Fabio Perini Spa v LPC Group Plc & Ors

This is an appeal and a cross-appeal against a decision of Floyd J, [2009] EWHC 1929 (Pat), in which he determined a substantial number of issues in a patent dispute. The case centred on two patents, EP (UK) 0481929 and EP (UK) 0699168 (“the 929 Patent” and “the 168 Patent” respectively), both registered in the name of the claimant, Fabio Perini SpA (“Perini”). The patents are concerned with machinery and methods for sealing the tail ends of rolls of paper, such as lavatory rolls and kitchen towel, so that they remain rolled up.

The paper is supplied to such machinery in the form of very large “parent” rolls, some three metres in length and about the same in diameter. These rolls are passed to an unwind stand at the beginning of a converting line of the machinery, from which they are rewound by a rewinder onto a cardboard tube to produce smaller diameter rolls called “logs”. These logs have the same diameter as the finished product, about 100mm in the case of a lavatory roll, but they still have the same axial length as the parent roll.

The logs have loose ends or “tails” where the tissue has been cut. If steps are not taken to seal this end to the log, the end could unwind, thereby interfering with subsequent converting steps. So the logs are passed to a tail sealer, which must be designed to keep pace with the flow of logs from the rewinder: otherwise it will hold up the overall speed of the plant. After the tail has been sealed, the log is cut into smaller axial lengths by a log saw, and the individual pieces are packaged.

Perini is one of the two major entities in the field of paper converting machinery of this type, and it brought the present proceedings against four defendants, claiming that certain machines, sold and marketed under the name Rotoseal, infringed the two patents. The second and third defendants, Paper Converting Machine Company Italia (“PCMC Italia”) and Paper Converting Machine Company Ltd (“PCMC UK”), are members of a group of companies which represents Perini’s major competitor in the field, PCMC Italia being the manufacturer of the Rotoseal machines. The fourth defendant, LPC (UK) Limited (“LPC UK”), is a customer of PCMC, who purchased and used two Rotoseal machines, and the first defendant, LPC Group PLC (“LPC Group”), is its parent company.

The defendants denied that the Rotoseal machines infringed either of the two patents, and contended that each of the two patents was invalid on grounds of lack of novelty, obviousness and insufficiency. The Judge concluded that the 929 Patent was valid and that it was infringed by two of the three Rotoseal machines (“the Original Rotoseal” and “the Proposed Rotoseal”), the 168 Patent was invalid on the ground of obviousness, but similarly infringed, and that PCMC Italia, but not PCMC UK or LPC Group, was liable for the infringement as a result of LPC UK’s infringing use of the Rotoseal machines.

In the notice of appeal, notice of cross-appeal, and respondents’ notices, the parties resurrected most, albeit not all, of the points which were taken below, but, as matters progressed, various of those points fell away. I propose to deal only with those points which were live by the end of the hearing. Given that there is no appeal by Perini against the dismissal of its claim against LPC Group, and that LPC UK takes no point on this appeal other than those taken by PCMC, I shall follow the Judge in referring to PCMC Italia, PCMC UK and LPC UK together as “PCMC”.

The scheme of this judgment is as follows. I shall first set out relevant details of the 929 Patent; I shall then consider issues of interpretation relating to claim 16 of that patent; I shall then turn to the validity of the 929 Patent; I shall next discuss whether the Original or Proposed Rotoseals infringe the 929 Patent. I shall then set out the relevant details of the 168 Patent, and then turn to the issue of its validity. Then there is the question whether PCMC UK and/or PCMC Italia are liable for infringement of the 929 Patent. Finally, there is an issue about costs.

The 929 Patent has a priority date of 17^th October 1990, and much of the summary which follows is taken from the judgment below, [2009] EWHC 1929 (Pat) paragraphs 29-35.

The specification is entitled “Apparatus for glueing the tail of a web to a log formed of the web material”. The specification acknowledges that apparatus “of this general type” is known from US 4,475,974, which describes a machine in which the logs are moved by a chain system between three stations on pairs of bearing rollers supported on a chain conveyor. At the first station, the tail is unwound and laid onto a flat surface; at the second station, glue is applied to the tail by a nozzle which moves longitudinally along the axis of the log; and at the third station, the log is rotated so as to wind the tail back up. The specification states that, while this system can achieve high outputs, the type of glue applicator has “a rather complex and costly structure”. One of the stated objects of the 929 Patent is “to reach high outputs per hour with an extremely simple construction and with a limited number of moveable parts”.

The first paragraph describing the claimed invention (column 1, lines 42-50) begins by referring to “the apparatus of the present invention”, in which “the means for applying the glue includes a dispenser device having means forming an upwardly orientated slit from which the glue is delivered by overflowing.” The paragraph then goes on to explain that “[t]he log transferring means discharges said log, with the tail unwound therefrom, causing it to roll over the slit while said tail is unwrapped from the log. …”. The specification then turns to describe various “particularly” or “distinctively advantageous” embodiments of the apparatus.

During the course of describing the first of those embodiments, the specification indicates that:

“The glueing occurs ... extremely swiftly and without the need for dispensing nozzles moving longitudinally along the log as in prior art machines. In fact by the simple rolling of the log over the slit, the glue is applied along the entire length of the log. Gravity, which causes the rolling of the log, causes the subsequent re-wrapping of the tail. … To transfer the log … a flexible conveyor may be provided, which includes a bearing means [which] may be made … to discharge the log … onto an inclined table which has an opening aligned with the slit of the glue dispenser” (column 2 lines 4-18).

When discussing another preferred embodiment, the specification teaches that:

“This type of supply has the advantage of not requiring a controlled dispensing means, such as a pump and nozzle and, moreover, it allows a continuous and easily adjustable flow for glue from the reservoir to the slit, while the excess of glue overflowing from the slit, which is not picked up by the logs rolling over the same slit, is collected in the underlying tank from which it can be easily recovered and made to flow back into the upper reservoir. This avoids the drawbacks exhibited by the nozzle-type dispensing means which, when the apparatus is stopped even for short periods, are clogged up by the drying of the glue, thus making it difficult to re-start the apparatus.” (column 2 line 53 to column 3 line 10).

Having described various preferred embodiments, the specification (column 3, lines 25-27) goes on to explain that “[t]he invention also refers to a method for glueing the tail of a log formed of web material onto the log itself, according to claim 16.” After mentioning that there are also method claims relating to “[f]urther advantageous features”, the specification explains the invention by reference to figures, which are said to represent a “feasible, not limiting, embodiment of the … invention” (column 3, lines 28-34).

Figure 1 shows the path of the log marked with a broad dashed line:

The log is delivered onto rolls 9 and 11. The tail is then unwound onto surface 15, and conveyed to the gluing area (region 45) where it is discharged from the vertical conveyor and rolls across a guide surface. During rolling, the log passes over a slit from which glue is supplied from underneath. As the log continues to roll, the tail is automatically rewound over the log and pressed against the log where the glue has been applied.

A more detailed view of the gluing area is shown in Figure 3, where the tail is marked with the letter L:

The two claims which are said by Perini to be infringed by the Rotoseal machines are claims 16 and 17 of the 929 Patent.

Claim 16 is in these terms:

“Method for glueing the outer end (L) of a web material to a log formed of said material (R), wherein:

the outer end (L) is unwound by a predetermined extent;

the glue is applied on a region of the web material which is still wound up on the log (R);

and the outer end is rewound on the log and glued thereon;

characterized in that

the glue is applied by rolling said log, with the outer end (L) being unwound therefrom, over a slit (63) from which the glue is dispensed.”

Claim 17 reads:

“Method according to claim 16, characterised in that

the log (R) is made to roll over a surface (65) along which said slit (63) is disposed, by such an extent as to allow the application of the glue and the rewinding of the outer end (L) around the log (R).”

