe on Friday 20 August 1999 (Black, 554R). He asserted that there was no conversation with Mr Rolfe until when Mr Rolfe came into the store and told him about a horse that had got sick and died. Indeed, Mr Frank van den Nieuwboer was certain there were no such conversations between him and Mr Rolfe about the chaff before the horse died (Black, 556H-T). Concededly, an honest witness following a general practice of giving such a warning to those who asked might forget that someone asked, such as, here, Mr Rolfe. But Mr Frank van den Nieuwboer's certainty that there was no such conversations at all with Mr Rolfe militates against his having forgotten that such a conversation had taken place. And if no conversation took place, he could not have given the warning, whatever his general practice; see further below, under "Grounds of appeal 8 and 9 - error in admitting evidence". 90 I would not accept, as the appellant submitted, that the trial judge erred in admitting this evidence, over objection, from Mr Frank van den Nieuwboer. That is, evidence of what he would have said in a conversation with Mr Rolfe in the store on 20 August 1999, when he denied and did not recollect any such conversation having taken place. However, I agree with the submission that the trial judge erred in concluding, on that evidence, that Mr Frank van den Nieuwboer warned Mr Rolfe in a conversation on Friday 20 August 1999 that "there was a problem with the Katunga chaff".

Could or did Mr Rolfe, when visiting the Robank store, see the sign on the counter warning customers of the recall? 91 I come now to the question of the sign on the counter which it was said Mr Rolfe should have seen on Friday 20 August 1999, warning of the recall. 92 The appellant correctly submitted that the evidence given concerning the existence of the one sign on the counter at Robank did not specify with any precision as to where it was located, how it was placed and where it faced, what size the sign or any printing on it was, how visible it was, what the sign said and whether it was in a location which Mr Rolfe as an account not cash customer, would ordinarily have gone to. Mr Rolfe being an account customer did not have to pay for his feed supplies at the counter. There was very little evidence as to the physical dimensions of Robank, where customers would ordinarily go in order to pick up feed supplies on their accounts and whether such customers might obscure the sign from Mr Rolfe on his likely path. I would agree with the above submission, based on a review of the evidence below. 93 Thus Mr Rolfe went to Robank on 20 August 1999 in order to make enquiries, and for this purpose spoke to Mr Frank van den Nieuwboer. I would accept the appellant's submission that this was at a location which does not suggest that he either had to go past a counter on which there was some form of sign, or that he should have been able to see the sign while he was there. 94 I agree with the appellant's submission that there appears to be some uncertainty as to what the sign said, as distinct from what the sign was intended to achieve. Thus if the sign said that purchasers should return bags which came from the "bad batch" of bags delivered to Robank by Katunga on 3 August 1999, the appellant fairly asks how would any customer know that any bags which he still possessed were covered by this description? I would accept that if no reason was given on the sign as to why such bags should be returned, it would be unlikely that the sign would necessarily command the attention and compliance by a customer, as there was no reason for urgency given on the notice. 95 Neither Mr Frank van den Nieuwboer nor Marianne Roughley gave any evidence concerning the placing of a sign on the counter. 96 Mr John van den Nieuwboer said that not all the customers were account customers, and customers were divided into categories of "cash sale" and "account" (Black, 658H-L). He said it was not possible to trace the names of the cash sale customers because it was cash over the counter and no names go with that. The preparation of the list was not restricted necessarily to one class of customers and not another. He gave as the reason why it was not restricted that "I organised, I instructed the employees to put a note on the counter if they bought chaff in a certain period, that they should return it and get a refund or replacement". The "certain period" which he specified on the notice, that went on the counter, was 1 August to 18 August 1999 to the best of his recollection (Black, 658M-V). 97 In re-examination Mr John van den Nieuwboer was asked more about the notice which had been put on the counter. He said that the counter was the point of sale "for the cash sale ones which we had no idea of" (Black, 704G-I). He agreed that the counter was the point of sale for the business for people who came in to buy over the counter (Black, 704J-L). 98 Mr Rolfe in his own evidence had said that Robank is "quite a big farm supply place with many employees" (Black, 169O). He said that when he went to Robank on Friday 20 August 1999 he went to Frank van den Nieuwboer's office. To get there you walked in the front door to the left hand side, there was no counter there, and Mr Frank van den Nieuwboer's office was directly off the first room that you walk into (which was where you complete your purchases) and to the left of that room (Black, 179E-K). 99 In cross-examination Mr Rolfe was asked whether, when he went on 20 August 1999 to Robank, he saw a sign put up in that area where the sales are conducted calling for the recall of the Katunga product. To this Mr Rolfe replied "There was no such sign". He added, "There was no sign that I visually saw, no" (Black, 185G-J). He was then asked whether, when he went in on Friday 20 August 1999 there was a sign "up there calling for the recall of the Katunga product?" Mr Rolfe replied, "That's incorrect." (Black, 185M). He was then asked whether, when he spoke to Frank van den Nieuwboer on Friday 20 August 1999, Frank told him that they had concerns about the Katunga product. Mr Rolfe replied, "If he had a sign up and he was concerned he would have told me about a concern with some contamination, but that wasn't the case" (Black, 185O-P).

