England and Wales Court of Appeal (Civil Division) Decisions

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE QUEEN'S BENCH DIVISION
The Hon Mr Justice Jack
[2004] EWCH 1101 (QB)

		Royal Courts of Justice Strand, London, WC2A 2LL
		12 May 2005

Lord Justice May:

Introduction

1. Christie's, the first defendants, held an important auction sale in London on 8^th December 1994. The sale was of the contents of Houghton in Norfolk, owned by the seventh Marquess of Cholmondeley. Lord Cholmondeley is the third defendant in these proceedings, but is not concerned with this appeal. The sale included 18^th and 19^th century French furniture and objets d'art. Lot 56 was described in Christie's catalogue as "A pair of Louis XV porphyry and gilt-bronze two handled vases". A photograph established that the vases had been owned by Lord Cholmondeley's grandmother in 1921, but there was no earlier established provenance. Ms Thomson, the claimant, bought the vases at the sale for £1,957,388 including the buyer's premium. The price exceeded Christie's estimate by some three times. In the final stages of the auction, there was one other bidder, believed to have been Ann Getty. After buying the vases, Ms Thomson put them into storage with Christie's. In October 1998, she heard a suggestion that the vases might not be Louis XV, but imitations made in the mid-19^th century or later. She communicated this fear to Christie's. They initiated a metallurgical investigation. This investigation, now discredited, concluded that the vases were made in the 19^th century.
2. In these proceedings, Ms Thomson claimed against Christie's and Lord Cholmondeley on the ground that the vases were made in the 19^th century and worth only about £25,000. She claimed that Christie's had owed her a duty of care which they had broken in not warning her of a risk that their judgment that the vases were made in the 18^th century might be questionable or wrong.
3. The action was tried by Jack J, who gave judgment on 19^th May 2004 – [2004] EWHC 1101(QB). His judgment is a masterpiece of evidentiary analysis, worthy to stand with any such exposition of works of art of this kind. It is an optimistic appellant who would try to have any of his factual judgments reversed.
4. The judge decided that the vases were made in the 18^th century. He so concluded on a balance of probabilities which he weighed as being 70%. He decided that Christie's had not been negligent in cataloguing the vases as Louis XV. But they had a special client relationship with Ms Thomson giving rise to a more exacting duty of care to her. The judge found Christie's to have been in breach of this duty. This entitled Ms Thomson to damages measured as the difference between what she paid for the vases, including the buyer's premium, and "their actual worth" (paragraph 200) plus interest. Their actual worth was their value at the time of purchase on the basis of the judge's finding as to the probable date (paragraph 201), the degree of probability being of the order of 70% (paragraph 204). If the parties were unable to agree a figure, there would have to be an enquiry as to damages.
5. Christie's appeal and Ms Thomson cross-appeals against the judge's decision.
6. Christie's main grounds of appeal are that:

i) the judge, having correctly found that they were not negligent in their cataloguing, was wrong and inconsistent to find them in breach of any duty owed to Ms Thomson; and

ii) the judge was wrong in his finding as to the measure of damages.

Ms Thomson, although forensically content with the judge's conclusion, contends that

iii) the judge was wrong not to find that the vases were made in the 19^th century; or

iv) the judge should at least have held that Christie's should have qualified their catalogue entry for the vases to describe them as "probably Louis XV" or by some such qualification.

7. Each party submits or acknowledges that the judge's finding that Christie's were in breach of duty to Ms Thomson sits uncomfortably with his finding that their catalogue entry was not negligent. In my view, the main eventual issue in this appeal can be expressed as whether the judge should have held that Christie's should have qualified their catalogue entry. This blurs some of the intermediate issues, but it sufficiently expresses the main point.
8. The judge's findings as to the date when the vases were made took account of extensive metallurgical and other scientific evidence put together for the purpose of these proceedings. Ms Thomson does not suggest that Christie's should have acquired information of this kind before the auction. It is not relevant to any question whether Christie's were in breach of duty. The judge regarded it as relevant to an assessment of damages. Christie's say he was wrong here. They say, and Lord Grabiner on behalf of Ms Thomson accepts, that any damages are to be assessed as at the date of the alleged breach of duty, that is December 1994. Christie's also say that any assessment of damages must leave out of account subsequently acquired information relevant to the dating of the vases which they were not obliged to acquire in 1994. Any damages, they say, must relate to the breach of duty established. They say that the value of the vases in 1994 was what Ms Thomson paid for them and that therefore she establishes no loss relating to the breach of duty which the judge found.
9. This court is not concerned with questions of limitation.
10. It is Christie's case that the vases were made in about 1765, that is in the last decade of the reign of Louis XV. It was, and still is, Ms Thomson's case that they were made after 1840 and probably after 1880. There was a revival of interest in earlier styles in gilt bronze during the Second Empire (1840-1870). Two technical features of the liners of the vases suggest dates not earlier than 1840 and 1880 respectively, and these are said to date the vases as a whole. Other metallurgical evidence also points strongly to a 19^th century date.
11. During the trial Christie's conceded that it owed Ms Thomson a duty to use reasonable skill and care in describing the vases as Louis XV and in making statements to her about their date and quality. The standard of care conceded was that of a reasonably competent international auction house specialising in the sale of fine art and antiques. As Mr Sumption explained to us, Christie's made this concession because, as discussed later in this judgment, its representative Mr Cooney gave Ms Thomson information about the date and quality of the vases, and because Christie's knew that she had no relevant expert knowledge and that she was not taking independent expert advice. Christie's did not concede that they owed such a duty generally to the bidding world at large. They did not rely against Ms Thomson on any of their Conditions of Business as excluding or precluding any duty of care to her. Ms Thomson contended for a rather wider duty than Christie's conceded.

The Houghton Vases

12. The judge gave a detailed description of the Houghton vases in paragraphs 17 to 19 of his judgment. A briefer description is as follows.
13. The bowl of each vase is made mainly of porphyry. The bowls are supported, and to a significant extent concealed, by elaborately fashioned gilded mounts made of brass. Gilded brass is sometimes inaccurately referred to as gilt bronze. It is also sometimes called ormolu. The mounts comprise two lions facing each other across the lip of the bowl. The lions' bodies and heads are in the nature of handles. The lions' claws hold in place gilt bronze drapery around each side of the vases from one lion to the other. The bowls are supported in a shallow basin of pointed gilded leaves. The leaves are supported by two gilded entwined snakes. Two small screws attach the leaves to the snakes. The snakes rest on a rectangular gilt base with recessed panels decorated with pouncing.
14. The bowls are not made entirely of porphyry. Their lower parts, concealed by the gilt mounts, are made of a cement like material, referred to as scagliola, in part painted to look like porphyry. The kind of porphyry of which the visible parts of the bowls are made came from Egypt and had not been quarried since Roman times. The porphyry bowls therefore had a previous existence in some form. The judge thought that they had probably been fashioned and drilled from pieces of Roman column (see paragraph 117 of the judgment). It would be easier to drill right through a piece of column and mend the concealed base of the bowl with cement than to make a complete bowl. Those advising Ms Thomson preferred the idea that the bowls had previously been mortars of the general kind to be seen in the Wallace Collection. This theory would mean that the maker of the vases decided to smash the bottoms of mortars already in existence. This seems to me to be quite unlikely and the proponents of this theory gave no satisfactory explanation for it. The judge discounted it, correctly in my view. I am not convinced, however, that this disagreement is centrally important. The debate arose because there was a measure of agreement, derived from the surface appearance of the exposed porphyry, that it was worked into its present condition before 1800.
15. Each bowl has a gilt bronze liner bolted in place through the base of the vase. The gilt rim of the liner lips over the internal edge of the porphyry. Each lion is held in place by a horizontal bolt passing through the liner. The head of each bolt is gilded.
16. The Houghton vases were made after a design by a French architect, apparently working in Parma, called Ennemond-Alexandre Petitot (1727-1801). Petitot, having trained in Paris, was appointed "architetto dello fabbrichi ducali" to the Duke of Parma in 1753. He died in Parma in 1801. The designs were stated to come from the "cabinet" of the first minister to the Duke of Parma. The designs were not necessarily or specifically for gilt ormolu objects. Some of them were made into garden ornaments for the first minister. The Petitot design is illustrated in an engraving by Bossi in a publication called "Suite des Vases" of 1764. There are some differences of detail between the Bossi engraving and the Houghton vases, but the main elements of the design are obviously the same.
17. The ducal court at Parma was a satellite of the French royal court of Louis XV. The duke and duchess were, so Mr Sumption tells us, first cousins once removed, both of or descended from the French royal family. The duchess was the eldest daughter of Louis XV. She visited Paris on occasions at great expense. Parma had been settled on the Bourbon Duke following the Treaty of Aix-la-Chapelle in 1748 which ended the war of the Austrian succession.
18. In paragraphs 23-25 of his judgment, the judge discussed the problem which may arise in dating objects such as the Houghton vases. In brief, there was from about the middle of the 19^th century a revivalist interest in 18^th century neo-classical style. During and after this period imitations (not fakes) of 18^th century objects were made. The imitations may be well made. But a 19^th century imitation is regarded as much less valuable than an 18^th century original.
19. It was part of Ms Thomson's case that there had been talk before the 1994 sale, which Christie's were or should have been aware of, that the vases were not 18^th century. The judge rejected this part of Ms Thomson's case, and there is no appeal from his finding.

The catalogue

20. Christie's had from June or July to September 1994 to prepare the catalogue for this substantial sale. Mr Charles Cator was in 1994 head of Christie's furniture and works of art department, which consisted of seven people. Orlando Rock, then a relatively junior member of Mr Cator's department, was mainly responsible for the detail of the catalogue entry and the research which preceded it. The terms of the catalogue were, however, a corporate decision. Mr Rock gave evidence for Christie's about the catalogue entry. He said that nobody in Christie's who examined the vases considered that they were other than 18th century.
21. There is a pair of vases, very similar to the Houghton vases, in the Getty museum in California. There was a significant amount of evidence about the Getty vases. There is a third similar pair of vases in Greece, about which there was less evidence. The Getty bowls have red marble bases in place of the Houghton cement. The porphyry of the Greek vases is also cut but their bases are of porphyry. Despite these and other differences, there is an underlying assumption, sufficiently based on evidence, that each of these three pairs of vases are probably of the same provenance and that they were probably originally in common ownership. This is a probable inference, not only from their common design, but from evidence about the gilding of the liners, which I discuss later in this judgment. These inferences may be stronger for the Houghton and Getty vases than the Greek vases. There was also some evidence of a fourth pair of similar vases sold in France in 1991 and at a further sale in Paris in 1995. These Paris vases are palpably 19^th century (see paragraph 86 of the judgment).
22. The only information as to the provenance of the Houghton vases available to Christie's was that they were shown in the 1921 photograph and that Lord Cholmondeley had been told by his grandmother that she had bought them in France shortly after the end of the First World War. Christie's had information about the Getty vases, which the museum had acquired from Sir Everard Radcliffe of Yorkshire. The museum had described them in its 1994 journal as "a pair of vases, Italian or French, circa 1765-1770". The Getty entry described its vases as apparently made after a design by Petitot, whose career is described. It was said to be likely that the porphyry bowls were cut in Italy and the quality of the mounts suggested Italian manufacture. The Getty journal entry was written by Gillian Wilson, the museum's Curator of Decorative Arts. She has a very high reputation in her field.
23. By 1993, Gillian Wilson had modified her opinion so that the museum's summary catalogue described the Getty vases as "Paris (possibly Italian), circa 1765-1770. After an engraving by Benigo Bossi of a design by … Petitot". A 1997 Getty publication has a further modified description of "(?) French (Paris), circa 1765-1770" with a suggestion that the porphyry bodies of the vases were probably made in Italy but that the mounts were likely to have been made in Paris. By 2001, the museum's handbook had removed the question mark and stated that their vases were "French (Paris)". As the judge pointed out (paragraph 89), the dating by the Getty museum of their vases had remained as 1765-1770 throughout. They had moved from attributing them as of Italian manufacture to French.
24. In September 1994, Mr Hardy of Christie's contacted Gillian Wilson, who provided him with a spare print of the Bossi engraving and further information on Petitot. They exchanged views. She stated her then view that she had always assumed that the Getty vases were made in Italy. Christie's had a copy of the Suite des Vases in their library which both Mr Hardy and Mr Rock had looked at.
25. The first draft of Christie's catalogue entry, prepared by Mr Cator, had the vases as Louis XVI (who came to the French throne in 1774), "possibly Italian" and "based on" a design of Petitot. A version of the draft catalogue entry dated 15^th August 1994, prepared by Mr Rock, still had them as Louis XVI, but the reference to "possibly Italian" was deleted. An abbreviated catalogue of October 1994, prepared for an exhibition for the sale in New York, had them as Louis XV, "possibly Italian" and "based on" a design by Petitiot. The sale catalogue itself described them as:

"A PAIR OF LOUIS XV PORPHYRY AND GILT-BRONZE TWO-HANDLED VASES designed by Ennemond-Alexandre Petitot"

The provenance was stated as "acquired by [Lord Cholmondeley's grandmother] before 1921". There followed quite a long essay about Petitot and Philip, Duke of Parma. This is quoted in full in paragraph 39 of the judgment. It stated that the vases were "designed by" Petitot either for his patron Philip, Duke of Parma, who was Louis XV's son in law, or for one of his leading courtiers. There is a brief biography of Petitot. He was said to be responsible for introducing neo-classicism to Parma. There is a reference to the Suite des Vases of 1764 and its Bossi engravings, to the vases in the Getty museum and to a third pair in a private collection in Greece.