Claims 16 and 17 are method claims. The broadest apparatus claim is claim 1, and, while it is unnecessary to set that claim out in full, it is relevant to record that it has limitations to a slit from which glue is dispensed “by overflowing”, as well as to the conveyor upstream of the glue applicator. Claim 3 is another apparatus claim, which refers to the log being discharged onto an “inclined surface having an aperture in alignment with the slit (63)”.

There are three issues of construction in relation to claim 16 of the 929 Patent. The first and second issues require one to concentrate on the words “a slit (63) from which the glue is dispensed”; the third issue centres round the words “the glue is applied by rolling the said log … over a slit (63)”.

The first issue is whether the Judge was right to reject PCMC’s contention that the expression “a slit” in claim 16 would have been understood by the reader of the 929 Patent (i.e. the notional skilled addressee) to have been limited to an opening formed by a pair of inclined plates forming the tip of a nozzle through which glue flows. The Judge held that the noun “slit” was “an ordinary English word which in context would be taken to mean a long narrow opening”, and should not be read as further qualified – [2009] EWHC 1929 (Pat), paragraph 97.

Subject to contextual indications to the contrary, it is hard to fault the Judge’s definition of the noun “slit”, at least provided the word “long” is read as being in contrast to the word “narrow”. It was suggested by PCMC that it was apparent from the short passage which I have quoted that the Judge had adopted a meaning based on what he thought was the ordinary meaning of the noun “slit”, and took no account of the context. Of course, an acontextual interpretation is quite inappropriate (and, at least in many cases, it is impossible). However, I think that it is clear from the very passage on which PCMC relies, with its express reference to context, that the Judge plainly did not purport to give the expression “a slit” a meaning divorced from its context.

More specifically, it seems to me that the Judge’s approach was quite appropriate. The process of construction has to start somewhere, and when the ultimate issue is the interpretation of a common English word, it is often helpful to begin with its ordinary meaning before one turns to its documentary context and other relevant factors. After all, issues of interpretation (whether arising in connection with patents or any other commercial documents) often require an intracranial iterative process, involving multiple factors, including natural meaning, documentary context, technical considerations, commercial context, and business common sense.

Looking at the immediate documentary context, reading claim 16 on its own does not provide any support for the conclusion that one should give the term “a slit” the rather artificially narrow meaning which PCMC suggests. Indeed, unless there is any other persuasive reason to the contrary, it would seem almost perverse to attribute to the patentee an intention to use the word “slit” in a rather artificially narrow sense. It involves attributing to the patentee a desire to limit unnecessarily the scope of the claim, and, therefore, the extent of his monopoly. (It is fair to say that if PCMC’s insufficiency argument, discussed in the passage at paragraphs 51-55 below, had been correct, that might provide a sound reason for its case on the meaning of “slit”, but, for reasons given in that passage, the insufficiency argument is bad.)

PCMC also argues that if one refers to certain passages in the specification of the 929 Patent, they justify the narrow interpretation of “a slit” for which it contends. The two primary passages on which PCMC relies are quoted in the judgment below at [2009] EWHC 1929 (Pat), paragraphs 51-52, and constitute descriptions of specific apparatus. But the fact that a particular type of slit in specific apparatus is called “a slit” in a document which includes a description of that apparatus, cannot justify the notion that, whenever the document uses the term “a slit”, it must have the same limited and specific meaning. As Perini says in its skeleton argument, just because “the word ‘slit’ is used when describing the gap between the plates of a nozzle … it does not follow that ‘slit’ is therefore used to mean that.”

In any event, it is not at all safe to assume that, when a patent is dealing with a process claim, it uses every expression to mean the same thing as when it is dealing with an apparatus, or product, claim: indeed, in this case, as the Judge said, the specification of the 929 Patent makes clear (at column 3 lines 25 to 27) that the method claims (which of course include claim 16) are distinct from the apparatus claims. The other passages in the specification which are relied on by PCMC are of even less assistance as they are concerned with specific embodiments of the claimed apparatus invention.

It is true that the title of the 929 Patent is “Apparatus for glueing the tail of a web to a log formed of the web material” (emphasis added), and that the bulk of the specification is given over to describing apparatus rather than method. However, the patent not only includes what are indisputably method claims, but, having discussed the claimed apparatus invention, the specification states in terms that “the invention also refers to a method … according to claim 16” (column 3, lines 25-27). At least in this case, the mere fact that the method claims are referred to relatively briefly in the specification seems to me to be no warrant for construing those claims artificially narrowly. The general nature of the technique involved in the method claims had already been set out in the specification, when discussing the apparatus claims.

PCMC also sought to support its contention by pointing to the fact that a reference – to 63 in figure 3 - in the drawings supporting the apparatus claims was included in claim 16. It seems to me that this is far too weak an indication to assist PCMC’s case: one must look at the words of claim 16 – a view supported by Virgin Atlantic Airways Ltd v Premium Aircraft Interiors UK Ltd [2009] EWCA Civ 1062, paragraph 17. The reference to 63 is, in any event, perfectly capable of referring, and more naturally refers, to 63 as an example of a slit – especially as figure 3 is an embodiment, or example, of apparatus claimed by the 929 Patent.

This is the most strongly pressed of PCMC’s points on the interpretation of the 929 Patent. PCMC argues that the words in claim 16 “a slit (63) from which the glue is dispensed” mean that (to quote from its skeleton argument) “the slit must participate in the act of dispensing the glue”, and that the claim therefore would not cover a case where the glue was simply dispensed through the slit. Like the Judge, I do not agree. In the context of the 929 Patent, what the words “from which the glue is dispensed” naturally convey in relation to the role of the slit is that the glue passes through the slit.

It is true that the apparatus claimed, and the embodiments described in the specification, involve the glue being dispensed by overflowing through the slit, but that does not mean that the process claim should be read as limited in this way. The embodiments are, characteristically, described as non-limiting, and “advantageous” or “particularly advantageous”, examples of the claimed invention, which rather underlines the point that it could be a mistake to assume that any feature of such an embodiment, which was not at least an express aspect of the claimed inventive concept in the method claims, was intended to be a limiting feature of those claims.

Further, as already mentioned, having discussed the claimed apparatus invention, the specification states in terms that “the invention also refers to a method … according to claim 16” (column 3, lines 25-27). I have already rejected, in paragraph 28 above, the suggestion that the briefness of this reference assists PCMC’s case. Further, the incorporation of claim 16 in that passage serves to emphasise the absence of any limitation on the glue-dispensing function of the slit in the method identified in the specification, when compared with its teaching in relation to the apparatus it describes, where the glue is always dispensed by overflowing through the slit – see paragraph 27 above.

Indeed, when one looks at the apparatus claims, they seem to me to undermine, rather than to support, PCMC’s contention. Claim 1 refers to “means for applying the glue [which] comprise a dispenser device … which form at least an upwardly orientated slit … from which the glue is dispensed by overflowing”. The nature and function of the slit there described is cut down by having to be a slit which is both “upwardly orientated” and from which “glue [can be] dispensed by overflowing”. No such limitation, either as to orientation or as to function, is imposed on the “slit” in claim 16.

For reasons already given in paragraph 29 above, I reject PCMC’s contention that its case is assisted on this point by the reference in claim 16 to the numbering in the figures of the 929 Patent. As to claim 17, it was originally suggested by PCMC that it had no purpose if the Judge was right on this point, but that contention is no longer pressed – quite rightly, in my opinion.

