email_pop up_pop




In 2012 Member States and the European Parliament agreed on the "patent package" - a legislative initiative consisting of two...

find out more

EPO Enlarged Board upholds partial priority for broadened claims – no more poisonous priorities or poisonous divisionals

Back in 2014 we wrote about the strict “same invention” rule used to establish whether a European (or UK) patent application is entitled to priority.  We wrote that the test is as strict as for added subject matter. We also wrote about the draconian consequences when one tinkers with the scope of the claims of a patent when filing a UK or EPO application claiming an earlier priority date.  In the worst case, the earlier application can itself be prior art to the patent application (e.g. a divisional) claiming priority from it.  This is the dreadful and bizarre concept of a “poisonous priority” or “poisonous divisional”. 

We are pleased to report that the EPO Enlarged Board of Appeal has issued its decision G01/15 on the subject of partial priorities and has now put to rest the notion of poisonous priorities and poisonous divisionals by relaxing its criteria for partial priority so that a later claim to an invention defined broadly can always benefit from partial priority to a part of its claim scope that has been described in a priority application

A claim that can notionally be separated into two claims of different scope, one entirely within the other, can have two priority dates.  The narrow disclosure in the priority application is not part of the state of the art under Article 54(3) EPC/Patents Act s.2(3) for the narrower claim, because the narrower claim is entitled to the priority date.  The rest of the later, broader claim is novel over the content of the earlier application but is only entitled to its date of filing.

The judgement is quite lengthy.  There were many amicus briefs, mainly in favour of abolishing the notion that a claim can be anticipated by the very application from which priority is claimed.  The judgement makes extensive reference to a memorandum of FICPI that formed part of the travaux préparatoires to the EPC back in 1973 and sets the problem out in a logical analysis of claims to the type “A or B” where only A is entitled to priority.  The decisions of the Boards of Appeal to date have taken different approaches to whether such a claim is entitled to partial priority. 

In a nutshell, the chemical and biotech divisions tended to apply a strict approach and only permitted partial priority where the broadening of the claim gave rise to a “limited number of clearly defined alternatives.”  Other boards, particularly electrical and mechanical boards, took a more relaxed approach.

The latter approach is perhaps best illustrated by the decision T135/01, which related to a method and apparatus for driving an electric motor and involved the claim expression “the interval between the first and second [switching] steps is between a quarter and three- quarters of the period (Tau) of a dominant frequency of vibration of the motor”.  In that case, the priority document described only that the period in question was “approximately = Tau/2”.  The question arose as to whether the broader claimed period could be notionally divided into a range of approximately in the centre that was entitled to priority (and everything else that was not).  If it were all-or-nothing, and the claim could not be seen to have partial priority, then the entire claim would lack novelty because the content of the priority document had in the meantime been made available to the public.  That board recognised partial priority to the narrower central range without requiring that it had to be spelt out in the claim. 

We are pleased to say that the Enlarged Board of Appeal has endorsed this more relaxed approach to partial priority, concluding that “the EPC does not contain any other requirements for recognising the right of priority beyond the same invention, whether for simple, multiple or partial priority”. 

The Enlarged Board did not quite admit that it was vacating its earlier decision in G2/98.  It regarded that decision as one that relates to multiple priorities and noted that it considers the partial priority situation to be a sub-group of multiple priorities.  So it may yet be that if an applicant is filing cascading applications from which different priorities are to be claimed, it would be wise to clearly define the different species or groups that are to be entitled to different dates.

G01/15 in pictures

Consider a priority application that contains specific subject matter A and a later application claiming priority from that priority application pursues a claim relating to subject matter B, a more general example of A:

The core of the test in G01/15, once the disclosure of the priority application has been correctly determined, is to conceptually divide the claimed subject matter B into a part entitled to priority, i.e. A, and a part not entitled to priority from the priority application, B-A:

B-A is only entitled to the filing date of the later application, however no self-collision occurs, as there is no overlap with the disclosure of A in the priority application for B-A (i.e. B-A is novel over A).

Note, however, that if A has in the meantime been published, it is necessary for B-A to be inventive over A.  This was the situation in T135/01.  In that case, perhaps surprisingly, the Board found the wider part of the range to be inventive over the disclosure of the narrow part of the range.  (The Board found that an interval outside the range “approximately Tau/2” of the dominant frequency of vibration would not have been obvious, in part because the published disclosure set out the importance of that value and dissuaded the skilled person from considering other values)

What would G01/15 have meant to the outcome of different cases in the recent past?

It would surely have made a difference to validity in Nestec v Dualit (reported in 2013 Patent issues).  In that decision, partial priority was denied and as a result, the priority application (filed in the UK) was part of the state of the art under section 2(3) of the Patents Act 1977.  A distinction was drawn over the earlier (and more generous) UK case of Unilin v Berry Floor (reported in Winter 2004 Patent issues) noting that the part of the Nestec claim that was not entitled to priority was not a clearly defined alternative.  But Unilin was not about clearly defined alternatives.  It was more about whether different features were separable.  In any case, the “clearly defined alternatives” notion is now defunct (an unnecessary condition) in the light of G01/15. 