26. It is Christie's case and was Mr Rock's evidence that Christie's were quite confident in 1994 before the auction sale that the Houghton vases were 18^th century, properly described as Louis XV and to be dated between 1765 and 1770. The basis for this opinion was visual inspection and research.
27. Christie's also produced a "Condition Report" which the catalogue said they would provide on request. Ms Thomson did not ask to see it, but it was seen by Mr Cooney who advised Ms Thomson on Christie's behalf about the vases. The report stated that the antique porphyry vases had been reworked by Petitot and the bases filled with a scagliola section. It said that the mounts were extremely finely cast and had an excellent colour. The condition was extremely good and the quality exceptional.

Ms Thomson and Mr Cooney

28. Ms Thomson is a rich Canadian. She is a keen private collector with no special training or expert knowledge. She had made a number of auction purchases since 1990, but never before to the value she bid for the Houghton vases.
29. Christie's had in 1994 a New York Client Advisory Group providing "Special Client Services". The role of the special client advisor was to provide the client with as much information as possible so that the client could judge whether to bid or not. The advisors would pass on information about the merits of a piece from Christie's specialists. Ms Thomson became a client of Miss Kim Solow, now Mrs Kilgore. She said that Ms Thomson insisted on "museum quality", an appropriate description no doubt for the Getty urns, when buying a serious piece, and that she was wholly reliant on Christie's. Ms Thomson regarded Christie's as expert advisors. Christie's knew that she never took her own independent expert advice. At some date in 1994, well before the Houghton sale, Ms Thomson was given the status with Christie's of a "special client".
30. In February 1994, Mr Patrick Cooney joined Christie's Special Client Services team. One of his tasks was to help VIP clients with buying at auction. Miss Solow introduced him to Ms Thomson as one of two advisors. Ms Thomson received the catalogue for the Houghton sale. She was shown some of the lots in storage, including the vases. She attended a private view on 1^st December 1994 and a lunch in Christie's boardroom the following day. She was impressed with the beauty of the vases and saw them again at an evening view. Mr Cooney had come to London. She met him at the view and spent most of the evening talking with him. She next saw the vases on the day of the auction at a private view arranged by Mr Cooney. He had discussed the vases with Mr Cator, who told him that they were wonderful objects with no problems. He expected the estimate (£400,000-£600,000) to be exceeded, perhaps even doubled. At a late afternoon view, Ms Thomson said she intended to bid for the vases and asked what they might go for. Mr Cooney was unable to say, except that they would go for well above the estimate. It was arranged that he would bid for her with instructions by telephone.
31. Paragraph 8 of Ms Thomson's re-amended points of claim pleaded that:

"At the viewing Mr Cooney showed Ms Thomson the Urns and praised their high quality. He drew Ms Thomson's attention to the quality of the gilding on the lions and serpents which he described as marvellous. He said that the chasing of the Urns was equally marvellous and the patina was equally fine. He explained that the quality of the work reflected the superb craftsmanship of the 18^th century which he described to Ms Thomson as a "lost art". Mr Cooney also told Ms Thomson that a very similar pair of Urns were in the Getty museum in California and that the Urns were of a similar museum quality."

The gist of this was admitted in Christie's defence. The judge said in paragraph 50 of his judgment:

"The gist was that Mr Cooney spoke very highly of the quality of the urns as examples of 18^th century craftsmanship, and did not suggest that there was anything other than certainty that they were what they were described to be. He also said in evidence that he thought that he had said that the urns were by far the best item in the sale that she had looked at, and that it was better to buy one big thing than a number of smaller items. … He said he tried to give her all the information he could. For her part Ms Thomson said … that she would not have made the purchase without expert advice because she did not have the understanding to make an informed decision, and that she looked to Mr Cooney together with what was in the catalogue for that."

32. At the auction, Ms Thomson entered the bidding through Mr Cooney at about £1m. From about then all except one other bidder dropped out. The judge accepted Ms Thomson's evidence that she made the final winning bid with reluctance, being encouraged to do so by Mr Cooney. The judge said that little turned on this except that it showed Ms Thomson's reliance on Mr Cooney.
33. As Christie's now, I think, appreciate, the role which they assigned to Mr Cooney and his undertaking of it in relation to Ms Thomson created an inappropriate situation of potential, and perhaps actual, conflict of interest. Christie's primary function at the Houghton sale was to act as agent for Lord Cholmondeley. Their duty to him was to maximise the prices at which his various treasures were sold, within the limits of the requirements that they should act honestly and in accordance with the requirements of good practice in the preparation of the catalogue. Christie's received a commission based upon the selling price, so that their interest coincided with that of their client. It is part of an auctioneer's function to provide prospective purchasers with opportunities to see and examine the items for sale and to provide any information requested over and above that supplied in the catalogue and condition report. If the auctioneer wishes to attract wealthy people as potential clients by providing them with a special contact within the firm, so be it.
34. However, if the auctioneer in his relationship with potential purchasers goes beyond facilitation, he does so at his peril. His role is to sell the object for the best price. His duty to his client (the seller) and his own interest preclude true impartiality. The existence of such a conflict of interest predicates a duty of care owed by Christie's to Ms Thomson, as to which there is now little dispute. The characteristics of the duty of care are not diluted because Christie's had other interests and owed other duties. The potential for breach of duty to Ms Thomson is, if anything, enhanced.

Duty of care

35. There is no dispute but that Christie's assumed a responsibility to Ms Thomson giving rise to a duty of care. There is some dispute as to the extent of the duty. Mr Sumption submits that the duty had no content beyond that subsumed in a reasonably careful catalogue entry by an international auction house of Christie's standing. Mr Cooney's advice to Ms Thomson only reflected what Christie's furniture department had said to him and in the catalogue. He did not assume any wider responsibility as to the general commercial wisdom of the bidding. If Christie's had, or ought to have had, any real doubt as to the authenticity, and in particular the dating, of the vases as described in the catalogue, it was their general obligation to qualify the catalogue description appropriately. If they did not have such doubt, there was by definition no relevant qualification which they were obliged to express to Ms Thomson.
36. Lord Grabiner submits that Christie's owned Ms Thomson a Hedley Byrne duty of care which on the facts extended to advising her that, notwithstanding Christie's own confidence in their dating of the vases, there were risks inherent in the nature of their opinion and of the purchase that their opinion might be, or might possibly be regarded as being, wrong. In the vernacular, Mr Cooney's advice was unjustifiably confident and encouraging. If his advice had been, as it should have been, more restrained, Ms Thomson would not have bought the vases. As it is, she bought vases which are at best no more than "probably Louis XV", as Christie's should have described them at least to Ms Thomson.
37. I record, first, that Christie's made no concession as to any duty they might owe, absent special facts, to the bidding public generally by reason of a catalogue description; and second, that the duty contended for by Ms Thomson is a single duty owed to her as a special client advised by Mr Cooney. A component of that duty arises from the unqualified terms of the catalogue. But Lord Grabiner contends for no separate duty arising from the terms of the catalogue alone.
38. Although dancing angels might be able to find some difference in substance between the duty conceded or contended for by each of the parties, I am unable to see such a difference. The law of negligence is often conveniently analysed in separate compartments of duty, breach of duty, causation and damage. But, as I shall elaborate later in this judgment, the single question in each case is whether the defendant assumed responsibility to the claimant to guard against the kind of loss which the claimant claims to have suffered. The four components have to be taken together. In the present case, there either were, or were not, in 1994 facts which should have led Christie's to be cautious about their dating of the vases. If there were such facts, they should (as they accept) have expressed them in the catalogue but certainly to Ms Thomson. If there were not, there no basis on which Mr Cooney should have qualified his advice.

Events after the sale

39. Ms Thomson was told in October 1998 of doubts about the 18^th century dating of her vases. She communicated this to Christie's. There followed much correspondence and other activity, much of which does no credit to Christie's. The details are in paragraphs 57 to 76 of the judgment.
40. Christie's, with Ms Thomson's permission, had the brass of the mounts of the Houghton vases tested. This resulted in a report of 22^nd March 1999, which concluded very clearly that the brass mounts were most probably made in the later 19^th century and that therefore the vases were likely to be a much later version of those illustrated by Bossi in 1764. It is now agreed that the method of data analysis used by this report was unsound and that its conclusions cannot be relied on. Christie's, however, believed at the time that the report was valid, but withheld its conclusion from Ms Thomson for more than a year.

The Proceedings

41. The first main issue in the proceedings was whether the Houghton vases were made in the 18^th or 19^th century. Eight experts gave evidence. The judge introduced them and their qualifications in paragraphs 94-101 of his judgment. Mr Paterson, called on behalf of Ms Thomson, was a generalist with no experience of cataloguing 18^th century French gilt bronze. Christie's, with some justification, decried his qualifications. The judge said that he had overstated aspects of his curriculum vitae. Mr Bourne, called on behalf of Christie's was a very well qualified expert in fine art who had handled a number of major sales when he was with Sotheby's. Mr Yates, called on behalf of Ms Thomson, was a generally well qualified expert on stone. Professor del Bufalo, called on behalf of Christie's, was an expert on stone, especially coloured marbles of ancient Rome. He was familiar with the Roman quarries in Egypt which had been until recently the only source of red porphyry. Mr Kingdon, called on behalf of Ms Thomson, was a foundryman expert in casting and the working of bronze. M. Moreira, called on behalf of Christie's, was an expert in casting and the working of bronze. He was in charge of restoration of bronzes at the Louvre and Versailles. Dr Northover was a well qualified metallurgical expert called on behalf of Ms Thomson. He is an expert in the analysis of the constituents of alloys. He had carried out various analyses of metal samples taken from the Houghton vases. Dr Bennett, called on behalf of Christie's, is a specialist in the history of metallurgy. Her main role was to comment on Dr Northover's conclusions.
42. Speaking generally, the judge was more impressed with the evidence of Mr Bourne, Professor del Bufalo and M. Moreira than with that of Mr Paterson, Mr Yates and Mr Kingdon. He did not accept some of the conclusions drawn by Dr Northover from the raw data of his analyses. He correctly said (paragraph 102) that his own task was to see what conclusions as to date could be drawn from the evidence as a whole.
43. Since research and visual examination, but not metallurgy, had been the basis of Christie's description of the vases, the judge considered these first. In summary his conclusions were as follows.
44. As to the design of the vases, the use of Petitot's design did not point conclusively to an 18^th century or 19^th century date, but general experience suggested that the first use of a design was more likely to be soon after the design became available than many decades later (paragraph 103). The design was original, bold and unusual. The use of liners was unusual. In paragraphs 105 to 107, the judge referred to and quoted from a number of publications, which research at about the time of the trial had turned up. These are interesting but, I think, peripheral to the appeal.
45. The judge rejected Mr Paterson's reliance on the preponderance of gilt over porphyry as more typical of the 19^th century. With these vases, it was something dictated by the design which was undoubtedly 18^th century (paragraph 108). The judge was also unpersuaded that such realism and accuracy as the lions have was inconsistent with 18^th century manufacture (paragraph 109).
46. No conclusion as to date could be drawn from the absence of maker's marks (paragraph 110). Absence of provenance before 1921 was a reason for caution. It was surprising that there was no secure earlier record of any of the 6 vases in either the 18^th century or the 19^th century, if they were indeed 18^th century. But there was evidence of securely dated 18^th century pieces which had no real provenance – it applied to 117 of the 214 French decorative objects in the Getty Summary Catalogue of 1993.
47. I have already referred to the judge's reasonably confident conclusion about how the porphyry bowls were fashioned. The finish of the porphyry is not good. There are visible turning marks and pitting. The polish is low except on the horizontal top edges. In paragraphs 118 to 121 the judge considered the respective opinions of Professor del Bufalo, Mr Paterson and Mr Yates as to what this suggested about the date of the vases. He concluded that an examination of the porphyry revealed nothing which was inconsistent with an 18^th century date or which militated strongly against such a date. He considered that an 18^th century neo-classical date was substantially more likely than a 19^th century date. His main reason was that the finish was more consistent with the earlier date. As the 19^th century progressed, workshop equipment improved and machine driven lathes became common. This enabled smooth surfaces and high polishes to be achieved more easily. I note that Professor del Bufalo and Mr Paterson had agreed at an experts' meeting that the porphyry of the vases had been fashioned before 1800. Mr Yates had not initially agreed this. The judge noted that there was really no dispute but that the porphyry must have been worked at the same time as the mounts were made, because each was made for the other. So the date of the working of the porphyry suggested the date of the whole.
48. Considering all the gilt bronze except the liners (that is, the entirety of the mounts) the judge discussed a variety of features in paragraphs 123 to 131, extending to some 4½ pages of closely typed analysis. The features were the method of casting; the number of pieces in which the mounts were cast; the thickness of the casting; casting accidents; methods of joining; fixing points; chasing and pouncing; the gilding; and holes, bolts and screws. The method or methods of casting did not help the dating. The number of pieces in which the mounts were cast favoured an 18^th century date, but should not carry much weight. The thickness of the casting did not assist. Casting accidents or faults in the mounts were an indication favouring an 18^th century rather than a 19^th century date. There were two possible methods of joining the cast pieces of the mounts – forge brazing and torch brazing. Forge brazing was the only method available in the 18^th century. Torch brazing was not invented until 1882. The judge accepted M. Moreira's evidence (in preference to Mr Kingdon's) that evidence of mitrailles (granules or clippings of brass) supported the use of forge brazing rather than torch brazing. This was consistent with 18^th century manufacture, but equally consistent with 19^th century because torch brazing was invented so late. The addition of fixing points by brazing did not help with the dating. As to the chasing of the lions, the drapery, the leaves and the serpents and the pouncing of the bases, the judge accepted M. Moreira's opinion that the precision of the tools used and the manner in which they were used were a strong indication of an 18^th century date – see paragraph 129 for the detail. But even here there could be no certainty, since old tools are not always thrown away and could have been used in 1850 or 1860 in an 18^th century manner.
49. The mounts are gilded by the mercury process, which became illegal in France in the second half of the 19^th century because it was poisonous. Electrolytic gilding (which I consider below in relation to the liners) was invented and patented in 1840 at about the same time as mercury gilding became illegal. M. Moreira considered that the mercury gilding indicated the 18^th century. The judge considered that it made an 18^th century date more likely. It did not rule out a date after 1850, although it would by then have been illicit.
50. Features of the holes, bolts and screws were consistent with 18^th century manufacture, but did not rule out a date after 1850 (paragraph 131).
51. The structure and other features of the liners loomed large before the judge and in this court. It was agreed that the liners were made at the same time as the rest, though not necessarily with all their present features. The liners are an integral part of the whole because the main bolts are attached to or pass through other parts.
52. The judge described the liners in detail in paragraph 133 and considered details of their manufacture in subsequent paragraphs. They were in three parts: the walls, the bases and the rims, each of which had to be joined. The features which could be indicative of their dating included:

(a) the circular plate attached to the underside of the bases to receive the bolt going through the base of the mount. The plate and its collar were attached by brazing. The evidence supported forge brazing rather than torch brazing, consistent with 18^th century manufacture and with 19^th century manufacture (paragraph 136).