PCMC makes much of the contrast between the word “slit”, on the one hand, and, on the other hand, “aperture” and “opening”, in the specification and claims of the 929 Patent. However, as Perini argues, the words “aperture” or “opening” are used in connection with the apparatus claims or description, where the glue is applied by surplus overflowing through a slit; there is, in such a context, a need for another space through or to which the overflowing glue can go. Hence the aperture or opening referred to, for instance, in claim 1. Therefore, the reference to an aperture or opening in the 929 Patent in relation to the apparatus is, in my view, of no assistance to PCMC’s case so far as the meaning of the method claims is concerned. Indeed, the mention of an aperture in claim 1 (and of an opening in the column 2 of the specification when discussing the apparatus claims) can be contrasted with the absence of any such reference in relation to the method claims, such as claim 16 (and in the brief discussion of the method claims in column 3 of the specification).

PCMC suggests that, to the notional addressee of the 929 Patent, the inventive concept it identified would have gone no further than the apparatus claims. However, Mr Ward, an engineer called by Perini, who the Judge described as “a very thorough and careful witness” (although not able to assist on common general knowledge), took the view that the specification disclosed a broader method of gluing a log by rolling it over a slit from which glue is dispensed and applied. He described the inventive concept of the 929 Patent as “the applying of adhesive to the tissue by rolling it over glue issuing from or adhering to an applicator that is itself stationary at the instant of application (the familiar and intuitive solution is to bring the applicator, with its load of glue, to a stationary log).” Of course, the interpretation of a patent is ultimately a matter for the court, but it is clear that the Judge agreed with this opinion, and took the same view – see [2009] EWHC 1929 (Pat), paragraph 120, also referred to in paragraph 51 below.

Particularly in the light of that conclusion, and as with the first issue of interpretation, it would seem little short of perverse to attribute to the words in question, “a slit ... from which the glue is dispensed” an artificially narrow meaning – see paragraph 25 above.

The third issue concerns the words “the glue is applied by rolling said log … over a slit”. PCMC contends that this means that it must be the act of rolling over the slit, and only the act of rolling over the slit, by which the glue is applied to the log. If that is right, then, if a particular process involves the log being the passive recipient of the glue, then that process would not be within claim 16 – for instance if the glue is delivered from a moving wire dispenser which passes through the slit. The Judge rejected this argument (at [2009] EWHC 1929 (Pat), paragraph 68), and in my view he was right to do so.

There is nothing in the words of claim 16, nor is there any technical reason, which justifies a distinction being made between any of the possible methods of delivery of the glue. It requires the glue to be delivered to the slit, and the log to roll over the slit at a time when the glue is passing through the slit, but it does not seem to me to go further than that. As already explained, there is no basis for restricting claim 16 to a method where the glue is delivered by overflowing. Nor is there anything in the point that, in the case of a machine where the delivery of the glue is applied by a moving wire raised into a slit, the log is not “rolling … over a slit”, but “over a wire”. The glue in such a case is applied by rolling a log over the slit as well as the wire, but that does not alter the fact that it rolls over the slit.

Again, as with the other issues of interpretation, particularly in the light of the Judge’s agreement with the evidence referred to in paragraph 36 above, it appears to me somewhat unrealistic to conclude that the patentee intended an artificially narrow interpretation of the expression “rolling said log … over a slit”, when it neither accords with the natural meaning of those words in claim 16 and there is no technical reason for such a narrow meaning.

I therefore conclude that the Judge was right in his views that in claim 16 neither “slit” nor “rolling said log … over a slit” had the relatively narrow meaning for which PCMC contends, and that the slit need not participate in the glue-application process.

At trial, there were three types of machine which were said to infringe, the Original Rotoseal, the Modified Rotoseal, and the Proposed Rotoseal. The Judge held that the Original and Proposed Rotoseals infringed the 929 Patent, and PCMC appeal in relation to both machines. The Judge concluded that the Modified Rotoseal did not infringe, and Perini do not appeal against that finding, so I shall say no more about it.

The Original Rotoseal was described by the Judge at [2009] EWHC 1929 (Pat), paragraphs 58-64. It is unnecessary to repeat that description: it is sufficient to concentrate on the grounds on which PCMC contends that it does not infringe the 929 Patent, which were identified by the Judge at [2009] EWHC 1929 (Pat), paragraph 65, and rejected in the ensuing three paragraphs.

The essential feature of the Original Rotoseal relied on by PCMC in this connection is that involves the glue being dispensed, and applied to the log, by a wire that moves up from an open-topped glue pan to the log, passing through an opening in the table. The first point made by PCMC is that the Original Rotoseal does not fall within claim 16, as, although the opening referred to is a slit in the ordinary sense of that word, it is not a slit within the meaning of claim 16, as it is not an opening in a glue dispenser formed between two plates. This argument is the same as the first issue of interpretation of claim 16 raised by PCMC, and, for the reasons already given, I would reject it.

PCMC’s second argument is that, even if the Original Rotoseal has a slit, it is not a slit “from which glue is dispensed” within the meaning of claim 16: it is from the wire that the glue is dispensed. That is really a repetition of PCMC’s case on the second issue of interpretation, which I have rejected. The fact that the glue is applied by a wire does not prevent it from being dispensed from a slit, if the wire passes through the slit.

Finally PCMC contends that in the Original Rotoseal, glue is not “applied by rolling the … log over [a] slit”, as it is not the act of rolling which applies the glue to the log. In this connection, PCMC point out that, if the wire did not move, no glue would be applied to the roll. But, as the Judge said at [2009] EWHC 1929 (Pat), paragraph 68, the fact remains that “[a]t the moment the log rolls over the slit the wire is in the uppermost position, so that there is glue present in the slit”. This is really the same point as PCMC’s third argument on interpretation, which, like the Judge, I have rejected. The log rolls over the slit in the Original Rotoseal, and while it is rolling over the slit, the glue is applied through the slit.

Accordingly, in agreement with the Judge, I consider that the Original Rotoseal infringes claims 16 and 17.

Again, it is unnecessary to explain this apparatus which the Judge described at [2009] EWHC 1929 (Pat), paragraphs 75-76; in the next two paragraphs, he went on to hold that it infringed claims 16 and 17.

PCMC contends that the Proposed Rotoseal does not infringe on the same three grounds as it relies on in relation to the Original Rotoseal, together with another ground. I would reject the same three grounds on the reasons just discussed. PCMC had a fourth ground, which did not apply to the Original Rotoseal, but it was abandoned during the hearing of the appeal.

Accordingly, I agree with the Judge that the Original Rotoseal and the Proposed Rotoseal both infringe claims 16 and 17 of the 929 Patent.

At the trial, PCMC contended that claims 16 and 17 of the 929 Patent, as interpreted by the Judge, were too broad, as they did not include a requirement that the glue is dispensed by overflowing, and hence the patent was insufficient. The Judge rejected that contention on the ground that the notional addressee of the patent would understand that the disclosure claimed by the patent, in particular by the process (or method) claims, including claims 16 and 17, was the idea of applying glue to a log while the log was rolling over the slit through which the glue was applied, and that, accordingly, the precise means by which the glue was to be applied through the slit was irrelevant – [2009] EWHC 1929 (Pat), paragraph 120.

PCMC’s contention was that this was wrong, because the 929 Patent only disclosed one method of applying glue through a slit, namely by overflowing, and, in particular it did not disclose the possibility of applying glue using a moving wire, which, as the Judge also found, was not part of the common general knowledge at the relevant priority date. In other words, PCMC relied on the propositions that “the specification [of a patent] must enable the invention to be performed to the full extent of the monopoly claimed”, and that a claim cannot cover ways of delivering the goods “which owe nothing to the teaching of the patent” - per Lord Hoffmann In Biogen v Medeva [1997] RPC 1, 48 and 50.

