Tuesday, 18 March 2014

The Mysterious Affair of Partial Priority and Poisonous Coffee

This post is an edited version of a presentation I gave to the CIPA Yorkshire meeting in Leeds on 14 March 2014. It is reproduced here for the benefit of those who didn't attend and to assist those who did and who want to find out more. It covers a subject that will be familiar to regular readers of this blog, but attempts to put it all together in a fairly coherent whole and, purely for entertainment value, with a silly comparison to a Poirot mystery.

Today I will be talking about partial priority and poisonous coffee. Since I am talking to patent attorneys, I suspect most of you will have a good idea of what this is going to be about. Some of you might already be inwardly groaning about yet another presentation on poisonous divisionals. I promise I won’t go into long boring detail about what the issue is, or why we should or shouldn’t be worried about it. Instead, I will be looking at what underpins the whole poisonous issue, which is partial priority under Article 88 EPC, and I’ll illustrate this with one particular example from last year. Along the way I hope to be able to show how both the High Court and successive Technical Boards of Appeal have got it completely wrong.

Two important landmarks happened last year, both of which were causes for sadness and celebration depending on how you look at it, or which side you were on. The first was David Suchet completing the entire set of Poirot mysteries for television, after playing the character for 24 years. In his final case, Curtain, Poirot returns to the setting of his very first case and, after solving the mystery, dies.

The second landmark was the end of Nestle’s monopoly on Nespresso coffee pods, which for some years has been making enormous amounts of money for Nestle. Unusually for a non-pharmaceutical case, Nestle's patents on what are quite straightforward mechanical inventions turned out to have quite high value. A lot of money has clearly been thrown at trying to use them against alleged infringers, and to defend the patents against attack. Despite their best efforts, they were not able to keep their patent protection nor were they able to win the argument on contributory infringement, and had an important European patent revoked first in the UK, and then finally at the EPO Board of Appeal.

Agatha Christie was keen on using poison as a way of killing off her victims, and used it around 30 times in her books in various ingenuous ways (see here for more). She had studied the subject in detail and was therefore very careful to make sure that the details of how the victim was poisoned were absolutely right in her books. She is once quoted as saying: “Give me a decent bottle of poison and I will construct the perfect crime”.

The very first novel Christie had published was The Mysterious Affair at Styles, first published in 1920. This introduced the Belgian detective Hercules Poirot, who became so popular that he was to feature in another 32 novels and short stories. The case involved the apparent poisoning of the victim, Emily Inglethorp, and the main mystery lay in how the poisoning was done. Once this was clear, it was straightforward to tell whodunnit.

Mrs Inglethorp was found at five in the morning convulsing in her bed, and died shortly afterwards. It turned out she had been poisoned with strychnine, which the police assumed had been in her coffee. The coffee cup was found broken on the floor by her bedside, but it was not just broken, it had been ground into the floor, probably by a shoe. She had taken her coffee to bed at around 8 o'clock the previous evening, after having a light meal. The effects of strychnine though are faster acting than this, and would usually take only about an hour to manifest. She had also taken to bed a bottle of tonic and her night-time cocoa in a flask. Although the tonic contained some strychnine, the concentration was so low that it would not have been enough to kill her. 

The question is then: was it the coffee that was poisoned, or was it something else? The answer to this question was key to how the mystery was solved. As we will see, the answer also turns out to be very similar to what happened in the Nestle coffee pods case, so if you know what happened there you should have a good idea of how the mystery was solved.

Now on to the coffee pods. Nestle have sold many millions of coffee pods. Last year nearly 200 million pods were sold in the UK alone, and Nestle accounted for a large fraction. Unfortunately for Nestle, any patents they had on the coffee pods themselves, which is the valuable item, expired a while ago. The earliest one was applied for in 1976, so is long gone. This US patent, showing the same basic construction of capsule, issued in 1979.