(b) M. Moreira and Mr Bourne considered that the manner in which the liners are constructed (described in paragraphs 137-139) confirmed an 18^th century date. If they had been made in the second half of the 19^th century they would have been shaped by compression, perhaps using a steam hammer. The liners of the admittedly 19^th century Paris vases appeared to have been made in this way. The contrast was striking, but not determinative.

(c) the evidence was inconclusive about whether the bronze sheet from which the liners were made was cast or rolled, but probability favoured casting. Rolled brass was only available in the 19^th century. The judge reviewed the differences of opinion here in paragraphs 141 and 142.

53. One of the liners has a visible repair. A small rectangular piece of metal was soldered into the wall near the rim. It is visible beneath the gilding on the outside and is not quite flush and clearly visible on the inside when the liner is removed from the bowl. It was agreed that the soldering was done with a blow torch and therefore after 1882. There are other imperfections associated with the repair (paragraph 143).
54. The judge said that it was hard to come up with a rational explanation for the repair. He reviewed various unpersuasive theories in paragraph 144, concluding that the true reason probably had not yet been imagined. It was Ms Thomson's case that the liner had been damaged in the course of its original manufacture. If it had been, and since on the exposed face the repair is covered by electrolytic gilding (see below), that would give a 19^th century date for the vases. But the judge considered that the strong probability was that the repair was carried out on an occasion subsequent to manufacture. He gave his reasons, which are persuasive, in paragraph 145. In short, if a repair had been needed at the time of manufacture, it would have been more skilfully done. I note that a clear alternative would have been simply to make a new undamaged liner. The judge found the evidence to which he referred far more cogent than the metallurgical evidence which he came to consider later in his judgment. If he was right, the electrolytic gilding of the liners was no guide to the dating of the liners and the vases.
55. It is now accepted that the present gilding of the liners was done electrolytically and that this must have been done after 1840. The judge accepted that this electrolytic gilding was done after the repair. But it was common for items to be regilded. If the vases are 18^th century, they must either have been regilded after 1840 or originally left ungilded. The latter possibility arose only during the trial, but the judge did not think this detracted much from it as a possibility, because it was Christie's case (and may originally have been accepted on behalf of Ms Thomson – see paragraph 147) that there were traces of earlier mercury gilding visible on the liners. If this were so, other possibilities did not then need to be considered.
56. The visual evidence about the gilding of the liners was consistent with their being 18^th century largely because of the possibility of regilding (paragraph 149). There was also the powerful argument in favour of the 18^th century in that the mounts are mercury gilded and the use of two different gilding processes (one of them illegal) at the same time was improbable. Having considered all the evidence, including metallurgical evidence, the judge concluded that there was a reasonably strong probability that the liners were never gilded by the mercury process in other words, that they were originally ungilded.
57. It is significant that Christie's did not notice before the 1994 sale that the liners are electrolytically gilded. Neither did M. Moreira on his first inspection. On the other hand, Mr Bourne said that it was obvious. It is accepted as probable that the liners of both the Getty and the Greek urns are also electrolytically gilded.

Metallurgical evidence

58. Those advising Ms Thomson spent much time and energy, both before the judge and in this court, trying to establish by metallurgical evidence not available to Christie's at the time of the sale that the vases were made in the second half of the 19^th century. The principal expert witness here for Ms Thomson was Dr Northover. The counterpart expert for Christie's was Dr Bennett. The judge considered this evidence with the greatest care. Taking this evidence with the evidence as a whole, he concluded that the vases were 18^th century. He assessed his personal confidence in that conclusion on the balance of probability as 70% and regarded this assessment as relevant to an assessment of damages to which he decided on other grounds Ms Thomson was entitled. By her cross-appeal, Ms Thomson seeks to persuade us that the judge's conclusion on all the evidence that the vases were probably 18^th century was wrong.
59. Since, as I shall explain, I consider that the metallurgical evidence and any conclusion to be drawn from it are irrelevant to any assessment of damages, I shall take this part of the appeal quite shortly. I should, however, make plain that, in my view, there is no proper basis for appeal against the judge's conclusion that Dr Northover's evidence did not prove that the vases must have been made in the 19^th century.
60. This was Dr Northover's and Dr Bennett's first venture into the field of 18^th century and 19^th century decorative brass. There were no established collections of metallurgical data with which comparison might be made. Metallurgical testing and the production of reports continued during the trial.
61. The judge said, correctly, that he had to look at Dr Northover's evidence as a whole in conjunction with all other evidence relevant to date, something which was beyond Dr Northover's remit. There were passages in Dr Northover's report where he had been a good deal more confident in stating his conclusions that the logical consideration of the available data could justify. Dr Northover had said that, if there was little available comparative data, you had to proceed on the basis of what there was. The judge considered that the right conclusion might be that you could not proceed in this way at all (paragraph 155).
62. Dr Northover's evidence covered the composition of the brass in the mounts and the liners and of the solder in the liners. He considered whether the liners were made of cast sheet or rolled sheet. He considered the brazing methods used. His view was that each pointed away from an 18^th century date, and that taken together they were conclusive (paragraphs 164-166). Having considered each of these separately (as I shall, briefly, below), the judge did not find Dr Northover's deductions persuasive, except as to the absence of mercury gilding on the liners (paragraph 177). His evidence did, however, have more weight taken as a whole than in its separate aspects.
63. As to the composition of the brass used for the mounts, Dr Northover had what he called a database of 22 samples taken (in effect) from 7 objects dated between 1700 and 1771. Put shortly, the judge considered (paragraph 159) this to be far too small to sustain any significant comparative conclusion. He was, I think, entitled so to conclude.
64. The level of impurities in the brass of the mounts was agreed to be low in comparison with the items in Dr Northover's database. The judge also made comparison with figures for impurities in 11 ormolu items from the Matthew Boulton foundry in England dated between 1780 and 1800. Dr Northover considered that the level of impurities for the Houghton vases suggested a 19^th century manufacture when refining techniques were better. The judge did not regard this as established (paragraph 163). Nor was he helped by comparison with the impurities in 19 French guns cast between 1750 and 1810 nor an 1857 boiler tube from a British steam engine (paragraph 161); nor various other objects referred to in paragraph 162. These conclusions are scarcely surprising. The judge said that the references underline the paucity of information on which to base any conclusion. Dr Northover's evidence was more confident than views which he had expressed in earlier reports. The judge did not find justification for the greater confidence.
65. The judge considered evidence about zinc levels and the cementation or smelter processes. He was not impressed, for the detailed and convincing reasons given in paragraph 167, with Dr Northover's thesis that the relatively high mean zinc levels in the samples taken from the Houghton vases pointed to the 19^th century; that they indicated that the brass was made by the smelter process, which was scarcely available in France in the 18^th century; and by comparison with the unhelpful database.
66. As to the composition of the brass used for the liners, the judge considered that no conclusion could be drawn from the relatively high zinc content of the alloy used for the liners and the low level of impurities (paragraph 169). The analyses showed that a distinctly different alloy was used for the liners from that used for the mounts with substantially more zinc. There were various inconclusive suggested explanations for this. The alloy used for the repair to one of the liners was very similar to, but not the same as, that used for the walls and the base (paragraph 170).
67. Dr Northover analysed the solders used for making the liners and for effecting the repair, concluding that they were exactly the same. The judge considered that the wide range of four of the five samples weakened this thesis, and only justified a conclusion that they were similar. No safe evidence was adduced to enable these solders to be attributed to either the 18^th or the 19^th centuries exclusively (paragraph 171). Similarly, no conclusion could be drawn from metallographic evidence relating to the brazing of the mounts (paragraph 172).
68. As to the sheets used for the liners, the judge briefly considered evidence from Dr Northover's examination of a single sample of the base and wall sections of the liners suggesting that the sheets were rolled, not cast (paragraph 173).
69. As to the gilding of the liners, it is accepted that the present gilding of the liners, including the gilding which covers the repair, was done electrolytically. The judge concluded for reasons which are persuasive (paragraphs 174-176) that the reasonably strong probability is that the liners were never gilded using the mercury technique, but only electrolytically. This precludes regilding electrolytically in the 19^th century liners which were originally gilded in the 18^th century. Accordingly, if the liners were made in the 18^th century, they were then originally ungilded.

The judge's conclusion as to date

70. The judge did not consider that the issue as to the date when the vases were made was to be determined on the balance of probability. The value of the vases would be so markedly different if they were certainly made in about 1765 (rather than in the 19^th century) than if they were probably made in about 1765, the degree of probability being, say, 55%. The judge regarded his task as the same as Christie's had been in 1994, that is to make a judgment as to the dating of the vases and to calibrate the extent to which that judgment was secure. As will shortly appear, I think that there is room for confusion here. It is necessary to be quite clear as to the purpose for which the question is asked. A conclusion as to date may be needed for purely factual purposes, as, for instance, if the true date of the vases needs to be determined in order to assess damages. For that purpose, a determinative balance of probabilities decision may be required. On the other hand, the different question may be whether Christie's were in breach of duty in their catalogue description of the vases or in the advice which they gave Ms Thomson. For that purpose, a degree of confidence expressed on a scale of 0 to 100 may be necessary. However, it does not follow that the degree of judicial confidence in reaching a balance of probabilities conclusion for the first of these purposes is also determinative of the measure of confidence which Christie's should have had for the second of these purposes.
71. The judge expressed his conclusion as to date in paragraphs 180-182 as follows:

"The following matters speak strongly in favour of an 18^th century date:

a) I accept that the chasing and pouncing have the signature of circa 1765 rather than of revivalist work. I am satisfied that when he saw the urns M. Moreira had no doubt about that, and I respect his skill and experience. He is a man who has spent his life working with such objects. M. Moreira was supported by Mr Bourne, who impressed me with his knowledge and as a careful expert witness.

b) The use of mercury gilding on all parts save the liners. If mercury gilding was used to create an 18^th century effect in the 19^th century, it would surely have been used on the whole ensemble rather than electro-plating the liners and so revealing the actual date. This suggests that the electro-plate was not original. In view of the tests on the liners to establish whether or not there was an original mercury gilding, the probability is that the liners were left ungilded following manufacture. I do not find that surprising. The surface area of the liners is comparatively large and only a small part is visible unless one looks down into them. It seems to me that they may well have been left as polished bronze, and perhaps lacquered, that is varnished, to preserve the shine, as was often done. … [considerations relating to the gilding of the liners of the Getty and Greek vases do not] … over-ride my conclusion that the gilding of the liners actually supports an earlier date for the urns.

c) If the electro-plate gilt is original, it is extraordinary that the repair inserted into the liner was done so badly and visible solder was left at a number of points. All this shows that the repair was subsequent of the manufacture. The strong probability is that the repairs consisting of the insertion of a new piece and the addition of patches of solder to smooth the surface were done at the same time as the electro-plating, perhaps by a workshop whose speciality was electro-plating rather than bronze working.

d) I have concluded that the porphyry is substantially more likely to have been worked in the 18^th century than later. This also dates the whole.

The main arguments for a date sometime after the mid-nineteenth century are those relating to electro-gilding, blow torch brazing and the metallurgical analyses. I do not think that any of these points can be brushed aside. They are matters which raise uncertainty. I have taken the various metallurgical aspects individually. There is nothing in them that supports an 18^th century date; they are all points against it. Taken together they carry more weight than individually, rather on the basis that a horse is more likely to fall if it has to clear six jumps than one.

I conclude that in particular the factors which I have just listed in favour of an 18^th century date outweigh those which are against it. I cannot be certain that the Houghton urns were made around 1760 to 1765, but I think that it is likely. The evidence establishes the position somewhere between certainty and more likely than not. If a figure must be placed on it, I would put it in the region of 70 per cent. I am of this view after having reviewed and reconsidered all the evidence and submissions which I have heard."

72. As to where the Houghton vases were made, the judge concluded that they may well have been made in Parma, but that it was also possible they were made in Paris (paragraph 183).