The Judge, however, said that, even on the basis of the reasoning of the House of Lords in Biogen [1997] RPC 1, the 929 Patent revealed “a principle capable of general application”, so that “the claims [such as claims 16 and 17] may be in correspondingly general terms”.

It is unnecessary to consider this argument in any detail: while PCMC does not, as I understand it, formally abandon its insufficiency argument, it realistically now accepts that, if the Judge’s conclusions on the issues of interpretation of the 929 Patent were correct (as I have concluded that they were), then its insufficiency argument is difficult to sustain. In my opinion, the Judge was right to reject the argument that the 929 Patent was insufficient.

Indeed, I incline to the view that, as Perini submits, the argument based on Biogen [1997] RPC 1 is even weaker than the discussion may have so far suggested. In that connection, it is, I think, unnecessary to do more than say that it appears to me that the answer to PCMC’s insufficiency argument is very well encapsulated in the following observation of Pumfrey J in Minnesota Mining and Manufacturing Company v ATI Atlas Ltd [2001] RPC 523, paragraph 62:

“when Lord Hoffmann speaks about the claim covering ways of delivering the goods which owe nothing to the teaching of the patent ..., he is talking about a claim which is not supported by the specification since it claims ... an obvious desideratum. He is not talking about a case in which there are ways of arriving at a device within the claim which themselves involved invention.”

PCMC argue that the Judge was wrong to reject its contention that the 929 Patent was invalid on the ground that it was obvious over a Japanese patent, the 562 Patent. (PCMC originally argued that the 562 Patent anticipated the 929 Patent, but the Judge also rejected that contention, and there is, rightly in my view, no appeal against that.)

The Judge described the 562 Patent at [2009] EWHC 1929 (Pat) paragraphs 83-86 in these terms:

“The 562 Patent was published in 1975 … . It describes a tail sealer. Figure 3 shows nearly everything one needs to know:

Having heard expert evidence from Mr Buxton for PCMC, and Mr Ward for Perini, the Judge held that the 929 Patent was not obvious over the 562 Patent. At least on the face of it, PCMC’s appeal against the Judge’s rejection of its case that the 562 Patent rendered the 929 Patent obvious faces an uphill task. An appellate court is reluctant to interfere with a trial judge’s finding on obviousness unless it can be shown that the judge went wrong in principle – see per Lord Hoffmann in Biogen [1997] RPC 1, 45. All the more so where, as here, the Judge appears to have expressly directed himself properly as to the law on obviousness – see [2009] EWHC 1929 (Pat), paragraphs 102-104.

PCMC’s contention, however, is that the Judge did go wrong in principle, essentially in three respects. In this connection, the argument before the Judge had centred, in the usual way, round the issue of whether it would have been obvious to the notional skilled person, as at the relevant priority date, in this case 17 October 1990, to take certain steps to adapt or modify the item of prior art in question, here the 562 Patent, so as to arrive at what is claimed by the invention of the patent in suit, namely the 929 Patent. Two specific errors are attributed to the Judge, namely that he wrongly concluded that it was not obvious to take a certain step, and, that, in any event, he wrongly held that a number of steps were required to justify the obviousness claim, and that the totality of those steps was not obvious. More generally, it is said that the Judge wrongly rejected the view of both experts that the 929 Patent was obvious over the 562 Patent.

Attractive though this line of argument seemed on reading PCMC’s skeleton argument and listening to the oral submissions made on its behalf, I consider that, properly analysed, the Judge reached a conclusion which was, to put it at its lowest, justified in the light of the evidence.

In general terms, PCMC’s case on obviousness suffers from two familiar defects (or, probably, two aspects of the same defect) identified by the Judge at [2009] EWHC 1929 (Pat), paragraphs 113. PCMC’s case is based (a) on hindsight (assuming the end, the invention claimed by the patent in suit, and then working out how to get there from the prior art) and, this being a case where there are a number of steps to get from the prior art to the patent in suit, (b) the assumption that, because each step is obvious over its predecessor, then the totality of the steps must be obvious.

In this connection, it is salutary to have in mind the well known remarks of Lord Diplock in Technograph Printed Circuits Ltd v Mills & Rockley (Electronics) Ltd [1972] RPC 346, 362, lines 34-47. As he said, “[o]nce an invention has been made, it is generally possible to postulate a combination of steps by which the inventor might have arrived at the invention that he claims”, and, in many cases, “[i]t may be that taken in isolation none of the steps, ... if taken in isolation, appears to call for any inventive ingenuity”. However, as he went on to explain, even in the “improbable” event that this was how the inventor made the claimed invention, it is often the case that the “inventive ingenuity [lies] in perceiving the final result ... was attainable from the particular starting point and in [the inventor’s] selection of the particular combination of steps which would lead to that result.”

In order to explain in more detail why PCMC’s appeal on this issue should, in my opinion, fail, it is convenient to set out the second (and, I think, main) reason the Judge rejected Mr Buxton’s evidence that no ingenuity would have been needed to move from the disclosure of the 562 Patent to the alleged invention of the 929 Patent. At [2009] EWHC 1929 (Pat), paragraph 113, having said that the move would have involved a number of steps, the Judge explained:

“Firstly, the 562 Patent is a compact design. The decision to elongate it by the addition of a downstream gluing station involves a rejection of that purpose. Secondly, deprived of the glue applicator, and without further adjustment the valley arrangement becomes an unwinding and winding up station. It delivers a completed wound up roll to the output conveyor. To turn it into an unwinding station alone is quite a major change to its function. Thirdly, the skilled person has to be able to see that a slit in the rolling surface, downstream of the unwinding station can be made to apply glue and seal the log when the log and its trailing tail are in a different configuration from that shown in the unmodified diagrams. Fourthly, the skilled person has to appreciate how to get a much longer tail than that shown in those diagrams. Finally, even then the skilled person would not necessarily come within the claims, because the downstream glue applicator that is designed would be more likely to stop the log before glue is applied, just as in the original proposal. A glue applicator which applied the glue "by rolling… over the slit" would not necessarily result.”

The Judge’s first reason effectively centred on the fourth of those steps: it involved appreciating that the rollers in the machine described in the 562 Patent would have to be stopped and reversed in order to ensure that there was an adequate tail length. Mr Buxton was taken to task for not having referred to this in his first report, and his answer was, in effect, that reversing the rollers was an obvious modification to deal with an obvious problem. The Judge was unconvinced by Mr Buxton’s evidence on this topic, and effectively held at [2009] EWHC 1929 (Pat), paragraph 112 that reversing the rollers would not have been obvious. This conclusion was in part based on his rejection of Mr Buxton’s evidence that the disclosure of another patent published in 1971, known variously as the Crown Zellerbach patent or the Janik patent, was part of the common general knowledge as at 1990.

On this first point, it was suggested that the Judge should not have found that Mr Buxton relied on this patent to support his view that reversing the rollers would be obvious, but I do not agree. Mr Buxton referred to reversing the rollers as being “as was prior art in the Janik patent” (Day 4, transcript page 521), and went on to say (on the next page) that the glue dispensing system in the 929 Patent was not in the 562 Patent, but “it was common general knowledge”. The day before (transcript page 380), he had referred to “common general knowledge through the [Janik] patent”.

In principle, a judge is entitled to reject an expert witness’s testimony that a particular step was obvious, or that the disclosure of an earlier patent was part of the common general knowledge. In the present case, that is reinforced by two factors. First, as Perini points out, the 562 Patent (on page 7 of the translated version) contained its own teaching as to how to find the tail of a roll, although it is one which does not work for PCMC’s purposes in this case. Secondly, Mr Buxton, the witness whose evidence is primarily relied on by PCMC, had, on the Judge’s finding at [2009] EWHC 1929 (Pat), paragraph 111, failed to consider, or at least “shelved”, the question of what to do about the problem of machinery working in accordance with the 562 Patent not producing an adequate tail length.