It would not have been much use to keep innovating the coffee pods and get new patent protection because any new designs would need to be backwards compatible with older machines, otherwise customers will get upset and stop buying them. So, what to do? Do Nestle simply accept defeat and allow others to start selling copycat coffee pods and chip away at their market. Of course not. Their main weapon is, of course, marketing, and on a big scale, but they certainly didn't give up on trying to keep some kind of patent protection and use it against their competitors.

The only way to keep some patent protection going was to keep innovating the machines rather than the pods. Here is an example of the latest version of Nestle’s ‘Essenza’ coffee machine. To make a coffee, all you need to do is pop a pod into the top of the machine and, after closing the lever, press the button to make the coffee. To eject the used pod and get the machine ready for the next one, you just lift the lever and the used pod automatically drops into a container underneath.

The way the coffee pods are inserted and removed was the subject of a patent application, EP09007962. In simple terms, what happens is that the capsule is inserted vertically into a space between a moveable part and a fixed part. The moveable part is then pushed forward by turning the lever arm, which presses the capsule against the fixed part and pierces the end of the capsule, ready for hot water to be passed through the capsule under pressure. The coffee then passes through a filter in the capsule, through the fixed part and out of the spout. To eject the capsule the lever is brought up again and the capsule drops down into a container.

A key part of the invention was that the axis of the capsule when first dropped into the machine would be offset from the axis of the moveable part. In the figure on the left, the two axes are shown with the capsule in this intermediate position. After the moveable part has been brought forward, the axes are aligned by the capsule being brought downwards so that it passes a pair of flexible stops on either side. This allows the capsule to drop into the container below once the moveable part is brought back. This is effectively what was claimed in the later application, and what was granted in 2010 as EP2103236B1.

The patent was opposed by six companies, who all wanted to sell competing coffee pods, or machines that used Nestle compatible pods, and each saw the patent as a barrier. Whether Nestle had any grounds for stopping others selling compatible pods using their patent is an issue I don’t have time to go into here. What’s of interest here is one of the arguments the opponents used to try to invalidate the patent. Claims 1 and 5 of the patent as granted (in the official English translation) read as follows:
1. Extraction system comprising a device for extraction of a capsule and a capsule (16) which can be extracted in the device; the capsule (16) comprising a guiding edge in the form of a flange, the device comprising
- a first fixed part (2),
- a second part (3) which is mobile relative to the first part,
- a housing (4) to receive the capsule and defining when in a position of closure of the mobile part against the fixed part an extraction position of the capsule about an axis (25) in said housing,
- a part for introduction and positioning of the capsule comprising guiding means (6,7) arranged such as to insert the capsule by gravity and to position said capsule in an intermediate position; the guiding edge being taken charge of in the guiding means (6,7);
said guiding means being introduction slides allowing the engagement of said flange;
- a system of outflow of beverage (19, 53), said mobile second part (3) being configured to displace the capsule (16) from the intermediate position to the extraction position upon closure of the device,
characterised in that the guiding means (6, 7) comprise stopping means (20) configured to retain the capsule (16) in an intermediate position; the flange pressing against these stopping means in the intermediate position, and in that the second mobile part (3) takes charge of the capsule to displace it from the intermediate position to the extraction position about the axis (25) of the capsule in the extraction position in said housing (4) such that in its movement, the mobile part acts on the capsule to bring it down, the flange of the capsule passing beneath the stopping means (20), and to push it about the axis (25) of the mobile part in its extraction position.
5. Extraction system according to claims 1, 2, or 3, characterised in that the stopping means retain the capsule in a inclined fashion with respect to the axis (25) of the capsule in an extraction position in said housing; the capsule tipping slightly with respect to the axis of the capsule in the extraction position (25).

The key parts of the patent are those underlined: the housing to receive the capsule, and the capsule being inclined, which is worth noting is only in claim 5.

In simple terms, the argument goes like this. The priority document, which was an earlier European patent application that also published (as EP1495702A1), disclosed the same coffee machine as in the patent. What it didn’t contain was a detailed description of the machine that was in the later application. There were two things the opponents argued were not disclosed in the priority application.

