In decision T 557/13, dated 17 July and published this week, EPO Technical Board of Appeal 3.3.06 has referred five questions to the Enlarged Board in a bid to resolve the current debate surrounding so-called “poisonous priorities”.  This is the name given to a controversial set of circumstances in which a European patent application can be deemed to lack novelty over its own priority document, its own divisional application or its own parent application.  The issue of “poisonous priorities” has been a source of much debate in the European patent profession in recent years and therefore the referral of this thorny topic to the Enlarged Board will be welcomed by patent attorneys and patent proprietors alike.

In this briefing note we review the background to the debate before summarising the facts of T 557/13 and the questions which have been referred.

Background: “Partial priorities” within claims

Under Article 88(2) EPC, a claim in a European patent application can have multiple priority dates.  According to Enlarged Board decision G 2/98, a feature in a claim is only entitled to an earlier priority date if both the priority document and the later application relate to “the same invention”.  This is assessed in the same manner as added matter: for a feature to be entitled to priority, it must be “directly and unambiguously derivable” from the disclosure of the priority document.

As a simple example, consider a first application (1) which discloses and claims feature “A”.  A later application (2) claiming priority from application (1) is then filed, with (2) additionally disclosing feature “B” as an alternative to “A”, and claiming “A or B”.  Feature “B” is not derivable from application (1).  Thus, the claim to “A or B” has two priority dates: feature “A” is entitled to the priority date of application (1), while feature “B” is entitled only to the filing date of application (2).  The claim in application (2) is said then to enjoy “partial priority” in that the claim is entitled to the priority of application (1) only to the extent that it relates to the same invention.

For example, application (1) could disclose an article made of copper and application (2) could disclose an article made of copper or steel.  “Copper” would be entitled to (1)’s date, but “steel” would only be entitled to the date of filing of (2).

In a case such as that above where features “A” and “B” are clearly individualised in the claim, it is straightforward to identify which portion of the claim is entitled to which priority date.  However, consider now that instead of claiming “A or B”, application (2) claims “C”, where “C” is a genus encompassing “A”.  In the example above, imagine that instead of “copper or steel” application (2) instead claimed an article made of “metal”.  Other examples of this type of situation are:

  • application (1) discloses a parameter having a specific value (e.g. a pressure of 15 Pa) and application (2) claims a range encompassing the specific value disclosed in (1) (e.g. a pressure of 5 to 20 Pa);
  • application (1) discloses a particular chemical substituent (e.g. methyl) and application (2) claims a generic group including that substituent (e.g. alkyl);
  • application (1) discloses a particular use (e.g. a method for coating the inner wall of a pipe) and application (2) claims a broader field of use (e.g. a method for coating the inner wall of a hollow body).

This type of situation was considered in G 2/98, in which the Enlarged Board remarked (at point 6.7) that a generic claim could be entitled to multiple priorities “provided that it gives rise to the claiming of a limited number of clearly defined alternative subject-matters.”

Divergent interpretations of G2/98

The present controversy arises from divergent interpretations of the Enlarged Board’s remark in G 2/98 concerning “clearly defined alternative subject-matters.”

The first, strict, interpretation essentially holds that the “clearly defined alternative subject-matters” must be separately identified in the claim.  So, in the above example, if application (1) claimed “copper” and application (2) claimed “metal”, the claim would only be entitled to the filing date of (2).  To preserve any claim to priority from (1), application (2) might instead need to claim “copper [which is entitled to the priority date] or a metal other than copper [which is entitled to the filing date]”.

The second, broader, interpretation instead argues that it is sufficient that the “alternative subject-matters” can be identified implicitly without needing to be specified as such within the claim.  So, reusing the example from above, if application (1) claimed “copper” and application (2) claimed “metal”, it is straightforward to envisage separating the later claim into “copper” and “metal other than copper” without needing to write it as such.  The claim then enjoys priority from application (1) to the extent that it encompasses copper, and takes the filing date of (2) to the extent that it encompasses other metals.

Although this consideration seems rather academic, it can have significant effects on the patentability of the claim, since different prior art is citeable against portions of the claim with different priority dates.

When do “partial” priorities become “poisonous” priorities?