As to the Judge’s second reason, the conclusion at the end of the passage quoted in paragraph 63 above, it is, at least on the face of it, a classic example of the application of Lord Diplock’s analysis in Technograph [1972] RPC 346, 362, and, subject to the question of whether there was some basis for the conclusion, it would appear to be unassailable. PCMC argues, however, that, with the exception of the fourth (which was said to be obvious, as just discussed) the five steps identified by the Judge in that passage were illusory.

In my view, although the first step was a very minor factor, and played hardly any part in the evidence, it was true, and therefore a fair point as far as it went, which, on its own, I accept, would not have been very far. As to the second step, it does involve redesigning the machine shown in the 562 Patent, effectively separating out the various functioning parts. The third step involves changing the glue dispensing system shown in the 562 Patent from one which is located between the two rolls to one that is a gap (whether or not a slit) in the table. It thus involves putting downstream of the gluing position in the 562 Patent what is in effect the inventive concept of the 929 Patent. The fourth step I have already discussed. As to the fifth step, it does appear to me to involve a problem which requires to be solved, although it is not the most significant of the five steps.

Accordingly, unless PCMC is right in contending that Mr Ward accepted that the 929 Patent was obvious over the 562 Patent, the attack on the Judge’s finding of obviousness fails on two separate grounds, namely that the fourth step identified in the passage quoted in paragraph 63 above was not obvious, and that, in any event, the five steps identified in that passage were, taken together, not obvious. (In fact, as Perini submitted, there appears to me to be a further reason for this conclusion, namely that the third step was not obvious, given the Judge’s finding on common general knowledge, but it is unnecessary to expand on that.)

That leaves PCMC’s contention that Mr Ward accepted its case on obviousness. It is unnecessary to decide whether that would be determinative of the issue (although it is right to record that I would take some persuading that it would be), because, properly analysed, Mr Ward’s evidence did not support PCMC’s case on obviousness. He said in cross-examination that there was “something inventive about the idea of a machine that changes speed and having a machine that reverses its rotation” (Day 1, transcript page 75). A little later, he said that “if I were going to build a tailsealer I would not start from [the 562 Patent]”, and “if I were a designer in 1990, I would not even have looked at [the 562] patent ... as a source of inspiration” (transcript pages 117-118). Insofar as there are some passages in his evidence which, taken on their own, appear to assist PCMC, they were all based on the proposition that he would have considered the 562 Patent with a view to attempting to convert it to the invention claimed by the 929 Patent – the very exercise whose value was doubted by Lord Diplock.

It seems to me, therefore, that Mr Ward’s evidence was by no means supportive of the notion that the 929 Patent was obvious over the 562 Patent. In my judgment, therefore, once one looks at the expert evidence, and the Judge’s analysis of the contention that the 929 Patent was obvious over the 562 Patent, PCMC’s appeal on obviousness is doomed to failure. Despite PCMC’s suggestion to the contrary, it is attempting to persuade an appellate court to overturn a trial judge’s conclusion on obviousness, which was, at best from the appellants’ point of view, a conclusion which the Judge was entitled to reach. I should perhaps add that, at least on the arguments we have heard and the evidence we have been taken to on this appeal, the Judge’s conclusion is one which appears to me to have been right.

The final issue relating to the 929 Patent is whether the Judge was right to reject PCMC’s case that the invention it claimed was obvious over an earlier patent registered by Perini, the 445 Patent, the equivalent of FR-A-2, 284, 548 referred to as part of the prior art in the specification of the 929 Patent.

I take the description of the 445 Patent from the judgment – [2009] EWHC 1929 (Pat), paragraphs 91-96:

“[The 445 patent] was published in 1977. Figure 1 shows the machine as a whole:

In its skeleton argument for this appeal, PCMC maintained its case that the 445 Patent was novelty-destroying, but that submission is now (realistically, in my view) abandoned. Having correctly directing himself as to the law on anticipation at [2009] EWHC 1929 (Pat), paragraphs 80-81, the Judge accepted Perini’s argument that the gap between the belts was not a slit according to claim 16, as its dimensions are not sufficiently described, and because there was no evidence from PCMC as to the size of the relative dimensions of the opening between the belts.

However, PCMC still contends that the 929 Patent is obvious over the 445 Patent, albeit on a different ground from that raised below. The obviousness argument runs as follows. While accepting that the Judge was right in holding that the 445 Patent did not anticipate the 929 Patent, as it contained no clear directions as to the size of the gaps between the belts or as to the number of belts, PCMC contend that, as there are no such directions, it would have been open to the reader of the 445 Patent to make a machine in accordance with its teaching with as many, or as few, belts as he wishes, and one obvious design would involve a narrow gap, i.e. a slit.

I am unimpressed with this argument. A contention that it would be obvious to implement the teaching of a patent in a certain way stands little chance of success if it is raised for the first time on an appeal. First, it is scarcely likely to be obvious if it did not occur to the party who so suggests for the first time, after a full trial, in which obviousness points were raised. Secondly, in the absence of any evidence from the experts, or any finding from the trial judge, on the issue, it is difficult to persuade an appeal court that it is appropriate to uphold an obviousness argument.

More specifically, as Jackson LJ said in argument, the 445 Patent’s discontinuous aperture, broken by a series of belts is very different in nature from the slit envisaged by the 929 Patent. And, as Hughes LJ observed, the development postulated by PCMC’s argument would involve abandoning the notion, taught by the 445 Patent, that the brushes should poke through the belts, and replacing it with the introduction of a slit. Given those points, it appears to me to be impossible for this court seriously to entertain the submission that the 929 Patent was obvious over the 445 Patent, given that these issues were not explored with the expert witnesses below. I might add that, in the light of the Judge’s reasoning in relation to the 562 Patent, it appears to me very likely that he would have rejected the obviousness argument based on the 445 Patent, had it been raised at trial: at least on a superficial impression, it appears to suffer from similar problems.

In these circumstances, I reach the conclusion that the Judge was right to reject the attacks on the validity of the 929 Patent.

The 168 Patent relates to an improvement over the invention of the 929 Patent. It has a priority date of 10^th December 1993. The summary which follows is taken virtually verbatim from the judgment below – [2009] EWHC 1929 (Pat), paragraphs 121-124.

Like the 929 Patent, the 168 Patent has apparatus and method claims. The improvement over the 929 Patent is a modified glue dispenser. Although there are others, the simplest form described is that shown in Figure 4:

The rolling surface has a glue container disposed below it. The upper wall of the container has a slit through which glue is dispensed by means of a moving member (181). The member (181) moves from a position where it is immersed in glue to a position where its extremity (157) emerges through the slit in the upper wall, which also forms the rolling surface for the logs. The motion is so arranged that the member is in the upper position when the log rolls across the slit, so that the glue adhering to the member is picked up by the log.

Claim 1 reads as follows:

“Device for gluing the tail end (LF) of a reel (L) of wound web material, comprising:

conveying means (5, 21) for moving the reel;

unwinding means (7, 9) for unwinding the tail end (LF) of the web material when said reel is in an unwinding position;

a dispenser (11) of glue (C) for applying the glue to the reel (L), arranged downstream of said unwinding position; and

means (21) for rewinding the tail end after the glue (C) has been applied;

said dispenser (11) including an upwardly oriented slit (155) from which the glue is dispensed;

characterized in that

said dispenser (11) includes a container (151) for the glue with said upwardly oriented slit

and a moving member (157) positioned inside said container, which is immersed in the glue contained in the container (151)

and moved towards said upwardly oriented slit (155) in order to dispense the glue to the reel as it rolls over said slit”.