Conclusion 100 I consider that there was no sufficient basis, in the evidence, for the trial judge to have drawn the inference he did that there was a prominent notice in the store (Red, 21Q). While the trial judge (Red, 21Q) referred to the sign, amongst other evidence, as one of "the pieces of acceptable evidence going to the knowledge of Mr Rolfe of the doubts concerning Katunga Chaff before he fed the horses on 23 August 1994 [sic]" he: (a) refers to doubts only, not more definite knowledge, (b) makes no finding, save by implication at best, that the sign was seen by Mr Rolfe or even as to what the sign said, No finding that the sign was prominent and actually informed Mr Rolfe of the contamination could properly be made on that evidence, whether by inference or otherwise, nor is that inference capable of being drawn from the other evidence relied on to show knowledge, including that below.

Other evidence relied on going to any knowledge on Mr Rolfe's part concerning contamination of Katunga chaff before he fed the horses on 23 August 1999. 101 I turn now to the other evidence relied upon as going to any knowledge on the part of Mr Rolfe concerning contamination of Katunga chaff before he fed the horses on 23 August 1999. 102 The remaining "acceptable evidence" relied on by the trial judge as going to knowledge was that "the plaintiff only bought NSW chaff after Wednesday 18 August 1999". Earlier in his judgment, the trial judge states that, "on 19 August the plaintiff purchased a bag of, significantly, NSW Lucerne chaff which further goes to the evidence that Katunga chaff was unavailable and that Mr Rolfe must have had some intelligence as to doubt surrounding the chaff since the purchase was that of his non-preferred NSW chaff." (Red, 20D-S). 103 This was too slender a basis either by itself or cumulative with the other "pieces of acceptable evidence" for the conclusion reached by the trial judge that "at the time of feeding the horses on the evening of 23 August 1999 the plaintiff knew that to feed the horse Katunga chaff was a very risky business" (Red, 21S-U). 104 The plain fact was that, as the appellant submits, if Katunga Lucerne chaff was not available for sale when Mr Rolfe attended Robank's store on Thursday 19 August 1999 he would simply have had to take what was available, namely NSW chaff. (Given that the new batch of Katunga Lucerne chaff was being "sold from sometime on" Thursday 19 August 1999, it must have been earlier that day, that is before the Katunga lucerne became available that Mr Rolfe would have bought the NSW chaff.) I agree that to find out that Katunga Lucerne chaff was unavailable merely established that fact. It did not justify the conclusion that he had concerns about the Victorian lucerne. Nor does it ground an inference that Mr Rolfe should have drawn that it was unavailable because of contamination from an earlier bad batch. It certainly would not of itself ground any inference that it was contaminated unless he knew that other horses had been killed in the area because of contamination from chaff affected by botulism and there is simply no evidence that he had that knowledge. With due respect to the trial judge, that could not ground an inference that Mr Rolfe "must have had some intelligence as to doubts surrounding the chaff". 105 And this is quite apart from whether "some intelligence as to doubts" suffices to obviate causation between the loss suffered and the contamination, or render resulting damage too remote. In my view neither conclusion can properly be reached. The situation is very different from that in Lexmead (Basingstoke) Ltd v Lewis (supra). There the House of Lords found that an accident from a design defect joining a Landrover to trailer was really the purchaser's fault. This was in continuing to use the defective coupling for a considerable period after the purchaser had noticed that the handle which operated the locking mechanism was broken. That reflects a degree of knowledge of the defect not found in any analogous sense in the present circumstances. Knowledge of mere rumour falling short of clear knowledge could not suffice to break the chain of causation. 106 To use the words of Mason CJ in March v E & M H Stramare Pty Ltd (1991) 171 CLR 506 at 522, the relevant question is whether the defendant's breach was so connected with the plaintiff's loss or damage that, "as a matter of ordinary common sense and experience it should be regarded as a cause of it." 107 Here, the sale of chaff contaminated with botulism might well generate rumours of Robank product being unsafe or even fatal. But mere rumour is not enough to sever the connection between breach, here of the warranty of merchantable quality, and the injuries suffered by Mr Rolfe's horses. In the words of the first limb of the rule in Hadley v Baxendale, damage of that kind for breach of contract is "such as may fairly and reasonably be considered … arising naturally … according to the usual course of things, from such breach of contract"; Hadley v Baxendale [1854] 9 Ex 341 at 354; 156 ER 145 at 151. 108 Likewise, in terms of the second limb of that rule, had the parties directed their minds to the probable result of the breach of warranty of merchantable quality, the injuries suffered by Mr Rolfe's horses was "such as may reasonably be supposed to have been in the contemplation of both parties, at the time they made the contract, as the probable result of the breach of it".