Partial priorities have the potential to cause a headache for patent owners in two main types of situation, provided that the first interpretation of G 2/98 is applied.

In the first situation, application (1) is filed as an EP application disclosing a particular species or range.  Application (2) is an EP application claiming priority from (1) but with a broader claim scope.  If application (1) is allowed to publish, it becomes citeable as prior art against application (2) under Art. 54(3) EPC.  If the first interpretation of G 2/98 is applied, application (2) is not entitled to priority from application (1).  Application (1) contains a disclosure within the scope of the claim of application (2), and therefore application (2) lacks novelty over its own priority document!

In many cases it may be rare that a priority application would be kept alive beyond the end of the priority year.  However, a second, more insidious, situation arises with regard to divisional applications.  This happens as follows: application (1) is filed as a priority application and application (2), with a broader claim scope, is then filed claiming priority from application (1), which is abandoned.  Subsequently, application (3) is filed as a divisional application, thus claiming the priority and filing dates of application (2).

The claims of divisional application (3) must be directly and unambiguously derivable from the content of the parent application (2) as filed.  However, if the disclosure in the parent application (2) is broader than the disclosure of its own priority application (1), as is often the case, then it is also possible for the divisional application to contain broader claims than the priority application.  Under the strict interpretation of G 2/98, the broad divisional claims are only entitled to the filing date of the parent application (2) rather than enjoying the date of the priority application (1).  If the parent application (2) contains a disclosure which is entitled to the earlier priority date and which falls within the scope of the broad divisional claims, then the divisional (3) then lacks novelty over its own parent (2).  Similarly, the reverse situation is also possible in which a parent application is deemed to lack novelty over its own divisional.  In fact, given the correct combination of disclosures and priority dates, it is conceivable that a parent application and a divisional application may be deemed to anticipate one another, or even that multiple divisional applications within a single family could deprive one another of novelty!

The adverse consequences of these situations for patent-holders have led to the adoption of the somewhat dramatic title of “poisonous” or “toxic” priorities (or, equivalently, “poisonous/toxic divisionals”).  In particular, decision T 1496/11 appeared to support the first interpretation of G 2/98, and since then the “toxic divisional” concept has become a relatively common line of attack in opposition proceedings.

Of course, if the broader interpretation of G 2/98 is followed – that is, if the broader claim is entitled to partial priority - the problem goes away.  To the extent that the claim encompasses the disclosure of the earlier document, it enjoys priority and therefore the earlier document is not prior art.  To the extent that the claim encompasses different subject matter from the earlier document, the earlier document’s disclosure does not fall within that scope and so cannot anticipate the claim.

At present, there is no uniformity in the approach of the EPO’s Technical Boards of Appeal with regard to the issue of partial priorities.  Thus a number of recent EPO decisions, as well as at least two judgments of the English courts, have followed the stricter first interpretation of G 2/98, while a number of other decisions (notably T 1222/11 and T 571/10) have supported the second interpretation.

Decision T 557/13

Infineum USA L.P. was the proprietor of European patent application 98203458.9, which was a divisional application.  Both the divisional application and the parent application claimed priority from an earlier British patent application.

A patent was granted on the divisional application with claim 1 directed to the use of a cold flow improver in enhancing the lubricity of a fuel oil composition.  The claim specified that the cold flow improver was an oil-soluble polar nitrogen compound with certain types of substituent group.  This was a generalisation of a more specific disclosure from the priority application.

The patent was opposed by Clariant Produkte (Deutschland) GmbH and revoked by the EPO’s Opposition Division.  According to the Opposition Division’s application of the strict interpretation of G 2/98, claim 1 of the divisional patent was not entitled to the date of the priority application.  T 665/00, T 1877/08, and T 1496/11 were cited in support of the application of the strict interpretation of G 2/98.  However, the parent application was deemed to enjoy the earlier priority date and thus constituted prior art under Article 54(3) EPC.  Example 1 of the published parent application fell within the scope of divisional claim 1 and therefore the divisional patent lacked novelty over its own parent.