The Judge accepted that the 168 Patent was invalid on the ground of obviousness over a disclosure which he concluded was made to an employee of PCMC UK, Mr Sigmund Casper, which he recorded in a memorandum (“the Casper memorandum”) prepared following a visit to Kaysersberg SA in Kunheim, a French paper manufacturer in 1992. Perini cross-appeals against that finding.

The relevant part of the Casper memorandum was in these terms:

“Perini has also developed two new tail sealing systems which are known as models 4630 and 4640.

Model 4630 is guaranteed at 30 logs per minute but to date, has only run in production at 28 logs per minute. However, model 4640 has been demonstrated at 40 logs per minute but it is unknown at this time how it is performing under production conditions.

This new type of tail sealer is reported not to have any moving parts and by that I believe is meant no traversing glue nozzle.

The basic operation of this unit is through the use of a glue bar which is equal to the log length. The glue bar sits underneath the log roll-off table which is slotted. When the log rolls down the table, it is momentarily halted over the slot at which time the glue bar moves up through the slot to deposit a line of adhesive along the whole log length in one small upward movement.”

The Judge concluded, at [2009] EWHC 1929 (Pat), paragraphs 153-160, that (i) the Casper memorandum was a genuine record of what Mr Casper was told in 1992; (ii) the disclosure which he recorded was not made to him on a confidential basis; (iii) the relevantly skilled person would have understood the memorandum to describe a machine which was within the claims of the 168 Patent, save that log stopped, rather than continued rolling, while the glue was applied to it through the slit; but (iv) it would have been obvious to the relevantly skilled person to develop the machine so that the log continued to roll over the slit. Accordingly, he concluded that the 168 Patent was obvious over the Casper memorandum.

Perini have not sought to challenge conclusions (i) and (ii), but they appealed against conclusions (iii) and (iv), although no oral argument was advanced to support its case on conclusion (iii). However, before turning to conclusion (iv), I shall deal with it, as it was not formally abandoned.

On conclusion (iii), the Judge held that the Casper memorandum “clearly discloses the use of a moving bar applying glue through the gap in the table to a log as it rolls down a roll off table” – [2009] EWHC 1929 (Pat), paragraph 155. Perini’s contention is that the memorandum, properly understood, was no more than a description of the machines then marketed by Perini in accordance with the teaching of the 929 Patent, known as the 65.50 and 560C machines, which had an overflowing glue nozzle, and that Mr Casper, viewing the machine from the outside, misinterpreted the way in which the machine functioned.

The contention that Mr Casper was confused is said to be supported by the acceptance by Mr Buxton (Perini’s expert witness) in cross-examination that PCMC, and Mr Buxton himself, understood the Casper memorandum to provide a “mangled” description of the 65.50 or 560 C machines, and the fact that PCMC never even attempted to design or manufacture a machine in accordance with what, according to the Judge’s reading, was the teaching of the memorandum.

The interpretation of the Casper memorandum is, of course, ultimately a matter for the judge, not for witnesses. A judge can rarely, if ever, determine such an issue simply by reference to evidence as to how a witness or a party understood it, although such evidence may be admissible and helpful in relation to the meaning of any technical language used: see per Lord Reid in van der Lely v Bamfords [1963] RPC 61, 71. The fact that PCMC may have understood the Casper memorandum to mean what Perini now contend is not enough to justify reversing the Judge’s decision to the contrary, if there were good grounds for that decision. In any event, two other witnesses, one of whom was Perini’s expert, Mr Ward, accepted PCMC’s interpretation of the memorandum, namely that “the log rolls down the table, over the slot, a glue bar comes up from underneath, touches the roll and applies the glue” (cross-examination of Mr Ward).

The Judge accorded the memorandum its natural meaning on the point at issue; indeed, the last paragraph (and, in particular, the last sentence) of the extract quoted in paragraph 84 above, is clear on the point. Of course, it is possible that Mr Casper had wrongly understood, and therefore inaccurately described, what he had been told, however, as the Judge said, the memorandum is a genuine record of what Mr Casper “must have been told” – [2009] EWHC 1929 (Pat), paragraph 153. I also note that it was not put to Mr Casper, when he gave evidence, that his memorandum contained an inept or inaccurate description of an overflowing glue nozzle, rather than of a glue application bar. Quite apart from this, I am by no means satisfied that Mr Buxton’s evidence, read as a whole, did amount to an acceptance of Perini’s case on this point. It is unnecessary, and therefore inappropriate, to add to the length of this judgment by going into the details of Mr Buxton’s evidence on this issue: even if he did make the concession attributed to him by Perini, it would not, in the light of the points made in the preceding paragraph of this judgment, be nearly enough to justify overturning the Judge’s conclusion (iii).

As to the fact that PCMC did not manufacture a machine in accordance with Mr Casper’s description, as interpreted by the Judge, that is merely arguable corroboration of the contention that PCMC, and in particular Mr Buxton, understood the memorandum in the way that Perini now contends, a point already considered.

I turn to conclusion (iv). As already mentioned, an appellate court is very slow to overturn a finding of obviousness. Perini’s case is that the Casper memorandum describes a machine with a moving bar which applied the glue to a log which had temporarily ceased rolling, and that it contains no teaching which would lead a person to develop the moving bar notion to a machine while the log continued to roll.

It seems to me that the Judge’s finding of obviousness was fully justified. Perini effectively contend that the Judge wrongly accepted that it was obvious to apply the idea which was disclosed in the Casper memorandum to the current Perini machines, such as the 65.50 and 560C machines, which functioned in accordance with the 929 Patent.

There was expert evidence that the relevantly skilled person would have thought the disclosure of the 168 Patent was obvious over what was revealed in the Casper memorandum. In his reports, Mr Buxton said that, as at 1992, the obvious way to implement the disclosure in the memorandum would have been to take the latest Perini machines on the market, namely the 65.50 and 560C machines, and replace the overflowing glue nozzle with the glue dispensing bar described in the memorandum; he was not challenged on this. In this connection, it is also germane to mention that the memorandum refers to the fact that it was “Perini” who “had ... developed two new tail sealing systems”. This would appear to lead the relevantly skilled person at least to consider that Perini had modified one of its existing machines, which, of course, involved applying glue to continuously rolling, not temporarily stationary, logs.

In cross-examination, Mr Ward accepted that it was “a logical extension” of his evidence that he “would appreciate” that one could “simply take the glue fountain away and replace it with the glue bar”. He also said in cross-examination that, if one was “familiar with the 560C”, the only difficulty in the notion that one would have considered that “the gluing could be achieved simply by rolling [the log] without stopping” was “the question of control and understanding whether the skilled person would have sufficient command of control systems in order to design that without invention”.

Having quoted the passage in the evidence of Mr Ward which contained that observation, the Judge said that “If there is some difficulty in arranging the control system, that is not something the 168 Patent gives any help on: rather it assumes that the skilled person has the necessary skills, as I am sure he or she would. ... In my judgment the Casper memorandum renders [claim 1] obvious” – [2009] EWHC 1929 (Pat), paragraphs 159-160. On the face of it at least, in the light of the evidence to which I have just referred, that was a view which, to put it its lowest, the Judge was entitled to reach.

The main point which Perini makes is that the Judge reached his conclusion purely on the basis of Mr Ward’s evidence, and that was impermissible, because Mr Ward had taken the view, which the Judge had rejected, that there was no significant difference between what one might call a stopping system and a non-stopping system. There are two answers to that point.