Grounds of appeal 8 and 9 - error in admitting evidence 109 I should elaborate now on one aspect of the evidence that was the subject of grounds of appeal 8 and 9. The appellant contends that the trial judge erred in admitting evidence, which was objected to, from Mr Frank van den Nieuwboer of what he would have said in a conversation with the plaintiff in the store on 20 August 1999, when he denied and did not recollect any such conversation having taken place; ground 8. Ground 9 is that the trial judge "erred on the facts in concluding that Mr Frank van den Nieuwboer warned the plaintiff in a conversation on Friday 20 August 1999 that 'there was a problem with the Katunga chaff'." 110 To place the matter in context, Mr Rolfe did give evidence that having heard on the rumour mill around Ebenezer from his landlady, Ms Bronson, that there was a rumour that a horse was ill, supposedly from eating some product provided by Robank, he went to Robank on 20 August 1999. His evidence was that he asked Frank (referred to by Mr Rolfe as "Frank Robanks") whether the rumours that he had heard were true (Black, 172-175, 182-186). According to Mr Rolfe, Frank assured him that there was no problem with the products that he was selling out of his farm supplies and that he was endeavouring "to find the truth out in the rumour" (Black, 174R). 111 Mr Rolfe denied that Frank told him in this conversation "we have been calling everyone to bring the product back" (Black, 174U). Mr Rolfe's evidence was that Frank clarified the rumour that Mr Rolfe had heard as being untrue (Black, 188V). 112 Frank van den Nieuwboer in examination-in-chief was asked about Mr Rolfe's evidence that he had a conversation with him on Friday 20 August 1999. Frank said he could not recall any conversation on that day (Black, 554R). After Frank stated that on Friday 20 August 1999 his belief was "that he had a problem" about the Katunga Lucerne Chaff, he was asked, "On the Friday would you have assured anybody (that the Katunga feed was alright)?" 113 This question was objected to by the plaintiff's Senior Counsel. Following argument, the trial judge allowed the question (Black, 555G). Frank was then asked "On the Friday when you had been back from Gunnedah and you had that belief that you've told us about concerning the Katunga chaff, would you have assured anybody that the Katunga chaff was okay or all right? (Black, 555H-J). Frank said, "I still don't understand the question." There was then further discussion in the absence of the witness. The trial judge ruled that the question could be asked in view of Mr Rolfe's positive assertion that he had a conversation with Frank on Friday 20 August 1999, and allowed the question (Black, 556B-G). Counsel for the defendant then asked Frank van den Nieuwboer the following questions: "Q. Mr van den Nieuwboer, Mr Rolfe has given evidence in these proceedings that he had a conversation with you on the Friday after you got back from Gunnedah? A. (Witness nodded)

Q. He has said that that conversation took place in your office at Robank Farm Supplies? A. I can't recall that, sir.

Q. He says that he told you something along these lines: that he'd heard some rumours around Ebenezer that some horses were getting sick from a product supplied by Robanks. He says that you said to him at that time there's nothing wrong with the feed that you were selling. Now, my question is, did that happen? A. No, sir.