The first was the option of the housing to receive the capsule being somewhere other than in the second part. As you can see from the drawing above, the housing is only in the moveable part. The patent did not define in the claim that the housing was in either part, which left open the possibility of the housing being in either or both parts.

The second was the extra option mentioned in the patent, but not the priority application, that the capsule in the intermediate position was held inclined relative to the axis when the capsule was in the extraction position.

Both of these issues raised the possibility of the patent not being entitled to priority. If so, the earlier application became prior art under Article 54(3) and made the patent invalid for lack of novelty.

Nestle managed to get their patent through opposition reasonably unscathed by making some amendments, but the parties appealed the decision. While the appeal was still pending, the patent came before Mr Justice Arnold at the High Court and the same issues about priority came up. Unfortunately, Mr Justice Arnold got it completely wrong on both points (see here for more).

I’ll deal with the second point first, as it’s easier to see how it was decided wrongly. This is the diagram presented at the High Court showing the different ways the pod could be arranged relative to an axis. The pod could be offset from the axis, as in the top figure, or it could be inclined as in the middle drawing. It could also be offset and inclined, as in the bottom drawing. The priority document only disclosed that the pod was offset, but the later application and the patent disclosed that the pod could also be inclined.

Justice Arnold reckoned that, because the priority document did not disclose the inclined option the patent covered more options than were present in the priority document and was therefore not entitled to priority. The problem with this logic is that the inclined option was only present in claim 5 of the patent, not in claim 1 (see previous slide). Whether this is Justice Arnold getting completely confused about priority or being misled by Nestec is not clear, but it probably didn’t help that Nestec argued the point that claim 1 covered the three different orientations shown in the drawings, when in fact it didn’t cover the middle one.

The other point was about whether the patent lost priority because the housing was only described as being in the moveable part in the priority document, but the patent didn’t specify which part the housing needed to be in. This is where Justice Arnold got it wrong again, but the reasons are not quite as clear. He referred to the EPO Enlarged Board decision in G 2/98 to figure out when it would be acceptable for a claim to have multiple priorities.

Just to remind you, this is what Article 88 EPC says about multiple priorities:

Multiple priorities may be claimed in respect of a European patent application, notwithstanding the fact that they originated in different countries. Where appropriate, multiple priorities may be claimed for any one claim. Where multiple priorities are claimed, time limits which run from the date of priority shall run from the earliest date of priority.

This is what the enlarged board in G 2/98 said about how to interpret Article 88:

The use of a generic term or formula in a claim for which multiple priorities are claimed in accordance with Article 88(2), second sentence, EPC is perfectly acceptable under Articles 87(1) and 88(3) EPC, provided that it gives rise to the claiming of a limited number of clearly defined alternative subject-matters. (reasons, point 6.7)

Justice Arnold’s reasoning was that there were three alternatives: the housing could be in the first or second part, or in a third alternative it could be divided between the parts. He concluded that the third option was not a clearly defined alternative to the first two, even though he had just defined it, and the patent was therefore not entitled to priority at all.

Reading the wording of G 2/98, on the face of it you can see how this kind of decision might seem to make sense. However, the enlarged board did not actually intend Article 88 to be so narrowly interpreted. As part of the reasoning in G 2/98, the board refered to a memorandum that led to Article 88 allowing multiple priorities for one claim, and stated that “the memorandum can be said to express the legislative intent underlying Article 88(2), second sentence, EPC”.

The memorandum gave three examples where it should be possible to allow multiple priorities in one claim. The first was where a chemical formula is broadened in a later application, the second where a numerical range was broadened, and the third where a field of use was broadened. In all cases, the memorandum stated that broadening should allow partial priority for the part of the claim that covered the earlier application, while the rest of the scope would be entitled to the later priority date.

The third example in the memorandum was about a priority document disclosing a method of coating the inner walls of pipes, which was then broadened to a method of coating the inner walls of hollow bodies. The later broader scope of claim would be entitled to partial priority for the case of pipes, and to a later priority for the case of hollow bodies other than pipes. So, when G 2/98 stated that multiple priorities were acceptable provided a generic claim gave rise to a limited number of clearly defined alternative subject matters, they must have had this in mind, along with the other examples.