First, although the Judge only referred to Mr Ward’s evidence in this connection, he did not say that it was the only evidence he relied on to support his conclusion on obviousness. This is important, because, as Perini accepts, and as I have mentioned, there was other evidence to support the Judge’s conclusion; the Judge no doubt was impressed with the fact that Perini’s own expert, Mr Ward, supported PCMC’s obviousness argument, although Mr Buxton also did so. So he simply quoted Mr Ward’s evidence, and then gave his own view. In a perfect world, the Judge might have referred to Mr Buxton’s evidence also, but, where both expert witnesses agree, it is scarcely surprising if the Judge specifically cites the expert who is called by the party who is supporting the opposite view.

Secondly, Mr Ward’s evidence was that there was no “clear functional distinction” between stopping and rolling systems. But, as Hughes LJ pointed out in argument, if (as the Judge concluded) there was a significant distinction between stopping and rolling systems, this would not necessarily help, and could not possibly be determinative, on the issue of whether it was obvious to develop from one to the other. Mr Ward’s evidence that there was no functional difference may have provided the basis for a fertile cross-examination (which did not happen), but it is not enough to render valueless his view on obviousness – particularly when it is supported by other evidence.

In these circumstances, the Judge’s conclusion that the 168 Patent was invalid on the ground that it was obvious over the Casper memorandum was right. There were a number of other issues raised by the parties on the 168 Patent, relating to interpretation, infringement, and obviousness over other patents. However, in the light of the fact that the 168 Patent is invalid over the Casper memorandum, the parties are agreed that none of the other issues raised in relation to the 168 Patent need be considered.

Two original Rotoseals (“the Rotoseals”) were purchased ex works from PCMC Italia’s Italian factory by LPC UK, who then imported them into the United Kingdom, where they were installed and used for over a year in LPC UK’s premises. Accordingly, there could be no question but that LPC UK was liable for infringement of the 929 Patent. The Judge found that PCMC Italia was jointly liable for the tort, but that LPC Group and PCMC UK were not. PCMC Italia appeals against the finding that it was so liable, and Perini cross-appeals against (a) the finding that PCMC was not also primarily liable for infringement, and (b) the finding that PCMC UK was not liable at all.

PCMC Italia not only installed the Rotoseals at LPC’s premises, but, as the Judge inferred from the terms of its contract with LPC UK, “there was use of the Rotoseal machines at start up under the close supervision of PCMC Italia’s employees”. Having pointed out that the infringed claims were method claims, the Judge went on to say that it was “clear from the contract that the machine supplied was to be one which operated in accordance with the patented method” and that PCMC Italia had “built the machine on LPC Limited's premises and caused it to work in accordance with the method claims”. As this was “done pursuant to a common design, evidenced by the contract between the parties”, the Judge concluded that “by installing the Rotoseal and causing it to operate according to the patentee's method in [LPC’s] premises, PCMC Italia have acted in furtherance of that design by building the machine on LPC Limited's premises and ensuring that it runs in accordance with the method. PCMC Italia are jointly liable with LPC Limited for infringement of the 929 patent” – [2009] EWHC 1929 (Pat), paragraph 179.

It is common ground, as the Judge accepted (at [2009] EWHC 1929 (Pat), paragraph 174) that if PCMC Italia had merely sold, exported and installed the Rotoseals in LPC’s premises, that would have been insufficient to fix it with liability. However, the Judge’s findings go further than this. The contract with LPC UK included obligations on PCMC Italia (a) to instal the Rotoseals, (b) to start up the machines, (c) to ensure that the machines were running properly, and (d) to train LPC personnel as to their use, two PCMC technicians being allocated to each of the tasks for two weeks. Having, unexceptionally, found that these obligations (“the services”) were duly performed by PCMC Italia, the questions are whether the Judge was right (a) to hold that this was sufficient to fix PCMC Italia with joint liability with LPC UK for infringement of method claims 16 and 17 of the 929 Patent, and (b) to refuse, in addition, to find PCMC primarily liable for infringement, on the ground that such liability was not pleaded by Perini.

So far as joint liability is concerned, both Perini and PCMC refer to a decision of this court (which went to the House of Lords, but not on the point at issue here), Sabaf v Meneghetti and MFI [2003] RPC 14, where, at paragraphs 58-59, the Court of Appeal rejected the contention that the supply of infringing goods from abroad into this country was sufficient to fix the supplier with liability even where the supplier “knew that [the goods] were going to be imported into the UK”. The supplier in that case had “merely been acting as a supplier of goods to a purchaser, which was free to do what it wanted with the goods”. Peter Gibson LJ said that, in order to be liable, the alleged joint tortfeasor must have “been so involved in the commission of the tort as to make himself liable for the tort”, and that he must have “made the infringing act his own”. While I agree with the decision, I must confess to finding the reasoning rather circular, which is not surprising as the circumstances in which joint liability arises are difficult, probably impossible, to define fully satisfactorily in abstract.

At least to my mind, the test propounded by Mustill LJ in an earlier patent case, Unilever v Gillette [1989] RPC 583, 608-609, is rather more helpful in the present case. At the end of a brief analysis of the principles (quoted by the Judge at [2009] EWHC 1929 (Pat), paragraph 177), Mustill LJ said that it was “enough if the parties combined to secure the doing of acts which in the event prove to be infringements”. Merely exporting a machine from another country to a third party in the UK, even helping to instal the machine in the third party’s premises in the UK, would not, at least in ordinary circumstances, amount to such an act, as it is the use of the machine (which, in that case at least, was a matter entirely for the third party) which constitutes the tort.

Interestingly, in that passage, Mustill LJ expressed himself on the basis that the issue for an appellate court would not, at least always, be the simple question of whether the judge was right or wrong as a matter of law. He formulated the issue as being whether “a trial judge directing himself correctly could reasonably come to the conclusion that ... there was a common design ... to do acts which ... amounted to infringement, and ... [one party] has acted in furtherance of that design.”

In this case, I think that the Judge was at least entitled to conclude that PCMC Italia’s actions, including as they did the services, went sufficiently far to render it liable as a joint tortfeasor with LPC UK. PCMC Italia did not merely sell the Rotoseals into the UK, knowing that they would be used in a way which would infringe the 929 Patent: their employees were specifically engaged (a) to supervise their assembly at LPC’s premises in the UK, (b) to supervise the starting up of the machines (and not merely for a day or so), and (c) to supervise the training of LPC staff in the use of the machines (again for a number of days). All these services, performed in LPC’s premises in the UK, were plainly undertaken to enable, indeed to assist, even to join in with, LPC’s use of the machines – i.e. to infringe claims 16 and 17 of the 929 Patent.

As mentioned, in Unilever [1989] RPC 583, 608-609, Mustill LJ approved the concept of “common design” as a helpful guide in this area, although he explained that such expressions were “not to be construed as if they form part of a statute”, and emphasised that there was no “need for a common design to infringe”. The Judge, rightly in my opinion, considered that, when seen as part of its obligations under the agreement to sell the machines, the services performed by PCMC Italia, at LPC’s premises in the UK, where the machines were obviously to be used, and were being used, for an infringing process, did amount to engaging in a common design with LPC UK so far as that use was concerned.

Even assuming that, if taken on their own, neither (a) the sale and installation of the machines, nor (b) the provision of the services, would have given rise to liability for infringement of claims 16 and 17 of the 929 Patent, that does not mean that the combination of the two sets of actions could not do so. It is also argued by PCMC Italia that it should not be liable for the approximately two years’ worth of infringing use of the machines undertaken by LPC UK. That may or may not be right: it does not impinge on the question of principle, namely whether PCMC Italia was liable to the extent of its actions under the contract, namely the supply of the machines and the provision of the services. As Perini says, the main purpose of the claim against PCMC Italia was to obtain injunctive relief, and the extent of its liability for damages was not canvassed, and will be a matter for assessment, which will take place in due course. An account of profits was also mentioned: I have real difficulty, in terms of both principle and fairness, with the notion that PCMC Italia should be obliged to account for profits made by LPC UK, but that, again, is not an issue which can fairly be determined at this stage.