Breach of duty

73. The judge approached the issue whether Christie's were in breach of duty in two stages: first, whether their catalogue description of the vases was negligent; second, if it was not, whether Christie's particular duty to Ms Thomson required them to advise her that their catalogue description might be wrong. Both parties, for rather different reasons, accept (as did the judge – see paragraph 187) that there is really only one issue. Lord Grabiner submitted that Christie's had assumed what he called a Hedley Byrne responsibility to Ms Thomson and that, in relation to Ms Thomson, a separate cataloguing duty did not arise. Mr Sumption accepted that Christie's had assumed a responsibility to Ms Thomson, but he did not accept that Christie's owed Ms Thomson or the bidding public generally a duty arising only from the terms of the catalogue. That said, Christie's do accept that, if there is material doubt as to the description or dating of a work of art offered for sale at auction, good auctioneering practise requires the auctioneer to articulate the doubt in suitable terms, for example "probably Louis XV".
74. The judge addressed the first stage of the issue by asking whether, in cataloguing the vases, Christie's exercised the skill reasonably to be expected of an international auction house of their standing. Having cited from Luxmoore-May v Messenger May Baverstock [1990] 1 WLR 1009, he said (paragraphs 188 and 189):

"It is well-established practice of auction houses and the art world generally to date items such as the Houghton urns by visual examination in conjunction with such information as to the item as may be already available or obtainable by such research as is reasonably to be expected of the auction house in the circumstances. The examination must be as thorough as the circumstances reasonably require and it must be done by persons of appropriate qualifications and experience. It will not, cannot, and need not be as exhaustive as the examinations which have been carried out for the purposes of this trial: that is impractical. The outcome will be an expression of opinion as to date, which may be expressed in absolute terms such as here, or in more guarded terms, using words such as "probably" or "possibly". Where it is expressed in absolute terms it remains nonetheless an expression of opinion, which may later be shown to be doubtful or wrong. That is the basis on which the market proceeds, and that is generally well-understood.

It is difficult to define the degree of certainty which an auctioneer should have before he ascribes a date without qualification, and it is probably unnecessary for me to try to do so. It may be whether, having reviewed all factors, he finds he has a sufficient positive basis for the view he has formed combined with an absence of matters which raise a real rather than a fanciful doubt by pointing another way."

75. As Mr Sumption demonstrated, these paragraphs were dependent on, and justified by, the evidence of Mr Bourne. In the light of criticisms of Mr Paterson's qualifications and credibility, there was no persuasive countervailing evidence. Nor was it suggested on behalf of Ms Thomson that Mr Bourne's approach was impermissible (paragraph 190).
76. The judge examined eleven reasons, advanced on behalf of Ms Thomson, why it was unreasonable for Christie's to have reached a "certain or definite opinion" that the vases were made in the reign of Louis XV. These were (paragraph 192):

"(1) The 18^th/19^th century problem arising from copies and imitations: this is a reason for care; it does not preclude the identification of a decorative object as 18^th century by means of its visual features.

(2) Lack of provenance: this is very common, but is a reason for caution.

(3) Electrogilding of the liners: Christie's did not observe this, nor did M. Moreira when he first saw the urns, even though Mr Bourne said that it was obvious; but even if it had been spotted, Christie's could have concluded that it was a re-gilding, perhaps following the repair to one liner. They did not then know that the Getty and Greek urn liners probably had also been electro-gilded.

(4) There were a number of unusual features – the proportion of gilding to porphyry, the use of scagliola, the use of liners, the lack of comparable design: these points largely arise from Petitot's design.

(5) The finish of the porphyry and the quality of the chasing: I refer to what I have already said.

(6) The features of the design: it was an 18^th century design; I do not find Mr Paterson's point as to the portrayal of the lions being typical of the 19^th century persuasive.

(7) Casting and chasing: I refer to what I have already said.

(8) Gillian Wilson's view: she gave a date to the Getty urns of 1765 to 1770, which in so far as it is relevant to the Houghton urns supports Christie's on date. Her uncertainty as to where they were made does not reflect on the date.

(9) The existence of 19^th century vases after the Petitot design: these were identified as 19^th century, which might encourage Christie's in distinguishing them and confirming their own dating.

(10) The comparison with the Getty urns made by Christie's was by photograph only. This was a reason for Christie's not to rely on the Getty dating of its urns: for without a full comparison Christie's could not know how similar the two pairs were and so whether they could draw any support at all for their dating. Christie's were entitled to conclude from the Getty literature and the correspondence only that there was another pair of similar urns which had been dated 1765 to 1770 by an institution of high repute and expertise. If Christie's concluded independently that the Houghton urns were also of that date, they could find some confirmation and encouragement in the Getty dating of their urns because the Getty dating strongly suggested that urns to Petitot's design were made in the 18^th century.

(11) The Bossi engraving: I refer to what I have already said about this aspect."

77. The judge had discussed these matters in detail in earlier paragraphs of his judgment. He summarised his views in the passage which I have just quoted. He then said (paragraphs 193 and 194):

"Mr Bourne was confident in his opinion that the urns were approximately contemporary with the Bossi engraving, and he was supported by M. Moreira, a man of greater expertise in some aspects. I do not think that they were negligent in how they reached that conclusion. It was one which they and an auction house such as Christie's could come to consistently with the exercise of the skill to be expected of an auctioneer of Christie's standing. Mr Bourne and M. Moreira considered the points on which Ms Thomson relies, and they had reasons for rejecting them which were reasonable.

I conclude that Christie's dating of the urns as Louis XV without qualification was an opinion which an auctioneer of their standing could reasonably reach, and that there was no breach of duty on their part in that regard."

78. In my judgment, Mr Sumption is correct to submit that an amalgam of paragraphs 188 to 194 of the judgment shows the judge's decision to be that Christie's at the time of the sale held, and were reasonably entitled to hold, the certain and definite opinion that the Houghton vases were 18^th century and correctly described without qualification as "Louis XV"; that there were no real rather than fanciful doubts pointing another way, and that Ms Thomson's eleven reasons did not raise any real doubts. These conclusions were convincingly supported by the confident evidence of Mr Bourne, M. Moreira and Professor del Bufalo in particular. Mr Sumption submits that the judge's conclusions here are not amenable to appeal. He goes on to submit that the judge's later conclusion that Christie's were nevertheless in breach of duty to Ms Thomson is inconsistent with these unappealable earlier conclusions.
79. Notwithstanding his conclusion that Christie's were not in breach of their duty as an auctioneer in dating the vases in their catalogue as Louis XV, the judge nevertheless concluded that Christie's cataloguing fell below the standard to be expected in certain respects (paragraph 195).
80. First, he did not think that Christie's were justified in characterising the vases as Louis XV without the qualification that they may have been made in Italy, in particular in Parma. The judge accepted that "possibly Italian", removed from an earlier draft of the catalogue, may well not have affected the price, but an unqualified description of "Louis XV" was "an unjustified firming-up of the catalogue description."
81. Second, the prose description that the Houghton vases were designed by Petitot for the Duke of Parma or one of his leading courtiers was a wholly unjustified boost. The most that might be said was that the vases were after a design by Petitot who had worked at the court of Parma.
82. Third, there was no ground for thinking that "the antique porphyry vases were reworked by Petitot", as was stated in the condition report. This report was not seen by Ms Thomson. But Mr Cooney had seen it and it would tend to increase his feeling of confidence.
83. These matters, said the judge, all provided a feeling of confidence and certainty about the vases which was unjustified. Mr Sumption criticised these findings. The prose in the catalogue made explicitly clear the Italian (Parma) connection with the design, which nevertheless did not detract from the appropriateness of the unqualified Louis XV description. The design was French. Petitot was French. The court at Parma was a satellite of the French court in Paris. The judge's second and third points were of little or no significance, especially since Ms Thomson did not see the condition report. The real battleground was between an 18^th century and a 19^th century date and neither of the points detracted from a secure 18^th century conclusion. The judge himself acknowledged that the first of his points would probably not affect the value of the vases. Neither, in my view, would the second or third.
84. The judge considered that Christie's case that they were just as entitled to describe the vases as Louis XV without qualification to Ms Thomson as to the readers of their catalogue generally did not take sufficient account of her relationship with them as a special client (paragraph 197).

"As Christie's knew, Ms Thomson was not taking any independent advice since she was looking to them. Christie's knew that this was a new field to her and that she had never made a comparable purchase before. Christie's took upon themselves to inform Ms Thomson through Mr Cooney of the very high quality of the urns as examples of the period, plainly with the object of encouraging her to bid for them. Christie's gave what was in the circumstances an incomplete picture and Ms Thomson was entitled to a fuller one … I am, however satisfied that the circumstances required them to tell her that the dating of objects in this area was an unusually difficult one because of the copies and imitations of 18^th century items which were made in the Second Empire and following, and of the risk that this entailed. Ms Thomson should have been told that Christie's were relying very largely simply upon the exercise of their judgment following visual inspection. She should have been told that the catalogue inflated what could properly be said about the urns and was likely to give a misleading impression about Christie's knowledge and the sureness of their judgment. The circumstances require that Ms Thomson be told the risk she ran in paying a large sum for the urns."

85. Mr Sumption took issue with each of the judge's three reasons. The first was inconsistent with the judge's earlier decision (in paragraphs 192(a) and 194 taken with 189) that the possible existence of 19^th century imitations was a reason for care, but did not in the case of the Houghton vases raise a real, rather than a fanciful doubt. Christie's were not obliged to express fanciful doubts to Ms Thomson. The second and third reasons were not supported by Ms Thomson's own evidence. She had accepted that Christie's view depended on an exercise of judgment. The judge had correctly described Christie's position in paragraph 188 of his judgment. There was no reason to detract from this in Ms Thomson's case. Ms Thomson had not said that she was influenced by features of the catalogue prose which the judge had criticised.
86. The judge held that, if Ms Thomson had been advised on the lines he had held that she should have been advised, she would not have bought the vases. She had said that, if she had had any suspicion that they were mid-19^th century and not as described in the catalogue – had she known of the slightest question mark over their authenticity – she would not have purchased them. The judge accepted that, if she had been told of the 18^th or 19^th century problem, she would not have bid. Mr Sumption points out that Ms Thomson's case was that the vases were 19^th century; that the judge had held that the existence (in general) of 19^th century imitations was a ground for caution, but not in this case for doubt; and that he had held that there were no real grounds to doubt the authenticity of the vases.
87. The judge went on to hold shortly (paragraph 199) that Ms Thomson was also entitled to succeed on her claim for damages under s.2(1) of the Misrepresentation Act 1967. Although Christie's were entitled to describe the vases in their catalogue as Louis XV without qualification, their duty to Ms Thomson required them to qualify to her the catalogue description. Christie's did not therefore have reasonable grounds for expressing to her the unqualified opinion which they did through Mr Cooney. It is the same point.
88. The judge said that there was agreement as to the measure of damages, but any agreement there may have been has not carried through into this court. The judge expressed the measure of damages for any breach of duty which caused Ms Thomson to purchase the vases as "the difference between what she paid including buyer's premium and their actual worth". He said that there were disputes as to the date at which the value should be taken and as to whether the metallurgical evidence should be taken into account. He considered that Ms Thomson would receive appropriate compensation if she had "the difference between the price paid [for the vases] and their value at the time of purchase on the basis of my finding as to the probable date" plus interest. There was no evidence that their value had altered significantly between 1994 and 2000. The fact that Christie's could not be expected to use metallurgical testing was irrelevant to damages. The metallurgical evidence was simply a factor which the judge had taken into account in his assessment of the vases' likely date. The judge then discussed damages for misrepresentation which were limited to a proportion of the buyer's premium.
89. As to the quantification of damages, there was no evidence of the vases' value now, since the parties' evidence had gone to their value if they were 19^th century. "Their value now has to be assessed on the basis that the Houghton urns are probably to be dated between 1760 and 1765, the degree of probability being of the order of 70%" (paragraph 204).

Grounds of Appeal and submissions

90. A summary of Christie's case on appeal is as follows:

(1) the judge found in Ms Thomson's favour, although he rejected her pleaded case in its entirety. She alleged that the vases were made in the 19^th century and that Christie's were negligent in describing them as Louis XV without qualification. The judge held that the vases were made in the 18^th century and that it was reasonable for Christie's so to describe them without qualification. This should have resulted in the judge dismissing her claim. He found in her favour only by reformulating her case in a way that was not even argued at trial.

(2) The successful reformulated claim was unsupported by the evidence. Christie's were confident in their unqualified dating of the vases. The judge found that their opinion was correct and reasonably held. He also found that Christie's had no reason to know of any contrary opinion. The judge nevertheless held them negligent for not warning Ms Thomson about the possible existence of opinions different from theirs. This is a very peculiar result. It is even more peculiar when the judge held that she was entitled to substantial damages, although she got what she was paying for and paid no more than the open market price.

(3) The judge's decision turned a professional obligation to exercise skill and care in stating an opinion into an insurance obligation against the possible existence of contrary opinions.