Infineum appealed this decision, arguing essentially that it was not permissible under the Paris Convention and the EPC for a divisional application to have a different priority date from its parent application even if the divisional application did not meet the “same invention” test.  Infineum further argued that partial priority should be acknowledged for claim 1 based on the reasoning in T 1222/11, which advocates the broader interpretation of G 2/98.  Infineum further argued that, taken to its logical extent, the strict interpretation of G 2/98 could even lead to patent applications being citeable as prior art against themselves if different portions within a single application had different priority dates.  Infineum also pointed out an alleged conflict with the position under German national law regarding partial priorities.

Clariant argued that the stricter interpretation of G 2/98 should prevail, and that claim 1 could only validly claim priority from the priority document if the claim could, in its entirety, be directly and unambiguously derived from the priority document.  Clariant further argued that there was no legal basis for a special link between parent and divisional applications which would prevent them from having different effective dates.

Both parties requested the referral of several questions to the Enlarged Board of Appeal.  The Technical Board of Appeal, having reviewed the legal framework for partial priorities and the development of the case law of the Technical Boards of Appeal, agreed that the conditions for allowing partial priority to be recognised were a point of law of fundamental importance and that the law had not been applied uniformly.  The Technical Board therefore agreed to refer the matter to the Enlarged Board.

The referred questions

The questions referred are as follows, in which the term generic “OR”-claim is intended to refer to a claim of the type discussed above in which the claim encompasses alternative embodiments without specifically naming them as such:

  1. Where a claim of a European patent application or patent encompasses alternative subject-matters by virtue of one or more generic expressions or otherwise (generic "OR"-claim), may entitlement to partial priority be refused under the EPC for that claim in respect of alternative subject-matter disclosed (in an enabling manner) for the first time, directly, or at least implicitly, and unambiguously, in the priority document?
  2. If the answer is yes, subject to certain conditions, is the proviso "provided that it gives rise to the claiming of a limited number of clearly defined alternative subject-matters" in point 6.7 of G 2/98 to be taken as the legal test for assessing entitlement to partial priority for a generic "OR"claim?
  3. If the answer to question 2 is yes, how are the criteria "limited number" and "clearly defined alternative subject-matters" to be interpreted and applied?
  4. If the answer to question 2 is no, how is entitlement to partial priority to be assessed for a generic "OR"-claim?
  5.  If an affirmative answer is given to question 1, may subject-matter disclosed in a parent or divisional application of a European patent application be cited as state of the art under Article 54(3) EPC against subject-matter disclosed in the priority document and encompassed as an alternative in a generic "OR"-claim of the said European patent application or of the patent granted thereon?

Questions 1 to 4 thus address the issue of “partial priorities” in general, while Question 5 addresses the question of whether parent and divisional applications can be cited as prior art against one another if partial priority is denied.

Question 2 is particularly interesting.  In the Board’s reasoning (at point 17.2.3(ii) of the decision) it is noted that G 2/98 specifically addressed the issue of “multiple priorities” within a claim, but that G 2/98 did not address the so-called generic “OR”-claim situation discussed above in which partial priorities can arise.  In asking this question, the Technical Board is therefore effectively asking whether “multiple priorities” and “partial priorities” are to be assessed according to different criteria.

It is also notable that the Board declined to refer a question on the interpretation of Article 76(1) EPC, which had been requested by both parties.  Such a question would effectively have asked whether a divisional application should automatically be deemed to enjoy the same right of priority as its parent application even if its claims relate to subject matter which would otherwise be deemed to lack priority.  The Board ruled that if the answer to referred question 1 is “no” – i.e. that partial priorities cannot be denied – then there would be no need for a separate question on Article 76(1); if the answer to question 1 is “yes”, then the answers to the remaining questions should be adequate to deal with the Article 76(1) issue.


The referral of the above questions to the Enlarged Board is very much to be welcomed in view of the present legal uncertainty surrounding partial priorities.  A decision from the Enlarged Board is not to be expected for at least a year, however.

Patent applicants should continue to exercise caution when drafting claims differing significantly in scope from the disclosure of the priority document, and when filing divisionals, to try to ensure that a “poisonous priority” situation does not become an insurmountable obstacle if the strict interpretation of G 2/98 is upheld.  Although the “toxic divisional” attack is an attractive weapon in opposition proceedings, opponents should continue to formulate supplementary attacks as a precaution against the Enlarged Board reaching a decision which closes off the “toxic divisional” route.