Q. My question is, had somebody asked you on the Friday, after you got back from Gunnedah, whether there was a problem with one of the feeds you were supplying, what would you have said to them? A. I would have said there's a problem.

Q. What would you have said the problem would be? A. Most problem - the chaff.

Q. Would you have said which chaff? A. Yes, sir.

Q. Which chaff would you have identified? A. Katunga chaff."

(Black, 556H-T). 114 In cross-examination Frank said that a second batch of Katunga Lucerne Chaff came from the First Defendant on Tuesday 17 August 1999 or Wednesday 18 August 1999, that his store continued to use its best endeavours to sell Katunga Lucerne Chaff from that second batch (Black, 559-604) and that more likely than not he would have said to anyone who walked into his store, who wanted to buy Katunga Lucerne Chaff from that batch, "There is nothing wrong with that batch of Katunga chaff" (Black, 600C-H). He said that his employees were told to sell Katunga Lucerne Chaff from the second batch, but not from the first (bad) batch, because he assumed and was satisfied there were no problems with the Katunga Lucerne Chaff he was selling at the end of the week ending Friday 20 August 1999 (Black, 602G-603B). 115 The cross-examination of Frank then turned to the conversation which Mr Rolfe alleged took place between Frank and Mr Rolfe, when Mr Rolfe visited Robanks' store at Ebenezer on Friday 20 August 1999. Frank said that he had no conversation with Mr Rolfe on that day, that the only conversation he had with Mr Rolfe was when Mr Rolfe came in and told Frank about the horse that had got sick and died, and that he was "certain" that there were no conversations between he (Frank) and Mr Rolfe about the chaff before the horse died (Black, 603-604, especially 604G-V). However, Frank repeated that if Mr Rolfe spoke to him on Friday 20 August 1999 "I would have said there would be a problem with Katunga Chaff" (Black, 603G), but later conceded that on this Friday he would have been selling Katunga Chaff from the second batch which arrived on 17 or 18 August 1999 (Black, 604G). 116 The trial judge in his judgment concluded that Frank warned Mr Rolfe in a conversation on Friday 20 August 1999 that "There was a problem with the Katunga Chaff" (Red, 20V-21C).

Conclusion 117 I would not accept the appellant's submission that what Frank van den Nieuwboer asserted he would have said in the conversation he denied having, was inadmissible on that ground or on a ground that he was responding to a hypothetical question. It is open to a witness to say what he would have done, in relation to a conversation he could not recall, as illustrating, for example, his state of mind, but based upon what he could reasonably assume he would have done. But here he ultimately went from not recalling the conversation to denying with certainty that any conversation took place concerning the chaff. It is true that Mr Rolfe gave evidence to the contrary that a conversation did take place between them on Friday 20 August 1929. But the content of the conversation was that he raised the question of rumours about the Robank's chaff, receiving an assurance that there was nothing wrong with the feed. In those circumstances I would not draw the inference of the trial judge that the strongly denied conversation took place and included a warning about the chaff.