91. Christie's say, with justification, I think, that Ms Thomson's pleaded case was that, contrary to the catalogue description, the vases were made in the 19^th century. Ms Thomson had alleged that, before the sale, Christie's were aware of queries and doubts in the market about the age and attribution of the vases, but the judge rejected this contention. They owed Ms Thomson duties to use reasonable skill and care in describing and dating the vases and to inform her of any doubts of which they were aware. They were in breach of these duties. They failed to take account of features of the vases which indicated or were only consistent with 19^th century manufacture. If they had advised her properly, she would not have bought the vases on the terms which she did. Their statements amounted to actionable misrepresentations. Ms Thomson was thus entitled as damages to the difference between what she paid and the value of the vases as 19^th century objects at the time of the sale.
92. As I have said, Christie's accepted during the trial that they owed Ms Thomson a duty to use reasonable skill and care in describing the vases as Louis XV and in making statements to her about their date and quality. They accepted that they assumed a responsibility to her for the accuracy of the information provided to her by Mr Cooney. Christie's maintain that Ms Thomson's case at trial was that a careful adviser should have advised her that there were serious reasons for doubt about the date of the vases and that she risked losing a lot of money. Her case remained that the vases were in fact worth no more than about £25,000. Her case did not extend to an allegation that Christie's were in breach of a duty to advise caution, even though their own opinion as to the dating of the vases was confident, unqualified and reasonably held. Mr Sumption points to exchanges between Ms Thomson's counsel and the judge to the effect that her case was that the vases were made in the 19^th century, and that, if she failed to establish this, the case failed "more or less". Mr Sumption contends that Ms Thomson's case never extended to saying that mere knowledge of a 19^th century revivalist fashion would have caused her not to bid; or that she had suffered any loss if the vases were and were reasonably described as 18^th century.
93. Mr Sumption submits accordingly that Ms Thomson's pleaded case failed completely. I think that this submission is largely justified. The main case advanced at trial was that the vases were 19^th century, although it shaded off on occasions towards a case that there were material doubts whether the vases were truly 18^th century. I am, however, persuaded that although the sub-species of breach of duty which the judge found was not much canvassed at trial, that it was nevertheless a theoretically legitimate sub-species of the breaches of duty which were canvassed. Mr Sumption contends that this is a point of substance, not mere pleading. If the case which succeeded had been advanced, different or additional evidence would have been adduced from Mr Rock, Mr Cooney and Mr Bourne and perhaps others. They would have been asked whether in their view it was a professional practice and obligation to warn a client of the obvious theoretical risk that others might not agree with Christie's own confident, unqualified and reasonably held opinion. Christie's should not have been found to have been professionally negligent when no allegation of such negligence in the respects found by the judge was made or put to the individuals concerned.
94. As will appear, I consider that the main point in this appeal turns on whether Christie's ought reasonably to have advised Ms Thomson of material doubts as to the dating of the vases. That predicates the question whether there were such doubts which Christie's ought reasonably to have held. I do not consider that Ms Thomson's case should founder on a formalistic pleading point. In so far as Mr Sumption's submissions here have substance, they contribute to a conclusion of substance.
95. People, including professional people, who, by giving information in the nature of advice, assume a responsibility giving rise to a duty of care, do not thereby normally undertake to draw attention to the obvious – see Tomlinson v Congleton Borough Council [2004] 1 AC 46. What is to be regarded as obvious depends on the characteristics and experience of the person receiving the information. Nor are they obliged to draw attention to risks which are fanciful, although of course some risks which are very small may be anything but fanciful, as, for instance, in cases of medical or surgical treatment.
96. It is Christie's case that they did not proffer, and Ms Thomson did not ask for, any opinion other than an opinion about the date and quality of the vases. She did not ask for, and was not entitled to expect, general advice about whether or not to buy the vases, nor as to any more general commercial risk in doing so. Ms Thomson was indeed a special client, but the services offered to special clients were generically the same as those available to any potential buyer.
97. As to the three respects in which the judge held Christie's to have been in breach of duty (paragraph 197 of the judgment) Christie's submissions are in summary as follows:

(1) The existence of a 19^th century revivalist fashion was a reason for caution, but not, on the facts of the present case, for doubt. The judge had himself so held in rejecting the first of Ms Thomson's eleven reasons for doubt in paragraph 192(1) of his judgment. This finding was fully supported by the evidence, particularly that of Mr Bourne. The relevant part of paragraph 197 is unjustifiably inconsistent with this. The point was not put to either Mr Cooney or Mr Bourne. There was no evidence that the existence of the revivalist fashion made dating 18^th century objects unusually difficult.

(2) It is obvious to any buyer, whether expert or not, that attributions as to date and views as to quality are matters of opinion and judgment. Christie's were under no duty to express the tautological platitude that their confident unqualified opinion was nevertheless an opinion. Mr Sumption referred here to Luxmoore-May v. Messenger May Baverstock at 1020F and Harlingdon Enterprises v. Christopher Hull Fine Art Ltd [1991] 1 QB 564 at 577E. Ms Thomson's complaint was, not that she was not told that Christie's were just expressing their opinion, but that there were serious reasons for not giving that opinion. She did not expect to be told of the process by which Christie's reached their opinion. There was no basis for the judge's conclusion that Ms Thomson should have been told that the opinion depended largely on visual inspection.

(3) As to the judge's view that the catalogue unjustifiably inflated what could properly be said about the vases, the points made were insubstantial and did not affect Christie's confident unqualified opinion as to the dating and quality of the vases. None of the matters in paragraph 195 of the judgment – "possibly Italian", "designed by Petitot for the Duke of Parma", "reworked by Petitot" – created a misleading impression of the sureness of Christie's judgment. They were sure for other reasons, and were supported in that by Mr Bourne, whose evidence the judge accepted. None of these matters were said to have been material to Ms Thomson's decision to bid. None of them affected the value of the vases. The case was that the catalogue entry was wrong because the vases were made in the 19^th century. There was no doubt as to the museum quality of the Houghton vases as the existence of the Getty vases demonstrated.

98. In summary, Mr Sumption submits that, however the duty is put, it can only be a duty to disclose matters which Christie's knew or should have known and which would have materially affected their confidence in their own opinion. None of the matters relied on by the judge were shown to have been material to Ms Thomson's decision to bid. Her case depended on the vases being 19^th century.
99. Mr Sumption submits that the judge's decision as to the measure of damage was wrong. If breach of duty is established, the measure of damage is the difference between what Ms Thomson paid and the value of the vases in 1994 as they should have been described to her. This results in no loss, since, on the judge's finding, they should have been described as they were described and the facts relied on by the judge did not affect their value. This is intrinsically just a different version of the case that there was no breach of duty. If on the other hand there were material doubts about the 18^th century date, so that, for instance, Christie's should have described the vases to Ms Thomson as "probably Louis XV", the measure of damage would be the difference between what Ms Thomson paid and the auction value in 1994 of vases so described (for which there was no evidence). The judge's 70% has no part to play in this. This was the measure of his own judicial confidence in his decision that the vases were made in the 18^th century. It also took account of the metallurgical evidence which was not relevant either to Christie's duty or to the appropriate measure of damage. For neither purpose would metallurgical evidence have been in play. These submissions apply equally to the misrepresentation claim.
100. Ms Thomson's case on Christie's appeal is essentially as follows. Christie's assumed a duty to give her a careful and balanced appraisal of the merits of purchase. The duty was not onerous. If there were grounds for uncertainty, Christie's should have told her of them. They eulogised the vases, giving the impression that they were undoubtedly 18^th century. But they really knew very little about them. Ms Thomson's case remains that on the evidence the judge should have found that the vases were made in the 19^th century. But, even if this is not established, there were numerous features about the vases which raised doubts about their true origin, history and age, such that Christie's could not properly have described them as Louis XV without qualification. At most, they could only properly have been described as "probably Louis XV". The risk of loss was very serious. If Christie's had expressed doubts or qualifications, Ms Thomson would not have bid. She has suffered real loss. If the vases are 19^th century, their value is no more than £25,000. If they are probably, but not certainly 18^th century, the judge's 70% route to assessment is not inappropriate, although Lord Grabiner's support for it was somewhat less than enthusiastic.
101. In support of the submission that Christie's did not hold, or should not have held, an unqualified view about the date of the vases, Ms Thomson points to evidence of Mr Rock in which he accepted that it is in the nature of the auction business that "you put your best foot forward" and that there is a range of possible view about all vases. This showed that Christie's were describing the vases in the best possible light. Mr Rock had agreed that the lot was among the most important and interesting ones in the sale and the thing Christie's knew least about. He did not agree that Christie's were desperate to bolster their description.
102. Ms Thomson relies on the judge's findings in paragraph 197 of his judgment as justified. By a Respondent's Notice, she contends that there were additional reasons to support the judge's decision. In summary, these are the eleven reasons considered and rejected by the judge in paragraph 192 of his judgment. She further contends that the judge was wrong to hold that Christie's were not negligent in their unqualified catalogue description. They had no reasonable basis for reaching their certain and unqualified conclusion that the vases were made in the reign of Louis XV.
103. Ms Thomson submits that the judge gave too little weight to the difficulty arising from the existence of the 19^th century revivalist fashion. Christie's distinction between a reason for caution and for doubt is artificial. The lack of provenance of the Houghton vases (and the Getty and Greek vases) was more than a reason for caution. It ought to have raised significant doubts. Christie's were negligent in not spotting the fact that the liners were electrolytically gilded. If they had done so, they would have discovered that the Getty liners were also electrolytically gilded. The judge should have concluded from this alone that Christie's could not reasonably conclude that there were no possible doubts about the description of the vases. In addition, the proportion of porphyry to gilding – the gilt bronze largely masking the porphyry – was atypical of the 18^th century. The porphyry has a strikingly poor finish. Petitot's design was for large stone garden vases, not ormolu decorative objects. There was no evidence linking the design with the date of manufacture. The casting, chasing and pouncing were not secure indications of date. 18^th century tools may have survived into the 19^th century. Gillian Wilson had expressed uncertainty whether the vases were French. Christie's view did not reflect this uncertainty. Christie's knew that there was at least one pair of vases of this model made in the 19^th century. They did not investigate this. Christie's relied too much on the Getty vases and their similarity with the Houghton vases, without sufficiently investigating them. Christie's could not reasonably draw any conclusion from the existence of the Bossi engraving which was no reliable indicator of the date of the vases. Taking all these points together, it was unreasonable for Christie's to reach any certain or definite opinion about the date of the vases.
104. Ms Thomson relies on the judge's conclusion that Chrsitie's were guilty of boosting their catalogue description. It was implicit in this finding that otherwise their opinion was less than confident. The prose in the catalogue wrongly suggested a factual anchor to the 18^th century. In all the circumstances, Christie's should have qualified their catalogue description.
105. In any event Ms Thomson says that the judge's conclusion about the age of the vases was wrong. In addition to the matters already summarised, she says that the judge wrongly assessed the evidence about the electrolytic gilding of the liners, the repair and the cumulative affect of the metallurgical evidence.
106. As to the gilding of the liners, Ms Thomson's case before trial, supported by Mr Paterson and Dr Northover, was that the liners were electrolytically gilded. Christie's experts maintained that the liners had originally been mercury gilded and had later been regilded to conceal the repair to one of them. There were, they said, visible traces of mercury gilding. Ms Thomson's experts disagreed that there were visible signs of earlier mercury gilding and pointed out that the unrepaired Getty and Greek vases also had liners which were electrolytically gilded. The judge was incorrect to say that Ms Thomson's case had changed here. Christie's case during the first days of the trial remained that the liners had been regilded. Further analysis by Dr Northover detected no mercury. There was also no diffusion layer, such as the heat of mercury gilding would produce. It was only on day 12 of the trial that Mr Bourne first produced the implausible, last gasp theory that the liners may originally have been ungilded. It was improbable that a manufacturer of such an important decorative object would have left the liners ungilded with unsightly seams. The judge's idea that the rims only may have been gilded originally was profoundly unlikely. The only reasonable explanation was that the liners were made in the 19^th century.
107. As to the repair, Dr Northover's opinion was that the metallurgical composition of the brass in the repair and the rest of the liner was identical, and that the silver solders were essentially the same. The patch was inserted using a blow torch, not invented until 1882. Dr Bennett had no convincing answer to this. The judge was wrong to reject Dr Northover's mean results. He gave too little weight to Dr Nothover's persuasive evidence. Similar points may be made about the way in which the liners were made. The judge's own point about the uneven thickness of the walls of the liners at the seam was unpersuasive.
108. Points about mercury gilding and the repair were two of the four matters which the judge regarded in his paragraph 180 as indicating an 18^th century date. The judge wrongly assessed the metallurgical evidence here which should have led to the opposite conclusion. The judge's remaining points – about the chasing and pouncing and the surface of the porphyry - were no sure guide to the date of manufacture.
109. Ms Thomson submits that the judge wrongly assessed the cumulative persuasive effect of the metallurgical evidence taken as a whole and Dr Northover's expert opinion based on it. The judge accepted that the liners were electrolytically gilded and that it was probable that they had never been mercury gilded. The liners of the Getty and Greek vases are also electrolytically gilded. It is very unlikely that the liners were originally ungilded. It is very likely that the liners were made from a sheet of rolled brass, a 19^th century innovation. It is very likely that the solder for the repair was the same as that used elsewhere in the liners and that the brass for the repair was also the same as that used for the liners themselves. It is very likely that the brass for both the mounts and the liners was made with electrolytically refined copper and with zinc from which the iron had been removed electromagnetically, both 19^th century techniques.
110. Ms Thomson submits that Christie's only real point in answer to Dr Northover's evidence was the suggestion that Dr Northover's database was too small to enable meaningful conclusions to be drawn. The judge should not have accepted this. On the contrary, he should have accepted Dr Northover's experience and opinion that comparisons with admittedly 18^th century objects in the database was properly capable of indicating probative similarities or dissimilarities. The objects in the database were not the only ones available for comparison. They and all the others had levels of impurities higher than the mean values obtained from the Houghton vases. The purity of the brass and solder was difficult to explain other than by reference to 19^th century developments.
111. Lord Grabiner accepted that damages should be assessed as at 1994. He contended that Mr Sumption's submissions misunderstood or misapplied South Australian Asset Management Corporation (SAAMCO) v. York Montague [1997] AC 191. The claimant has to establish a duty to guard against the kind of loss which she has suffered. The kind of loss was loss arising from buying comparatively worthless objects in the false belief that they were extremely valuable. The judge therefore correctly decided that the measure of damage was the difference between what Ms Thomson paid for the vases and their actual value. Their actual value is their value as 19^th century objects, if the court so finds on Ms Thomson's cross-appeal; or as objects which cannot certainly be said to be 18^th century upon the judge's finding. The judge's 70% approach was sensible and one which Ms Thomson could "live with". Unless the vases are found conclusively to be 18^th century, Ms Thomson has suffered some loss and the court's task is how best to assess it. There are no losses, such as there were in SAAMCO, attributable to factors extraneous to Christie's breach of duty. In any event, Mr Cooney gave advice to Ms Thomson. His role was not limited to providing information.
112. In response to Ms Thomson's appeal against the judge's findings relating to negligence and the date of manufacture of the vases, Christie's say that these require the Court of Appeal to re-evaluate the whole of the factual and expert evidence. They say that there are no grounds for embarking on such a re-evaluation. It cannot be sufficient to take pot shots against part of the evidence. The appeal is a review, not a rehearing.
113. There was a mass of evidence to justify the judge's conclusions. He analysed and weighed the evidence in great detail. Christie's expert witnesses were of great distinction in their fields. They expressed firm and reasonable conclusions. It was open to the judge to accept them. The fact that Ms Thomson's experts expressed different views is irrelevant to an appeal.
114. There was in fact no reliable evidence to support a contrary view of Christie's cataloguing. Mr Paterson's evidence was embarrassing and did not begin to measure up to Mr Bourne's.
115. As to the actual date of the vases, the cross-appeal relies myopically on metallurgical evidence, mainly that relating to the liners. It ignores the other mass of evidence indicating an 18^th century date. If the court were to reopen the judge's 70%, this court should find that the probability of an 18^th century date is very much higher than that.
116. On the question of Christie's alleged negligence, the issue was whether, by the standards to be expected of international auctioneers, Christie's unqualified opinion that the vases were Louis XV was reasonable. As Slade LJ said in Luxmoore-May at 1020F, the valuation of pictures of which the artist is unknown pre-eminently involves an exercise of opinion and judgment which may be fallible – see also Bolam v. Friern Hospital Management Committee [1957] 1 WLR 582 and Maynard v. West Midlands Regional Health Authority [1984] 1 WLR 632. A claim for professional negligence requires distinct proof of matters which are for expert evidence. As Lord Browne-Wilkinson said in Bolitho v. City and Hackney Health Authority [1998] AC 232 at 243:

"It is only when a judge can be satisfied that the body of expert opinion cannot be logically supported at all that such opinion will not provide the bench-mark by reference to which the defendant's conduct falls to be assessed."

By this standard, Ms Thomson would not establish that Christie's were negligent, even if there were evidence, persuasive if taken alone, contrary to that of Christie's experts. In fact, there was no such persuasive evidence.

117. Christie's say that the balance of Ms Thomson's case has shifted. At trial, her main case was that the vases were 19^th century and that Christie's should have so concluded and advised. Now her main case is that the vases are no more than "probably Louis XV" and Christie's should have so concluded and advised. She continues to rely on the eleven reasons, which the judge rightly rejected. Most of these reasons depended on the opinion of Mr Paterson which was discredited and unpersuasive. Ms Thomson's written case on appeal does not mention Mr Paterson once, yet his evidence was essential to her case. By contrast, the combined weight of the evidence of Mr Rock, Mr Cooney, Lord Hindlip, Mr Bourne, M. Moreira and Prof del Bufalo, taken with documentary and historical links with the Petitot design, fully sustain the judge's conclusion.
118. Christie's submit that the judge was right to be unpersuaded by Dr Northover's metallurgical analyses. No one, including Dr Northover, had previously done work on the composition of French 18^th century metal work. The metallurgical evidence was unpersuasive because:

(1) the 18^th century database which Dr Northover used for comparison consisted of 22 samples from only 7 objects. This could not begin to show that brass with other characteristics was not made during the period.

(2) the reference points for Dr Northover's opinion about copper impurity were a study of 19 French guns (1750-1810) and one 1857 boiler tube from a British steam engine. These were an entirely insecure guide to copper used in foundries making decorative brass.

(3) the 18^th century Flemish and Dutch instruments produced by Dr Northover showed impurities at a low level.

(4) without secure information about the sources of copper and the process by which brass was made in 18^th century France, it could not be concluded that low impurity levels in the copper were unusual for the 18^th century.

(5) no comparative conclusion about zinc content could be drawn from Dr Northover's database. He had not made any study of the methods of brass manufacture in 18^th century France.

(6) Dr Northover's 1999 and 2000 reports were not consistent with his reports to the court nor the confidence with which he expressed his conclusions.

(7) Dr Northover was a good deal more confident in his conclusion than the logical consideration of the available data justified.

(8) whatever its weight, the metallurgical evidence was no more than one factor in an overall assessment.

119. As to the liners and their electrolytic gilding, the evidence was introduced late and developed up to the end of the trial. Ms Thomson's case had to be that the repair to one of the liners was carried out at the time of manufacture, which, since the repair was affected by torch brazing, had to have been after 1882. Ms Thomson's case also was (1) that there was no evidence of mercury gilding on the liners; (2) that the brass and solder in the repair and in the walls and base of the liners are identical; and (3) that the liners were made of rolled brass, a process invented in the 19^th century.
120. Although the judge found that the liners had probably never been mercury gilded, M. Moreira, Mr Bourne and Dr Bennett all gave evidence that they saw indications of it. Dr Northover had said the same in 1999, although he had withdrawn this. Dr Northover only showed that there was no mercury on the tiny samples he tested. More importantly, there was no reason to suppose that the liners were originally gilded at all and Mr Bourne gave evidence that they may not have been. The judge accepted this evidence, as he was entitled to do. He was right to be impressed by Mr Bourne as a witness. Mr Bourne was cross-examined at length. His opinion had support from Sir Francis Watson, a leading authority on 18^th century French furniture. He had stated in his introduction to the Wallace Collection catalogues that gilding was only applied to the finest bronzes. There was a convincing explanation for the fact that the idea only surfaced on day 12 of the trial. It is a plausible explanation which avoids serious inplausibilities in Ms Thomson's case. These include (a) that a damaged liner was used at manufacture; (b) that the mounts and the liners were gilded at the same time by different processes, one of them illegal; and (c) that the maker acquired and then vandalised valuable pre-1800 porphyry mortars to make the vases. Christie's say that Ms Thomson has never had any answer to the evidence that forge brazing was used to make the mounts and the liners (apart from the repair) and that Dr Northover accepted that no craftsman would use forge brazing if torch brazing was available. Nor has she any answer to the accepted fact that the mounts are mercury gilded. By 1882, mercury gilding had been illegal for many years.
121. As to the similarity between the brass and solder used in the repair and to join the walls and base of the liners, the judge did not find that they were identical. This did not prove anything as there was no evidence about the respective compositions in the 18^th and 19^th centuries. The judge was entitled to be sceptical about Dr Northover's mean analyses and about the British Museum's recognition of his methods. The judge was entitled to accept M. Moreira's opinion that the brass used for the liners was cast, not rolled. The evidence in favour of rolling was exceptionally weak.
122. Christie's submit that Ms Thomson's case on appeal as to the date of the vases does no more than try to eliminate pieces of the judge's decision. For this court to decide that the judge was wrong would require a re-evaluation by this court of the whole of the evidence, a process which the judge undertook. Such a re-evaluation should not be entertained. But if it were, the conclusion should be that the judge's 70% was too low. With the sole exception of the electrolytic gilding of the liners, the judge did not find a single feature of the vases which positively indicated a 19^th century date. If the judge had stood back and considered the composite effect of his own conclusions, he would have appreciated that there was only one answer. The differences of opinion were more apparent than real. The judge's assessment of the cumulative effects and shortcomings of the metallurgical evidence cannot be faulted.

Discussion

123. I consider the appropriate measure of damage first because, if Mr Sumption is correct about the measure of damage, the metallurgical evidence is irrelevant for all purposes. It is agreed that it is not relevant to the case that Christie's were in breach of duty. For the purpose of this discussion only, I assume that Christie's were in breach of duty to the extent found by the judge. It is also appropriate to assume alternatively that Christie's were in breach of duty in not describing the vases as "probably Louis XV" or with some such qualification; that is to assume that Ms Thomson's appeal succeeds in part.
124. It is now agreed that damages should be assessed as at the date of the sale. Ms Thomson's case is that the measure of damage is, as the judge held, the difference between what Ms Thomson paid for the vases and their actual value taking account of what is now known. Christie's case is that it is the difference between what Ms Thomson paid and the value of the vases if they had been correctly described or if she had been properly advised. This would be, or would approximate to, the price which would have been paid at the auction if the catalogue had described the vases as "probably Louis XV".
125. In any negligence claim, there is an essential intrinsic link between the breach of duty alleged and the damage caused by it. I cite for convenience a passage from my judgment in S v. Gloucestershire County Council [2001] Fam 313 at 337:

"A negligence claim is habitually analysed compartmentally by asking whether there was (a) a duty of care; (b) breach of that duty and (c) damage caused by the breach of duty. But damage is the essence of a cause of action in negligence and the critical question in a particular case is the composite one, that is whether the scope of the duty of care in the circumstances of the case is such as to embrace damage of the kind which the plaintiff claims to have suffered. As Lord Bridge of Harwich said in the Caparo case [1990] 2 AC 605, 627; "it is never sufficient to ask simply whether A owes B a duty of care. It is always necessary to determine the scope of the duty by reference to the kind of damage from which A must take care to save B harmless." Lord Oliver of Aylmerton emphasised the same point in Murphy v. Brentwood District Council [1991] 1 AC 398, 486 when he said:

"The essential question which has to be asked in every case, given that damage which is the essential ingredient of the action has occurred, is whether the relationship between the plaintiff and the defendant is such … that it imposes upon the latter a duty to take care to avoid or prevent that loss which has in fact been sustained."

This question necessarily subsumes the question whether the acts or omissions of the defendant caused the damage relied on."

The converse of this is that the measure of damage has to relate to the duty of care and its breach.

126. The leading case here is South Australian Asset Management Corporation (SAAMCO) v. York Montague [1997] AC 191. As Lord Hoffman said at page 213C:

"Rules which make the wrongdoer liable for all the consequences of his wrongful conduct are exceptional and need to be justified by some special policy. Normally the law limits liability to those consequences which are attributable to that which made the act wrongful. In the case of liability in negligence for providing inaccurate information, this would mean liability for the consequences of the information being inaccurate."

The loss must have a sufficient causal or connection with the subject matter of the duty (page 214B). A claimant has to prove both that he has suffered loss and that the loss falls within the scope of the duty (page 218B).

127. Lord Hoffmann stated the principle at 214C as follows:

"I think that one can to some extent generalise the principle upon which this response depends. It is that a person under a duty to take reasonable care to provide information on which someone else will decide upon a course of action is, if negligent, not generally regarded as responsible for all the consequences of that course of action. He is responsible only for the consequences of the information being wrong. A duty of care which imposes upon the informant responsibility for losses which would have occurred even if the information which he gave had been correct is not in my view fair and reasonable as between the parties. It is therefore inappropriate either as an implied term of a contract or as a tortious duty arising from the relationship between them.

The principle thus stated distinguishes between a duty to provide information for the purpose of enabling someone else to decide upon a course of action and a duty to advise someone as to what course of action he should take. If the duty is to advise whether or not a course of action should be taken, the adviser must take reasonable care to consider all the potential consequences of that course of action. If he is negligent, he will therefore be responsible for all the foreseeable loss which is a consequence of that course of action having been taken. If his duty is only to supply information, he must take reasonable care to ensure that the information is correct and, if he is negligent, will be responsible for all the foreseeable consequences of the information being wrong."

At page 216D, he contrasted this with the measure of damage where the defendant warrants that his information is accurate.

"The measure of damages in an action for breach of a duty to take care to provide accurate information must also be distinguished from the measure of damages for breach of a warranty that the information is accurate. In the case of breach of a duty of care, the measure of damages is the loss attributable to the inaccuracy of the information which the plaintiff has suffered by reason of having entered into the transaction on the assumption that the information was correct. One therefore compares the loss he has actually suffered with what his position would have been if he had not entered into the transaction and asks what element of this loss is attributable to the inaccuracy of the information. In the case of a warranty, one compares the plaintiff's position as a result of entering into the transaction with what it would have been if the information had been accurate. Both measures are concerned with the consequences of the inaccuracy of the information but the tort measure is the extent to which the plaintiff is worse off because the information was wrong whereas the warranty measure is the extent to which he would have been better off if the information had been right."