Nestec must be decided differently, because either the later claim was broader (in which case it could be awarded partial priority under G01/15) or it was different (in which case the priority application was not novelty-destroying). 

It would probably not have made any difference in Hospira UK Generics Ltd. v. Novartis AG, [2013] EWCA Civ. 1663, because in that case Lord Justice Floyd took the view that the problem for Novartis in seeking to establish that its claim (claim 7) was entitled to priority was that the disclosure of the priority document was either too general or too specific.  The individual features of the claim were present in the priority document, but there was no disclosure linking them together.

Turning to the three Samsung cases discussed in our 2013 edition of Patent issues (see panel below), these are cases in which the content of the respective priority applications had in the meantime been published by submission to a standards organization.  In such cases the publication is prior art in relation to the part of the claim that is not entitled to priority.

Only the first of the three decisions might have had a different outcome, but then only if the inventive step analysis fell in Samsung’s favour (as for T135/01 where the analysis fell in favour of the patentee).  The broad claim to “QoS parameters” was probably supported by “data rate” (perhaps also by service type or input frame length). A claim limited to “data rate” could be entitled to priority.  But the claim to “QoS parameters [other than data rate]” would have to be inventive over the disclosure to the public standard submission of data rate (in the context). 

Brief summaries of three Samsung v Apple cases:

  • In Samsung I, the priority document disclosed a number of parameters service type, input frame length and data rate, whereas the later claimed generalized to “Quality of Service parameters” and it might well have been decided that at least one of the disclosed parameters (e.g. data rate) was a species of QoS parameter whereby the claim was divisible into a narrow part that was entitled to priority and a broader part that was not.
  • Samsung II would not have been decided any other way – the priority document was completely silent on one of the features of the claim and the standard that was published in the meantime included that feature.
  • Samsung III is a case where the claim was in one respect broadened but in another respect narrowed - from a scheme for allocating I/Q channels and OVSF codes to a scheme that was silent on allocating OVSF codes (broader) but focussed on a particular I/Q channel allocation (narrower).  The particular I/Q channel allocation was disclosed in the priority application, but only in combination with allocation of I/Q channels and OVSF codes.  This was not like Unilin.  It was more like Hospira (although that decision came later).

And where does this leave UK Patents Act 1977, section 6?

This section of UK domestic law purports to say that “for the avoidance of doubt” anything you do between the priority date and the later filing date in relation to matter disclosed in the priority application does not count against you.  So you should be free to publish the content of the earlier application without fear of it being later cited against you.  Fair enough?  “Well – no” said Justice Jacob (as he then was) in Beloit v Valmet [1995] RPC 705 – only if there is an effective priority declaration.  This brought PA section 6 into line with EPO decisions.  Jacob J. said the words “for avoidance of doubt” means that section 6 is mere clarification and has no substantive effect.  His decision was not challenged on appeal at the time and has remained good law.  The CIPA Guide to the Patents Act cites Nestec v Dualit as being consistent with Beloit.  Whereas Nestec v Dualit is now in doubt following G01/15 (see above), this does not bring section 6 back into full force and effect.

As we can see from the examples above, if the priority claim is invalid, the publication of the contents of the priority document will constitute part of the state of the art in respect of the part of the claim that is not entitled to priority and can invalidate the patent. 

This will remain so if the later application claims something more specific than the disclosure of the priority application (as in Hospira and probably Samsung III) or simply different from the disclosure of the priority application (as in Samsung II), while the content of the priority application will not be novelty-destroying, but can be cited for obviousness if published in the interim (as in the Samsung I scenario and in T135/01).

So it is far from true to say that publication between the priority date and the later filing date does not count against you. Section 6 remains moribund.


Our take-away advice does not change much.

  • Don’t assume that the filing of a priority application means it’s safe to publish. Only if the invention is not going to be developed further is it safe.
  • Be very wary about tinkering with the scope of the claims when filing in the UK or the EPO based upon an earlier priority application.

What changes is this.

  • You’re not entirely stuck with the scope of the disclosure of the priority application.  A generic expression can be supported by specific examples and it is not necessary that the claim language itself calls out the generic and specific claim scopes.
  • This does not mean you can broaden a feature of the claim and narrow some other feature, even when the narrow feature is disclosed in the priority application. Partial priority in such a claim may be denied if the features in the priority application were so linked that they cannot be considered in isolation.
  • You can add further embodiments, examples or claims (dependent or independent) to the later application within the priority year, provided there is unity of invention with the claims that are entitled to priority, and we can now broaden the claims to encompass the new embodiments or examples, BUT when all the dust has settled, if the claims that are ultimately granted do not encompass the originally described invention (e.g. because they draw from later filed embodiments), there will be no partial priority, and this can be fatal if the subject matter has in the meantime been disclosed to the public or has otherwise entered the public domain.  (In any event, new claims to further embodiments or examples will not benefit from priority and indeed the further embodiments or examples will give rise to a new right of priority that can be claimed later - Paris Convention Article 4F.)

Friday, March 10, 2017