SUMMING UP 118 In Alexander v Cambridge Credit Corporation Ltd (1987) 9 NSWLR 310 at 361, McHugh JA (as he then was) cited the following passage from the High Court's decision in Mahoney v J Kruschich (Demolitions) Pty Ltd (1985) 156 CLR 522 at 528: "... A line marking the boundary of the damage for which a tortfeasor is liable in negligence may be drawn either because the relevant injury is not reasonably foreseeable or because the chain of causation is broken by a novus actus interveniens : M'Kew v Holland & Hannen & Cubbitts ... But it must be possible to draw such a line clearly before a liability for damage that would not have occurred but for the wrongful act or omission of a tortfeasor and that is reasonably foreseeable by him is treated as the result of a second tortfeasor's negligence alone: see Chapman v Hearse ... Whether such a line can and should be drawn is very much a matter of fact and degree ...." [McHugh JA's emphasis] 119 Here, Mr Rolfe may have been less than prudent in using chaff about which he had heard rumours concerning its possible danger. But to my mind that does not amount in fact and degree to the kind of knowledge which would, using the language in the cited passage, permit a clear line to be drawn as would break the chain of causation resulting from the breach of warranty of merchantable quality by Robank. 120 I should refer here to the observation of McHugh JA that "Mahoney makes it clear (at 529-530) that the exacerbation of an injury by a negligent medical treatment though reasonably foreseeable will constitute a novus actus interveniens if the treatment is inexcusably bad or improper" (at 361). Again, mere knowledge of doubts or rumours concerning the relevant product yet feeding it to his horses, would not of itself be "inexcusably bad or improper" so as to break the chain of causation. 121 I have earlier indicated where I do not, with respect, consider the trial judge's findings can stand. I would accept, on the evidence, that Mr Rolfe may well have had some doubts concerning Katunga Lucerne chaff before he fed his horses with it on 23 August 1999, given the rumours. But that cannot suffice. I recognise the advantages enjoyed by the trial judge "that cannot be fully recaptured by the appellate court" (Annikin v Sierra (2004) 79 ALJR 452 at 458 [37]). But this is, as the trial judge himself recognised, fundamentally a case of circumstantial evidence and therefore one which depends on inference. Here an appellate court is in as good a position as the trial judge to decide on the proper inference to be drawn from the facts (Warren v Coombes (1979) 142 CLR 531 at 551). Moreover, as in State Rail Authority of New South Wales v Earthline Constructions Pty Ltd (in liq) (1999) 73 ALJR 306 per Kirby J at 321 [63], the circumstance of impression of witnesses does not "preclude a Court of Appeal from concluding that, in light of other evidence, a primary judge had too fragile a base to support a finding …". This is insofar as such findings are to be taken as amounting to knowledge of contamination beyond mere doubt based on rumour. 122 Particularly telling is that, had Mr Rolfe known of the clear likelihood of danger to his horses from feeding the Katunga chaff, it would be entirely irrational for him knowingly to inflict damage on his own business in that way. That militates strongly against him doing so, or acting with a deliberate hiding of the eyes. His horses were too valuable to him to make that credible.

OVERALL CONCLUSION 123 The appellant must therefore succeed in his appeal, having established a breach of warranty of merchantable quality in relation to the Katunga chaff. There is no novus actus interveniens preventing the appellant from demonstrating the necessary causal connection between Robank's breach and the damage suffered by Mr Rolfe from the injury to his horses from being fed the contaminated chaff. Nor have any losses so suffered been shown to be too remote to qualify for compensation. 124 As to costs, I would not award the appellant all his costs, given that the appellant should have focussed on the only tenable argument he had, namely merchantable quality, and not strayed otherwise into contract and tort liability. 125 Accordingly I would propose orders as follows: (1) Judgment and orders in the court below in favour of the respondents/defendants to be set aside. (2) In lieu thereof judgment for the appellant/plaintiff against the second respondent/defendant on the issue of liability, damages to be assessed. (3) Matter to be remitted to the District Court of New South Wales for: (a) a new trial on the issue of damages; (b) a determination of the cross-claim brought by the cross-claimant, Robank Farm Supplies Pty Limited, against the cross-defendant, Katunga Lucerne Mill Pty Limited. (4) The respondents/defendants to pay 80% of the appellant's costs of the appeal and of the trial in the court below. 126 McCLELLAN AJA: I agree with Santow JA. **********

DISCLAIMER - Every effort has been made to comply with suppression orders or statutory provisions prohibiting publication that may apply to this judgment or decision. The onus remains on any person using material in the judgment or decision to ensure that the intended use of that material does not breach any such order or provision. Further enquiries may be directed to the Registry of the Court or Tribunal in which it was generated.