In the light of my view upholding the Judge’s conclusion that PCMC Italia was a joint tortfeasor, it is common ground that it is unnecessary for us to decide whether the case that PCMC Italia was primarily liable for infringement was properly pleaded. In those circumstances, it seems to me inappropriate for us to address that issue.

It is only right to add that, both at first instance and in this court, there was only limited discussion about the difficult issues thrown up by the question of PCMC Italia’s liability. Accordingly, my analysis of the points thrown up by the question has been relatively brief.

The Judge rejected Perini’s contention that PCMC UK was liable for infringement of the 929 Patent. Perini’s case was essentially based on PCMC UK’s apparent involvement in the contract under which PCMC Italia agreed to sell the Rotoseals to LPC UK, and to supervise their installation and starting up, and the training of LPC staff, as just discussed. In effect, Perini says that, as the name of PCMC UK and one of its employees, Trefor Hughes, appears in a number of places in the contract, it is for PCMC UK to prove that they were not involved with the acts of infringement in the same way as PCMC Italia, at least to the extent of being liable as a joint tortfeasor.

On the facts of this case, I consider that the Judge was right to reject this argument. The burden of proof was clearly on Perini, so that it was for Perini to come up with evidence which established liability on the part of PCMC UK, not on PCMC UK to adduce evidence to show that it was not liable. Of course, there could come a point where there was sufficient evidence, which could be direct or circumstantial, expressed or implied, written or oral, to discharge that burden, so that, unless PCMC UK had produced convincing evidence in response, its liability would have been established. But that point had not arrived in this case. The mere fact that PCMC UK and one of its employees were mentioned in the contract whose implementation gave rise to a finding of infringement on the part of a company in the same group was insufficient to discharge the burden on Perini.

Accordingly, I consider that the Judge was right to hold that PCMC Italia was liable, and PCMC UK was not liable, as a joint tortfeasor with LPC UK, for infringement of claims 16 and 17 of the 929 Patent.

The final point to be decided concerns the basis on which Perini’s costs should be paid by PCMC in relation to a particular issue. In its Defence, PCMC alleged that Perini had waived the right to allege infringement of the 929 and 168 Patents against PCMC as a result of a settlement agreement entered into in 1998 between Perini and PCMC’s holding company (“the 1998 agreement”). This argument (“the waiver argument”) was abandoned shortly before the trial. PCMC argued that there should be no order for costs, and Perini contended that PCMC should pay its costs on the indemnity basis, of the waiver argument. The Judge held that PCMC should pay Perini’s costs of the issue on the standard basis. Perini contends that the Judge should have awarded costs on an indemnity basis.

Perini realistically accepts that it faces an uphill task at least in principle, as an appellate court should be very circumspect about interfering with a judge’s exercise of discretion on an issue relating to costs. Unless satisfied that the judge went wrong in principle, misunderstood an argument, failed to take into account a relevant factor, or took into account an irrelevant factor, or that there was some sort of serious misunderstanding or procedural mishap, an appellate court should not interfere with a first instance decision on costs. In this case, Perini’s task is particularly difficult, as it is common ground that the Judge applied, or at least purported to apply, the right test, namely whether the facts relating to the waiver argument brought the matter “out of the norm”, in such way as to justify a departure from the standard basis.

In very summary terms, the Judge’s view was as follows. The point that Perini had lost the right to claim infringement due to the 1998 agreement was pleaded by PCMC in good faith; PCMC then amended its pleading to add that the 1998 Agreement was subject to Wisconsin law, principally in order to be able to rely on documents which would have been inadmissible on an interpretation issue in English law (because they were concerned with negotiating the 1998 agreement); this proved something of an own goal, as the documents effectively destroyed PCMC’s argument; within less than two weeks of appreciating this, PCMC formally abandoned the waiver argument; in those circumstances, the Judge had no doubt but that PCMC should pay Perini’s costs of the issue, but concluded that it would be inappropriate to require those costs to be assessed other than on the standard basis.

In his judgment on costs, the Judge briefly recited the history of the issue, and said that it was “completely clear that [PCMC] should pay the costs of this issue”. However, he refused to order that those costs be paid on the indemnity basis. As he put it, PCMC first relied on English law, but “they then sought to bring in Wisconsin law ... , and, ultimately, that proved their undoing”, and he did not “see in that course of conduct anything which takes this particular litigation outside the norm.” He also thought that PCMC had dropped the point “relatively rapidly after appreciating that the point was a bad one.”

As Perini accepts, there is plainly nothing wrong with that analysis as far as it goes. As Floyd J pointed out, if the court is too ready to penalise a party in costs for abandoning a bad point before trial, there is a danger that litigants will be reluctant to abandon bad points when they should do so. (I should add that this should not be taken as intending to discourage judges from awarding costs on an indemnity basis when points are abandoned, if the circumstances justify it.)

Perini’s case was that the Judge failed to take into account other factors, and that, had he done so, he would have reached a different decision. Perini contends that the Judge failed to take into account the fact that PCMC’s principal US patent attorney, Mr Chestnut, who negotiated the 1998 agreement, should have checked his papers relating to the 1998 agreement before the waiver argument was raised; instead, he only checked his papers shortly before trial, whereupon he realised that it was hopeless, and caused it to be abandoned in June 2009. Indeed, Perini suggests that the waiver argument was not taken in good faith. Perini also rely on the facts that the argument was not taken in negotiations before the issue of the instant proceedings, and that, well before it was abandoned, Perini’s solicitors alerted PCMC’s solicitors to the insurmountable problems faced by the point.

It is true that the Judge did not deal expressly with these points in his judgment, but it is clear that he concluded that PCMC raised the waiver argument in good faith, and that the reason that it was dropped was a result of PCMC considering the negotiating documents, in the possession of Mr Chestnut and of Perini, once it had raised the argument that Wisconsin law applied to the 1998 agreement. That assessment of the Judge’s view is confirmed once one looks at the evidence and at the transcript of the argument on costs.

The conclusion was one which Floyd J was plainly entitled to reach. There was evidence to support it from Mr Chestnut and from the solicitor acting for PCMC in these proceedings. No doubt, it might have been better if Mr Chestnut had looked at the documents in his files before the point was taken, but his failure to do so until Wisconsin law was pleaded is excusable on the basis that such documents would have been inadmissible. Further, it appears that perhaps the most important document was one which Mr Chestnut had thought was purely internal to PCMC (and therefore not admissible, as I understand it, in Wisconsin law), whereas the disclosure process revealed to PCMC that the document had been sent to Perini during the negotiations leading up to the 1998 agreement. The pleading of Wisconsin law actually assists PCMC on bad faith, as it is scarcely likely that it would have alleged that Wisconsin law applied, thereby rendering negotiating documents admissible, if it had appreciated that this would destroy its case.

For my part, I would go along with Perini’s argument to the extent of accepting that there was material which would have persuaded some tribunals at least to consider seriously making an order for indemnity costs against PCMC. However, I am in no doubt that the Judge was entitled to take the view that such an order was not one that he was prepared to make; Perini’s appeal on this issue should, in my view, accordingly fail.

Accordingly, my conclusions are as follows:

I believe that it follows from this that, if Hughes and Jackson LJJ agree, both the appeal and the cross-appeal are dismissed. The parties should try to agree the consequential matters as far as they can, and should also provide the court with a proposed order; as to those aspects which cannot be agreed, the parties should make their submissions in writing in the normal way.

I agree.

I also agree.