Share this article:

Rethinking the Novelty of Selection Inventions: Board in T 1688/20 Reverts to First Principles

By Michael Ford, Partner

Assessing the novelty of a patent claim in the materials science space (as well as in chemistry and the life sciences) often involves determining whether a “selection invention”, as it is known at the European Patent Office (EPO), has been made. A selection invention arises where the prior art describes subject-matter in a generic form as a set of elements or a numerical range and the patent or application claims the matter more specifically by selecting individual elements, subsets or subranges. The EPO has developed various rules for determining whether or not a given type of selection is novel. However, as discussed in our May 2022 materials science webinar (a recording can be requested here), these rules are not necessarily consistent with one another and there are conflicting decisions by EPO Boards of Appeal on the matter. A new decision in case T 1688/20 potentially throws another spanner in the works.

Rules according to the Guidelines

The Guidelines for Examination define the rules which EPO examiners are expected to follow in normal circumstances in first instance examination and opposition proceedings. These rules are intended to reflect the established EPO case law at any given time. According to the present Guidelines, three general types of selection invention are possible.

Type 1: Selection of individual elements from lists of elements

Basic Rule: The selection of an element from a single list does not confer novelty, whereas combining elements selected from two or more lists of a certain length, in the absence of a pointer to the combination, is novel.

Type 2: Selection of a single sub-range from a broader numerical range

Basic Rule: The selected sub-range is novel if it is (a) narrow compared to the known range and (b) sufficiently far removed from any specific examples disclosed in the prior art and from the end-points of the known range. There is no longer a requirement that the sub-range is selected purposively from the broader range, in the sense that there must be some technical effect associated with the narrower range, for novelty.

Type 3: Selection of a first range which overlaps a second range

Basic Rule: Novelty of the first range is destroyed by an explicitly mentioned end-point of the second range, explicitly mentioned intermediate values, or a specific example in the area of overlap. Novelty cannot be achieved simply by excluding specific novelty-destroying values known from the prior-art range, it must also be considered whether the skilled person would seriously contemplate working in the area of overlap.

Conflicting case law

The Boards of Appeal at the EPO are not required to follow the Guidelines and they are instead bound only by the European Patent Convention and decisions of the Enlarged Board of Appeal. Boards therefore can and do sometimes come to decisions which appear to conflict with the general rules expressed in the Guidelines.

In our webinar last year, we discussed recent cases which throw doubt on the rules applied when assessing the novelty of Type 2 and Type 3 selections. For example, in decision T 0261/15, the Board expressed the view that the end-points of a known range, although explicitly disclosed in the prior art, are not to be treated in the same way as the examples. According to this Board, the skilled person would not necessarily contemplate working in the region of the end-points of a prior art range without some further teaching in this direction. This is because the end-points are not normally representative of the gist of the prior art teaching. The Board therefore disputed the legal basis for the requirement in the Guidelines that the claimed sub-range must be far removed from the end-points of the known range. It is thus somewhat ironic that the Guidelines actually cite T 0261/15 as providing basis for the rules given for Type 2 selection inventions, presumably to support the omission of a requirement for a purposive selection in the current Guidelines (this third criterion was also criticised in T 01261/15).

Detailed discussions of various conflicting lines of case law regarding selection inventions can be found in the epi (the Institute of Professional Representatives before the EPO) journal here and here. The arguments set out in these articles previously led the epi to request the President of the EPO to seek clarification from the Enlarged Board of Appeal regarding the novelty of selection inventions. The request was refused.

T 1688/20

In this new decision T 1688/20, a Board of Appeal has departed entirely from the tests set out in the Guidelines for selection inventions.

In this case, the patent EP3056283 claimed a rotary atomizing electrostatic applicator for spraying paint onto automotive bodies configured in such a way to cause air discharged through air holes to be twisted at an angle of 56° or more and 59° or less. The Board had to determine whether this claimed range was novel over a prior art disclosure of twisting air through an angle of 50° or more and 60° or less, as well as a disclosure of the specific value 55°.

The Opposition Division had decided during first instance proceedings that the prior art disclosure was novelty-destroying by applying the criteria in the Guidelines: 56° to 59° was not narrow compared to 50° to 60°; 56° to 59° was not sufficiently far removed from the end-point 60° or the specific value 55°; and there was no purposive selection (this criterion was still in the Guidelines at the time of the initial decision).

In the appeal decision, however, the Board found that the terms “narrow” and “sufficiently far removed” as used in the Guidelines could not provide objective, solid and consistent criteria for establishing the novelty of a selected sub-range. Indeed, the Board stated that these terms are generally open to such a broad interpretation that a decision whether these criteria are met not only depends on the factual circumstances of each case, but could also depend on the subjective perception of the deciding body.