Comparing this example with the one in the Nestec case, it’s quite clear that making the claim scope broader by not specifying which part the housing is in fits quite well. The option of the housing being in the moveable part would be entitled to the earlier priority date, while the other options of the housing being in the fixed part or divided between the parts would be entitled to the later date.

To be fair to Justice Arnold, the same mistake was made in at least three decisions of the EPO boards of appeal, all of which thought they were following the guidance in G 2/98 but none of them referred to the memorandum.

In T 1877/08, the Board rejected a range of numerical values as not corresponding to distinctive alternative embodiments (point 2.4 of the reasons). The same point was made in T 476/09 (point 5.6 of the reasons). This is in direct contradiction to example b) of the memorandum, which specifically refers to numerical ranges being a possibility for multiple priorities in one claim. Both of these decisions must therefore had been wrongly decided on this point.

In T 1127/00, the Board rejected a claim to partial priority for a generic formula encompassing an earlier narrow range in the priority application because, as they said, the claim did not embrace a limited number of clearly defined alternative subject-matters in the form of an “OR” claim which could be split up into groups of different priorities (point 6 of the reasons). This is, however, obviously contradictory in terms of both examples a) and b) of the memorandum.

The only board of appeal decision so far that has actually looked at what G 2/98 actually meant by referring to the memorandum is T 1222/11. In point 11.8 of the reasons, after going through the various examples in the memorandum the board set out in general terms how multiple priorities in a single claim should work. I will paraphrase here, as the board was quite wordy. They said:

In so far as a subject-matter disclosed in a priority document and encompassed by a broader claim of a European application is concerned, the decision on whether priority can be acknowledged for this subject-matter is independent of whether said subject-matter or embodiment disclosed in the priority document is identified in the broader claim as a separate alternative embodiment. Also, when an application on the basis of which a priority date is claimed encompasses a narrower subject-matter already disclosed in an earlier application, the decision on whether the claim to priority on the basis of the later application is valid does not depend on whether the narrower subject-matter disclosed in the earlier application is identified in said later application.

The need to have specific basis for the “limited number of clearly defined alternative subject-matters” in the application is therefore explicitly rejected in T 1222/11. Instead, it is enough for a claim to be theoretically divided up into a first part having a valid earlier priority and a second part having a different priority. For the case of the method of coating hollow bodies, the later application does not need to basis for a claim that states a method of coating hollow bodies other than pipes, because this is already covered by theoretically dividing up the claim into two parts with different priorities.

What should now be obvious, but I haven’t mentioned yet, is how this way of looking at partial priority is the solution to a hypothetical problem that has been discussed for a while. The issue is whether a priority document that is published and becomes Article 54(3) prior art for later filed application can knock out the later application if it claims a broader scope than what is in the priority document. In the Nestec case, the issue was publication of the priority document, which was also a European application. In other cases it might be a divisional application, which could be argued to be Article 54(3) prior art for its parent once it publishes.

Taking the narrow view of partial priority, which Justice Arnold did, means that your own priority document might knock out your patent because it falls within the broader claimed scope of the patent. Taking the broader view, however, it should never be possible for a priority document to be used in this way. This is because the priority document is not prior art for the part of the claim that has a valid claim to priority, while the rest of the scope is at least novel so cannot be knocked out by the priority document alone.

It would be nice if this issue could have been resolved by the board of appeal or the Court of Appeal in the Nestec case. Unfortunately, the EPO board of appeal decided to revoke the patent when they came to hearing the case in October last year and didn’t even get to discuss the issue of partial priority (see here for more). Instead, the board found that the patent failed for added matter, as a result of amendments the proprietor had made during the appeal. The possibility of getting Justice Arnold’s dodgy decision overturned on appeal was then removed, as there was no longer any patent to argue about. 