128. Ms Thomson does not contend that Christie's warranted their information. Lord Grabiner does however say that they were advisors and not mere providers of information. In so far as this matters, I do not think that Lord Grabiner is right here. Christie's duty was to provide advice in the nature of information to enable Ms Thomson herself to decide whether to bid or not. There were numerous elements of that decision which did not come within the scope of Christie's duty at all, such as her ability to afford the likely purchase price and the relative merits of spending the money on these vases rather than on something quite different.
129. I do not, however, think that a distinction between an advisor and a provider of advice in the nature of information is critical to determining the proper measure of damage. As I have said, the measure of damage has to relate to the scope of the responsibility assumed and to the breach of the resulting duty of care. Christie's duty did not extend to carrying out metallurgical analyses and the measure of damage should not embrace the result of such analyses. The breach of duty which the judge found related only to the provision of information.
130. I conclude that, on the assumptions which I am making, the measure of damage in this case would be the difference between what Ms Thomson paid for the vases and their value at auction in 1994 if Christie's had described them as "probably Louis XV". There is no evidence of what this second value would have been. So, if the assumptions represent the outcome of this appeal on matters of negligence, a further hearing to assess the damages would be needed.
131. My view accords with the decision of this court in Phillips v. Ward [1956] 1 All ER 874. This was a case in which a surveyor had negligently overlooked death watch beetle and worm in the timbers of the house he was employed to survey. The court held that the measure of damage was the amount of money which would put the purchaser into as good a position as if the surveying contract had been properly fulfilled. This was the difference between the value of the property in its assumed good condition and its value in the bad condition which should have been reported to the client – see Denning LJ at 875I-876A; Morris LJ at 877I and 878B; and Romer LJ at 879B. The court also held that the damage should be assessed at the date when the damage occurred, that is when the property was purchased. The judgments contemplate the possibility that, if proper advice had been given, the client would not have proceeded to purchase at all. The principles in Phillips v. Ward were applied in Perry v. Sidney Phillips [1982] 3 All ER 706, another surveyor's negligence case – see also County Personnel v. Pulver [1987] 1 WLR 916 at 927G. It is, I think, implicit in each of these cases that the surveyor would not be liable for, and the measure of damage would not embrace, a concealed defect which the surveyor did not spot, but without negligence. It would be supposed that the defect would remain concealed and that the notional sale price which would establish this part of the damages calculation would not take it into account. The surveyor does not warrant the value of the property surveyed – see Denning MR at 708d and Oliver LJ at 710c in the Perry case.
132. Mr Sumption submitted that, in assessing the market value of the vases in 1994 on the assumption that Christie's had given proper advice and information the court should not attribute to the market knowledge which it would not have had at that date, that is that the court does not apply hindsight. He relied on Lynall v. Inland Revenue Commissioners [1972] AC 680; Gardner v. Marsh & Parsons [1997] 1 WLR 489; and de Balkany v. Christie Manson & Wood Limited (1995) TLR 163.
133. Lynall is, I think, directly relevant. The Finance Act 1894 provided a method of valuation of property for estate duty purposes by reference to what the property would fetch if sold in the open market at the time of death. The property was a holding of shares in a private company. Confidentially there was a flotation in contemplation, but this was not public knowledge. The possibility of flotation would increase the value of the shares. The Revenue contended that the possibility should be taken into account. The House of Lords rejected this contention. As Lord Reid explained at page 694, the decision turned on what knowledge hypothetical bidders for the shares must be supposed to have had. The court had to consider what would have happened if the imaginary sale had taken place. One thing that would not happen would be that bidders would be omniscient. They would derive their knowledge from facts made available to them by the shareholder exposing his shares for sale. So here, the appropriate value of the vases would be their auction price in 1994 on the assumption that Christie's had described them as "probably Louis XV".
134. Gardner v. Marsh & Parsons is a case in which the majority of this court applied Phillips v. Ward in an orthodox way in a surveyor's negligence claim. Its interest is that the court rejected a submission that a hypothetical sale price should not be taken because the seller would not have sold at that price – see Hirst LJ at 496E, Peter Gibson LJ at 504 and Pill LJ at 511. In the present case, Ms Thomson's evidence was that she would not have bought the vases if Christie's had qualified their advice. That, however, does not, in my view, affect the measure of damage. She did in fact on any view acquire objects of value and it is for the court to assess their value on appropriate principles. It is not suggested that she became locked into a deal from which she could not escape.
135. In de Balkany, Morison J at first instance decided in an auctioneer's negligence claim that the measure of damage would be the difference between the market value of the painting as described and its value if carefully described – see page 183 of the report.
136. Lord Grabiner in effect contends that he court should value the vases on the basis that they suffered from a latent defect which has subsequently come to light, but which was not known at the time of the auction. The metallurgical evidence establishes, he says, that the vases are 19^th century, or, if not, it casts greater doubt on their 18^th century manufacture than was available at the time of the auction. That should be taken into account to establish the true value of the vases at the auction date. He relies on Naughton v. O'Callaghan [1990] 3 All ER 191 and he mentioned Smith New Court v. Citibank [1997] AC 254.
137. Naughton v. O'Callaghan is a fairly short first instance decision by Waller J, as he then was, decided on the citation of limited authority. The plaintiff bought a horse at the Newmarket sales upon an incorrect description of its pedigree. If the horse had been correctly described, it would have fetched only slightly less than the price which the plaintiff paid. But by the time the misrepresentation was discovered, its value had fallen very considerably because of its lack of success after the sale on the racecourse. Waller J held that the appropriate measure of damage was the difference between the sale price and the horse's value at the time the misrepresentation was discovered. He cited authority to the effect that damages are normally assessed as at the date of breach, but that this is not invariably so. He regarded the case before him as "different from the norm" for reasons which he gave at page 197j. For these reasons, which would not bear upon Ms Thomson's case, he regarded it as unjust to take the normal date of assessment. In the present case, it is agreed that the date of assessment should be 1994.
138. Smith New Court v. Citibank was a case of fraudulent misrepresentation in which the House of Lords endorsed the extended ambit of an award of damages in a case of fraud in Doyle v. Olby (Ironmongers) Limited [1969] 2 QB 158. So far as is relevant for present purposes, the essence of the decision was that the claimant must give credit for benefits which he has received including the market value of the property acquired normally as at the date of acquisition. But, as Lord Browne-Wilkinson said at page 267B:

"… such general rule is not to be inflexibly implied where to do so would prevent him obtaining full compensation for the wrong suffered; (5) although the circumstances in which the general rule should not apply cannot be comprehensibly stated, it will normally not apply where either (a) the misrepresentation has continued to operate after to the date of the acquisition of the asset so as to induce the plaintiff to retain the asset or (b) the circumstances of the case are such that the plaintiff is, by reason of the fraud, locked into the property."

None of this applies in the present case. It is not a fraud case. The assumed misrepresentation did not continue to operate after Ms Thomson acquired the vases. She was not locked in.