Going even further, the Board also stated that the criteria set out in the Guidelines simply did not need to be assessed for the question of novelty to be answered. The Board’s argument rested on the fundamental principle that the concept of disclosure must be the same for the assessment of novelty as it is for the assessment of entitlement to priority or added subject-matter. The disclosure test for determining whether or not an amendment adds subject-matter is known as the “gold standard” of G2/10 and requires consideration of what a skilled person would have derived directly and unambiguously, using common general knowledge, from the description, claims and drawings of the European patent application, seen objectively and relative to the date of filing. The Board concluded that a finding of lack of novelty should also require a direct and unambiguous disclosure in the prior art of subject-matter falling within the scope of the claim.

In addition, according to the established case law of the Boards of Appeal, any test developed to aid in determining whether an amendment adds subject-matter must not lead to a different result than would be obtained when applying the “gold standard”. In T 1688/20, the Board thought that the same should apply to the assessment of novelty. Thus, the Board considered that the criteria set out in the Guidelines were not necessary and that novelty could be assessed directly by applying the fundamental “gold standard”.

In this case, the Board found that the prior art did not directly and unambiguously disclose an angle value within the claimed range of 56° to 59°, applying the general principle that generic information (e.g. a broader range) cannot anticipate a more specific technical feature (e.g. a narrower range).

The Board went on to find that the selected range also involved an inventive step over the prior art because the claimed twist angle produced an effect of using the force of the atomization air to the maximum extent for the forced atomization of paint particles, thus improving coating efficiency and particle diameter. Even though the same technical effect could have been achieved across the broader range disclosed in the prior art, the presence of such an effect was not disclosed and so the selection was considered to be purposive. Moreover, there was no hint in the available prior art to suggest restricting the angle to the range claimed.


This case highlights that different outcomes are possible before the Boards of Appeal as compared to the first instance divisions of the EPO, and not simply because of a different take on the same facts or issues but because the deciding bodies are bound by different sets of rules. While it is unusual for Examining or Opposition Divisions to deviate from the Guidelines, Boards of Appeal are free to interpret the EPC and case law as they see fit. The Boards are also more inclined to revert to first principles when the application of a rigid rule could lead to nonsensible decisions.

We tend to see this pattern most often in decisions concerning the assessment of added subject-matter. Examining and Opposition Divisions are more likely to follow set rules such as the essentiality or three-point test for deleting claimed subject-matter (described in the Guidelines at H-V, 3.1) or the criteria which must be fulfilled to achieve an allowable “intermediate generalisation” (H-V, 3.2.1). In contrast, Boards of Appeal are more likely to be persuaded by arguments based on the “gold standard” for disclosure, according to which the most important question is whether an amendment presents the skilled person with new technical information. For example, in T 1906/11, the Board emphasised that classifying an amendment as an “intermediate generalisation”, an “omission of an originally claimed feature” or a “multiple selection from two groups of alternative features” did not permit the drawing of any conclusion about the allowability of the amendment; the only relevant question is whether a skilled person faced with the amended version of the application or patent, as compared to a skilled person having seen only the version originally disclosed, would derive from that amended version any additional technically relevant information.

As mentioned above, we have already seen other cases in which Boards of Appeal have questioned the selection invention tests set out in the Guidelines. The most recent edition of the Case Law of the Boards of Appeal continues to list various conflicting decisions, including T 0261/15. Given this most recent decision, different threads of case law on this matter seem likely to continue unless we receive a clarifying decision by the Enlarged Board.

It will also be interesting to see whether the specific arguments put forward in decision T 1688/20 are picked up by other Boards, in particular those with more of a technical focus on chemical compositions or biotechnological inventions. Although numerical ranges can of course appear in claims in any technical field, selection inventions are most commonly encountered in chemistry and the life sciences.

Practice points

For now, this decision is unlikely to have much of an impact on examination or opposition proceedings, as EPO examiners will continue to follow the selection invention criteria set out in the Guidelines.

However, it is another weapon in the applicant or proprietor’s arsenal which could potentially be deployed during appeal proceedings. Indeed, the issuance of decisions like this could provide hope that in challenging cases a Board of Appeal may be more likely to find a selection invention type claim novel than a first instance examiner.

We also note that the next update to the Guidelines will be available for preview in February and should come into force in March – it remains to be seen whether any changes are made to section G-VI, 8 regarding selection inventions.

This is for general information only and does not constitute legal advice. Should you require advice on this or any other topic then please contact or your usual HLK advisor.

HLK bubble graphic HLK bubble graphic

Stay connected with HLK

Keep up-to-date with the latest IP insights and updates as well as upcoming webinars and seminars via HLK’s social media.