I can now go back to the question I asked at the beginning: was the coffee poisoned? In the Nestle case, even though the decisions went against them, and the patent was probably invalid anyway for other reasons, I think the answer is that the coffee was not in fact poisoned. It should not have been possible for the priority document to be used against the application. The fact that Justice Arnold and several EPO boards of appeal have found otherwise does not mean that I am wrong, just that the right argument was not presented. If it had been, the conclusion would have been that the coffee was not poisoned and the patent might have survived. In the end though, the patent was killed off by the EPO board of appeal by added matter.

Strangely enough, this is also the answer to the Poirot mystery. If you remember, I mentioned that the victim Mrs Inglethorp was found to have died of strychnine poisoning. Poirot had the coffee analysed and found there was no strychnine in it, which he suspected anyway because if it had been poisoned it would have acted much more quickly. The coffee therefore wasn’t poisoned after all. What was poisoned was her cocoa, and Mrs Inglethorp had done this herself by adding her tonic to it. Although the tonic should have had only small amounts of strychnine, and not enough to kill her, someone else had added bromide to it, which had reacted with the strychnine and made it precipitate out to the bottom of the bottle. The effect was to make the last dose fatal, as it contained most of the poison. The solution to the mystery, as in the Nestec case, was that it was added matter that did it, and not poisoned coffee.


  1. I was surprised that your previous posting, Tufty, on technical effect, got no comments because I thought it excellent. So I rush to make a comment here, straightaway, to the effect that I found the post very entertaining and thank you for making the effort.

    Just one small quibble though.

    You write that Nestec's Claim 1 "didn't cover" the middle one (inclined) of your three pod ways. That statement of yours raised my eyebrows. Can you get them back down again, perhaps?

  2. The reason I say that the middle one is not covered is that claim 1 requires the capsule axis to be offset in the intermediate position (because it has to be displaced from this position to be in the extracted position). In the middle example the capsule is only inclined and is not offset. This I think is where Nestec's counsel went wrong, and where Arnold J was misled.

  3. OK, thanks for that answer. If I understand you, Tufty, you find the feature "offset from the axis in the intermediate position" to be implicit in the claim.

    I can see that the claim requires displacement of the capsule from its intermediate position to its extraction position. What does "displacement" mean? Is a mere rotation of a capsule a "displacement" of that capsule, or is there displacement only when it is moved in translation? I don't know, but we might need to discuss that point before going on to construe the rest of the claim.

    Moving on, I regret that am still confused.

    I see that in the extraction position the capsule is "about the axis". What I'm still not seeing is that the intermediate position must necessarily be not "about the axis".

    And what if the capsule is offset but not by a distance (as in two of the three drawn dispositions) big enough to carry the axis of the mobile part to a point outside the volume of the capsule. In such a case, the capsule is both "offset" from the axis of the mobile part and also "about" the axis of the mobile part. In your eyes, are all "offset" dispositions within the scope of the claim, or only those offset positions where the axis of the mobile part does not pass through any part of the interior of the capsule?

    Sorry to be so difficult. It is just that I am sceptical when I read that a judge has misunderstood something. I expect to find a miscommunication rather than a misunderstanding.

  4. I don't think the possible alternative interpretations of the claim alters things. If you read the judgment you will see that Nestec proposed an amendment to claim 1 that made explicit the offset feature. Arnold J still thought that this resulted in the claim not being entitled to priority because it covered additional options. The point regarding the inclined option only being present in claim 5 still stands, and Arnold J was therefore still wrong on this point. He seemed not to understand that priority would only be lost for claim 5 if the additional inclined option was not derivable from the priority document, not for claim 1. This was argued by Nestec (see para 103 of the judgment), but perhaps it wasn't argued strongly enough.

  5. Thanks for that, Tufty. I'm lazy, preferring reading the blogs and asking you to ploughing through the case file to find the answer myself. I should now say that the way you now explain it I find convincing, and agree that further discussion over where exactly the periphery of Claim 1 lies does not change the basic priority point.

  6. That's fine. Glad to be of help. I glossed over that point in my presentation to save time and avoid unnecessary detail.

  7. "The solution to the mystery, as in the Nestec case, was that it was added matter that did it, and not poisoned coffee."

    Bravo, bravo.