139. My view about the measure of damage means that the judge's decision that the actual value of the vases should take account of his assessment that they were probably 18^th century, the degree of probability being 70%, was wrong. It was wrong for two reasons; first, because it took account of the metallurgical evidence; and second, because a measure of the judge's personal level of confidence in his own decision made in 2004 has no bearing on the market value of the vases upon any assumption at any date, let alone as at 1994 upon an assumption that they were to be described as "probably Louis XV".
140. My view about the measure of damage also means that the metallurgical evidence and any decision as to date which takes account of it has no bearing on any issue in the appeal or cross-appeal. I shall nevertheless deal with these matters briefly because considerable time was spent on them, and because general considerations relevant to them are also relevant to other parts of the appeals.
141. The judge rejected Ms Thomson's case that the metallurgical evidence established that the vases were made in the 19^th century. The cross-appeal challenges this finding. For this challenge to succeed would mean that this court would have to be persuaded to reverse a finding of fact made by the judge upon a very careful composite evaluation of a large body of oral evidence, mainly expert evidence. It is a commonplace that this court will only rarely disturb a judge's finding of fact reached after hearing oral evidence whose credibility is in issue. It may sometimes be somewhat less difficult to depart from an evaluative judgment of expert evidence, if the evidence is mainly in writing and its evaluation an intellectual process largely unaffected by personal credibility. But that is not this case. The judge's factual conclusion that the vases were made in the 18^th century was a composite amalgam of a large number of particular judgments where an assessment of the weight and persuasiveness of each of the experts as witnesses was of critical importance. The judge was obviously impressed with Mr Bourne, M. Moreira and Professor del Bufalo. He was not impressed with Mr Paterson for reasons which may readily be articulated. He was impressed with Dr Northover's experience and expertise, but less so with his logic and judgment. It would, I think, be a travesty of the appellate process to interfere with the judge's composite judgment here as to the date when the vases were made unless it were shown that he made one or more glaring and important misjudgments. This is scarcely likely for such a thorough and careful judgment and the cross-appeal does not, in my judgment, begin to do so. The judge's analysis of the evidence, apart from the metallurgical evidence, in support of an 18^th century date is, in my view, unimpeachable. His conclusions are in each instance supported by credible evidence. The existence of other evidence does not render these judgments wrong nor does it persuade me that he should have reached an opposite conclusion. His assessment of the metallurgical evidence was equally careful and his rejection of Dr Northover's opinion was persuasive. Restating evidence and submissions which the judge rejected does not persuade me that the judge was wrong, when the judge's conclusions were supported, as they were, by the opinions of Christie's experts. I accept Lord Grabiner's general submission that some individual matters on which the experts expressed a view were scarcely matters for their professional expertise, but rather questions of common sense or logic. I accept in general that the court is as able to consider such matters as the experts. These would not truly be amenable to expert opinion at all. And I grant that there are examples of matters of this kind in the evidence, as, for instance, the likelihood that the liners of the vases may originally have been ungilded. Another, perhaps better, example is the sufficiency for persuasive comparative purposes of Dr Northover's database. On this point, my own understanding and assessment agrees with the judge's. Samples from a mere 7 objects can scarcely claim to be representative of 18^th century techniques. Nor is the exercise improved by adding material from some French guns and a mid-19^th century English steam engine. But, even accepting that individual points such as these are amenable to judicial appellate evaluation whatever the expert opinion, no appellate court should cherry pick a few such points so as to disagree with a composite first instance decision which, in the nature of a jig-saw, depended on the interlocking of a very large number of individual pieces, each the subject of oral expert evidence which the appellate court has not heard.
142. I do not propose to plough through all the individual submissions made in support of or opposition to the cross-appeal as to the date of the vases taking account of the metallurgical evidence. I shall, however, address briefly two of the subject matters which formed part of the judge's evaluation – the porphyry and the gilding of the liners. I trust that these subjects are reasonably representative. The purpose of this is to demonstrate that the judge's conclusions were persuasive (or at the very least tenable) and that they were supported by evidence which he was entitled to regard as credible.
143. Particular consideration of the porphyry features in paragraphs 112 to 121 and 180(d) of the judgment, which I have summarised above, and in paragraphs 192(4) and (5) in brief answer to two of Ms Thomson's eleven points.
144. Ms Thomson's case in outline was that the surface of the porphyry was poor. Although Mr Paterson and Professor del Bufalo agreed that the exterior surfaces were turned at some time before 1800, Mr Yates thought that the surface was or could have been produced in the 19^th century. The porphyry had an intermediate existence as, for instance, mortars of poor quality. Such poor quality porphyry would not have been used in the 18^th century – compare classy 18^th century items of porphyry in the Wallace Collection. The judge's conclusion in paragraph 121 of his judgment was wrong, since he says nothing to show that the quality of the porphyry was consistent with an 18^th century date and inconsistent with a 19^th century date. Porphyry was highly valued in the 18^th century when much more of it would have been revealed. By the 19^th century, fashion had changed. There is no other example of a rough filler material in the base of a vase. All in all, the state of the porphyry was not good enough to be 18^th century. The fact that the design was 18^th century was not persuasive, since the design did not dictate a particular material.
145. Christie's case in outline was that the consensus of the evidence – Mr Paterson, Professor del Bufalo and, eventually, Mr Yates – was that the porphyry was fashioned into its present state before 1800. The bowls and all the metal parts were made at the same time, since each was made for the other. Admittedly you have to start with something. But it was far more likely that the bowls were core drilled at the time of manufacture from a column than that they had a previous independent existence as bowls. Why smash the base of 6 perfectly good existing bowls? There was nothing to suggest that this quality of porphyry would have been rejected in the 18^th century. Both Mr Bourne and Professor del Bufalo considered that the porphyry gave the vases a real 18^th century feel. Mr Yates was unable to refer to any example of 19^th century porphyry with the degree of pitting on the Houghton vases. In the 19^th century, there was an improved workshop ability to achieve smooth surfaces and high polishes.
146. The judge's decision here is plainly supported by evidence. It was not merely tenable, but, in my view, persuasive. Given that the porphyry was very probably fashioned before 1800 and that all the components of the vases were made at the same time, it seems to me highly improbable that existing bowls were smashed at the bottom and the whole then replaced with inferior fill material. The poor finish generally was more consistent with earlier workmanship when tools and machines were less developed. The opinions of Professor del Bufalo and Mr Bourne justified the judge's conclusion.
147. The judge considered the gilding of the liners in paragraph 174-176, and in paragraphs 180(b) and (c). He concluded that the reasonably strong probability was that the liners had never been mercury gilded, but only electrolytically. By contrast, the mounts were mercury gilded, a strong indicator in favour of an 18^th century date. It was not likely that two different methods of gilding were used at the same time. So the probability was that the liners were left ungilded at the time of manufacture.
148. Ms Thomson's case is that the electrolytic gilding of the liners is obvious and should have been spotted by Christie's. Metallurgically, there is no trace of previous mercury gilding. The idea that the liners may have originally been ungilded did not occur to anyone until the 12^th day of the trial. It is highly improbable. Original ungilded liners would have been unsightly with an obvious seam and inconsistent with the obvious importance of such important and supposedly valuable objects. Mr Bourne's idea that the vases may have been for display above eye level is scarcely persuasive. Most people decide where to put things after they are made. The judge's notion that the rims only may have been gilded was a misguided frolic of his own.
149. Christie's case was that there are visible indications of previous mercury gilding of the liners. That was the evidence of their witnesses, although the judge found that the liners had probably never been mercury gilded. Christie's did not therefore have reason to look for an alternative explanation for the electrolytic gilding until a fairly late stage of the trial. The judge thus accepted that the late introduction of the idea that the liners may originally have been ungilded did not affect its credibility. Mr Rock and Mr Cooney both gave evidence that regilding happened a lot. Reference was made to some account books of George IV. M. Moreira thought that regilding had taken place after previous mercury gilding. Mr Bourne considered that regilding was not a reason to affect the reasonableness of Christie's description. Mr Paterson never had an answer to the point about the repair, which provided a sound explanation for the electrolytic gilding. Taking Mr Bourne's evidence as a whole, he was convinced for a lot of sound reasons that the vases were 18^th century. The electrolytic gilding of the liners was odd and indicated that this gilding had been done in the 19^th century. But it did not detract from his firm view that the vases were made in the 18^th century.
150. The judge's conclusion here – that the liners were probably ungilded originally – was reasonably open to him on the evidence. It is not amenable to appeal. More importantly perhaps, the conclusion that, on the evidence taken as a whole, the vases were made in the 18^th century, was the result of a balanced amalgam judgment which the evidence as a whole fully justified. Taken in isolation, the decision about the gilding of the liners is perhaps mildly surprising, but this part of the evidence was not to be taken alone. The judge no doubt thought it intellectually necessary to reach conclusions every strand of which tied together neatly. But in some cases, this may be reaching for the moon. A rough edge or two or a point which is less than completely explained may perhaps be seen as natural in a case such as this. I would not see it as detracting from the judge's main conclusion.
151. I do not, as I say, propose to plough through all the points made in support of the cross-appeal on the date of the vases taking account of the metallurgical evidence, since, in the light of my view about the measure of damage, a decision on this part of the case has no bearing on any issue before this court. However, taking the matter compendiously, I am entirely satisfied that the judge's decision on the strong balance of probability that the vases were made in the 18^th century was correct and impervious to appeal. There was ample evidence to justify his conclusions, both on individual points and as a whole. He heard the evidence and we have not. His judgment was balanced and amply reasoned. There is no proper basis for this court to interfere.
152. I have also already explained why in my view the judge's 70% was wrongly used by him on the question of damages and is not material to any issue before this court. There is no need to reconsider it. I have referred to it as a "strong" balance of probability both because, short of certainty, 70% is not far short of the highest that any assessment of this kind is capable of going, given the intrinsic vicissitudes of life and litigation; and also because my personal assessment (for what it is worth) of the strength of the evidence in favour of and against an 18^th century date points very strongly to the 18^th century conclusion. I think that paragraph 180 of the judgment, taken alone, understates the case. The 18^th century date was, I think, very strongly supported by the combined evidence of Mr Bourne, M. Moreira and Professor del Bufalo, against whose experience and opinion there was little or nothing of any weight apart from the metallurgical evidence. I am no more impressed by the sufficiency of Dr Northover's database than was the judge. Dr Northover's own overconfidence in this insufficient material did not assist the other elements of his opinion. The judge's judgment resounds with understatement. Left to myself, I would be more confident perhaps on the evidence as to the 18^th century date than was the judge. This view is not, however, critical to any issue in this appeal.
153. I move to the related issues raised by Christie's appeal against the judge's finding that they were negligent, and Ms Thomson's appeal against the judge's finding that Christie's catalogue entry was not negligent. For reasons which I have largely anticipated, I consider that the judge's conclusion that Christie's were entitled to describe the vases in the catalogue as Louis XV without qualification sits most uncomfortably with his conclusion that nevertheless they were in breach of duty to Ms Thomson. I have no difficulty with Lord Grabiner's submission that Christie's owed Ms Thomson a Hedley Byrne duty beyond any duty arising out of the catalogue alone. Christie's essentially accepted this and the judge essentially so found. The nature of the duty does not require sophisticated analysis or justification. Nor is there a necessary logical difficulty with Christie's being in breach of the Hedley Byrne duty, but not of a catalogue duty. The duty which Christie's assumed to Ms Thomson was, as I have indicated, potentially in conflict with their duty to Lord Cholmondeley and perhaps any duty relating to the catalogue. The difficulty lies with the facts which the judge found relating to the catalogue and the obligation which Christie's through Mr Bourne accepted in relation to their catalogue obligation.
154. Accepting, as I do, that Christie's owed Ms Thomson a single duty, I nevertheless take the catalogue first, as did the judge.
155. As to the catalogue, there is no proper criticism of the judge's analysis of the standard of care in paragraphs 186-188 of his judgment. Applying that standard and accepting the evidence of Mr Bourne, the judge asked himself whether Christie's had or should have had "a real rather than fanciful doubt" as to the catalogue description (paragraph 189). He rejected Ms Thomson's eleven reasons why it was unreasonable for Christie's to have reached their certain or definite opinion. He relied here, as he was entitled to, on Mr Bourne's confident opinion, supported by M. Moreira (paragraph 191-3). He concluded that Christie's dating of the urns as Louis XV without qualification was an opinion which an auctioneer of their standing could reasonably reach and that there was no breach of duty in that regard.
156. An amalgam of paragraphs 189-194 produces a decision that Christie's dating of the vases as 18^th century was certain and definite, without any real rather than fanciful doubt, and that it was not negligent. If they had had any real, rather than fanciful doubt, it should have been expressed by some appropriate words in the catalogue. This was in substance Mr Bourne's evidence.
157. If Christie's reasonably held this certain and definite opinion, it is difficult to see how they were obliged to express anything other than confidence to Ms Thomson. They would by definition have been expressing fanciful, unreal doubt about the age of the vases, when they had no real doubt. I do not see that they had any obligation to do so. I shall examine the judge's reasons for finding that Christie's were in breach of duty shortly. But, insofar as they concern the date of the vases, they must either give rise to no more than fanciful and unreal doubt or be inconsistent with the judge's own earlier finding.
158. Ms Thomson repeats in this court, without, I think, improving, the eleven reasons which did not persuade the judge that Christie's catalogue description was negligent. The judge's individual and composite decisions here are not readily amenable to appeal. They depend on and are justified by the opinion of Mr Bourne supported by M. Moreira and Professor del Bufalo, whom the judge heard, and we have not heard. The relevant evidence is systematically catalogued for each of the eleven reasons in Appendix 1 Table 1 of Christie's skeleton in response to the cross-appeal. With the exception of the third reason – electrolytic gilding of the liners – to which I shall return, I do not find any of the eleven reasons persuasive. The important point, however, is that the judge's conclusions were firmly based on evidence which he was entitled to accept. This applies as much to the electrolytic gilding of the liners as to the other ten reasons. In addition to particular points, there was a number of general reasons in support of the judge's conclusion that Christie's unqualified opinion about the date of the vases was professionally sustainable. These included the fact that it was the collective unqualified view of Christie's staff; the exposure which the vases had to the art world generally because of the high profile of the sale, and the fact that Christie's description of the vases did not attract adverse comment; the fact that a number of people bid for the vases at prices well above the estimate; and the clear documentary support for the fact that the design originated in the 18^th century at a date entirely consonant with the experts' view about the date of manufacture. There was also Gillian Wilson's consistent view about the date of the Getty vases. Christie's say that it was bold of Dr Northover to contend that the Getty vases were 19^th century. Ms Thomson's case had originally been that the Getty vases were 18^th century, but that there were material differences between them and the Houghton vases. In fact they are very similar.
159. I have summarised earlier in this judgment the judge's reasons for finding that Christie's were in breach of duty notwithstanding their confident opinion as to the date of the vases. He concluded that Christie's gave what was in the circumstances an incomplete picture and Ms Thomson was entitled to a fuller one. He gave three reasons. After anxious consideration, I have concluded that each of them is insubstantial and at variance with his earlier finding that Christie's were entitled to hold and express a confident unqualified opinion that the vases were made in the 18^th century.
160. The judge's first reason was that Christie's should have told Ms Thomson that the dating of vases such as these was unusually difficult because of the existence of 19^th century copies and imitations. This was contrary to the judge's own earlier finding (paragraph 192(1)) that problems arising from 19^th century copies and imitations did not, if you are cautious, give rise to real, rather than fanciful, doubt and that this was so in this case. The earlier finding seems to me to be intrinsically correct in this case. If Christie's were justifiably confident in their opinion, it could only mean that they had considered and confidently rejected the possibility that these vases might be 19^th century imitations. The existence of obviously 19^th century imitations, such as the Paris vases, does not detract from a justifiably confident opinion that the Houghton vases are 18^th century. The existence in general of 19^th century imitations was acknowledged, but this did not make Christie's experts less than confident in their opinion about the Houghton vases.
161. The judge's second reason was that Ms Thomson should have been told that Christie's were relying very largely upon the exercise of their judgment following visual inspection. This was a tenuous basis for finding negligence for at least three reasons. First, this was well understood, as the judge himself had found in paragraph 188 of his judgment. Second, there was no evidence or probability that Ms Thomson was so naïve that she needed to have this commonplace point made to her. Third, if, as the judge found, Christie's were justifiably confident, the basis for their confidence was not a reason to dilute the confidence.
162. The judge's third reason was that Christie's should have told Ms Thomson that the catalogue inflated what could properly be said about the vases and was likely to give a misleading impression about Christie's knowledge and the sureness of their judgment. Leaving aside catalogue inflation for the moment, this too is an intrinsically tenuous basis for finding negligence. If Christie's were, as the judge had held, justifiably confident (sure) in their judgment, extraneous indications of confidence do not make more apparently sure that which is sure already.
163. The judge dealt with catalogue inflation in paragraphs 195 and 196. He concluded that the three matters to which he referred "provided a feeling of confidence and certainty about the urns which was unjustified". Yet he had in substance concluded only two paragraphs earlier that Christie's feeling of confidence and certainty was justified, being reached without any real, rather than fanciful, doubt. This depended on matters other than those discussed in paragraph 195 and was justified without reference to them.
164. The first of the three matters was the removing of the words "possibly Italian" from the catalogue description of the vases. The judge described this as "an unjustified firming-up of the catalogue description". I find this entirely unpersuasive, when the prose following the description makes the Italian connection entirely clear. The judge accepted, as I would, that an item made by a French craftsman and a French architect in the Parma of the 1760s – a satellite of the French court – could arguably be described as Louis XV. But there was in substance no removal of "possibly Italian" from the catalogue read as a whole.
165. The second of the three matters was that the catalogue description that the Houghton vases were designed by Petitot for the Duke of Parma or one of his courtiers was wholly unjustified. I agree that the evidence did not justify a more specific statement than that the vases were from a design by Petitot, illustrated in a publication of 1764, which refers to Petitot as the principal architect to the Duke of Parma and Petitot worked in Parma. But I am unpersuaded that the difference between this and what was said in the catalogue had any bearing on Christie's justifiable confidence and certainty about the date of the vases. The vases are from a design by Petitot which was published in 1764 with reference to the Duke of Parma. Insofar as this contributed to Christie's opinion, the opinion was not enhanced by the more definite statement in the catalogue. There was no evidence that Ms Thomson was or would have been interested in or influenced by any such difference. As the litigation has amply demonstrated, the important thing was the date, as to which this criticism is immaterial. Further, there was no evidence that this or the earlier "Italian" point were in any way material to the value of the vases.
166. The third of the three matters was the unjustified statement in the Condition Report that the antique porphyry vases were reworked by Petitot, giving the further impression that Petitot had designed the actual vases. I do not consider that this stands differently from, or adds to, the second matter. It did not make Christie's justified confident opinion more confident. Importantly, however, the condition report was not seen by Ms Thomson or referred to her. Mr Cooney, who did see it, already had a justified feeling of confidence. I am quite unpersuaded that, absent this statement in the Condition Report, he would have been less confident or would have advised Ms Thomson any differently.
167. For all these reasons, I am persuaded that the judge's finding that Christie's were in breach of their duty to Ms Thomson cannot stand upon the reasons which the judge himself gave. I have considered whether this conclusion goes against my general view that this is an appeal in which the judge's composite conclusions on oral expert evidence, which he heard and we have not, are not readily amenable to appeal. I consider that in this instance some appellate modification of the judge's decision is both justified and necessary. This is because the judge has, in my view, reached two essentially inconsistent conclusions. In short, if Christie's were, or should have been, less than completely confident about their description of the vases, they should have qualified their catalogue description of them. Either Christie's were not negligent in advising Ms Thomson or they were in breach of duty at least to her in their catalogue description.
168. In these circumstances, I have considered whether the correct appellate decision should be that Christie's were not in breach of duty to Ms Thomson or that they should have qualified their description with an expression such as "probably Louis XV". I have explained why I do not consider that the judge's reasons sustain the second of these. I have also considered anxiously whether the fact that the liners were electrolytically gilded should have raised a sufficient question mark over the dating to require some qualification of the description. This seems to me to be the only one of the eleven reasons which requires further scrutiny.
169. The argument is as follows. The liners were gilded electrolytically. This was obvious. Mr Bourne spotted it, although M. Moreira at first did not. Christie's were negligent in not spotting it. The electrolytic gilding cannot have been done before 1840. It could be an indication that the vases were not made before 1840. But even if all or most of the other indications pointed to an 18^th century date, the later gilding of the liners raised a query and demanded an explanation. No satisfactory explanation was readily forthcoming, as this litigation has amply demonstrated. There thus was a more than fanciful doubt about the dating of the vases. This should have been reflected in Christie's catalogue description.
170. I was at one stage in two minds whether this submission should be accepted. The judge dealt with it very shortly in paragraph 192(3). He said there that Mr Bourne had said that it was obvious, but that even if Christie's had spotted it, they could have concluded that it was a re-gilding, perhaps following the repair to one liner. The judge's conclusion, accepting the evidence of Christie's experts, in particular Mr Bourne, was that this was not a reason to dilute Christie's confident opinion. In other words, it did not raise more than a fanciful doubt.
171. In the end, in contrast with my view of the judge's inconsistent reasons for finding Christie's in breach of duty, I have concluded that it would not be right to disturb the judge's finding here. It would be cherry picking one element of a composite judgment to give it a prominence and weight which the judge did not think it had. Mr Bourne was, after all, quite confident in his dating of the vases and he was well aware, not only of this point, but of all the metallurgical evidence as well. It was open to the judge to accept Mr Bourne's opinion. As I have said, it may have been reaching for the moon to try to fit every piece of relevant evidence into an entirely unproblematic whole.
172. If, as I think, Christie's were not in breach of duty to Ms Thomson, her claim for misrepresentation fails also.
173. For all these reasons, I would dismiss Ms Thomson's cross-appeal, but allow Christie's appeal.

Lord Justice Jonathan Parker:

174. I agree.

Lady Justice Smith:

175. I also agree.