The case law on the novelty of generically defined compounds and particular examples of these was summarised in decision T 12/90. The board had to consider the novelty of a vast family of chemical compounds defined by a general structural formula, where the prior art also disclosed a vast family likewise defined by a general structural formula, the two families having a large number of products in common. The board pointed out that a distinction had to be drawn between two situations:

(a) If the subject-matter of the invention was a particular compound, whereas the prior art disclosed a family of compounds defined by a general structural formula including this particular compound but not describing it explicitly, the invention had to be considered novel (see T 7/86, T 85/87, T 133/92).

(b) If, with the same prior art, the subject-matter of the invention was a second family of compounds partially covering the first, the invention was not new (see T 124/87).

As regards case (a) the board said: "That case is not comparable with the present one in which a distinction must be drawn between the novelty of a group of substances defined by a general formula and a second group of substances partially covering the first and defined by another general formula, because the concept of individualisation naturally only applies to the structural definition of a single compound, not a collection of compounds".

Case (b) was extensively discussed in T 124/87 (OJ 1989, 491). This decision dealt with the problem of assessing the novelty of a class of compounds defined by parameters within numerical ranges. The patent in suit claimed a class of compounds defined by parameters within numerical ranges while the prior document disclosed a process by which a class of compounds could be prepared – comprising those claimed in the patent in suit – having the combination of parameters required by the main claim of the latter. In that particular case, the example specifically described in the prior document did not disclose the preparation of any particular compounds within the class defined in the claims of the disputed patent. However, it had been accepted by the patentee that a skilled man would have no difficulty in preparing such compounds within the class defined by the claims of the disputed patent using the process described in the said prior document, in combination with his common general knowledge, so that the disclosure of the prior document had to be regarded as not only limited to the particular compounds whose preparation was described in the examples, but also as comprising the general class of compounds made available to the skilled man in that technical teaching, even though only certain compounds within this class were described as having been prepared. Since the compounds as defined in the claims of the disputed patent formed a major part of this general class, they formed part of the state of the art and therefore lacked novelty.

In T 133/92 the question to be answered in examining novelty was whether the selection of the alkyl group as defined in claim 1 of the disputed patent had been made available to the public with regard to the disclosure of a prior document. By citing T 666/89 (OJ 1993, 495), the patent proprietors contended that the legally correct approach for deciding selection novelty was identical or very similar to that employed in determining obviousness. In particular, they argued that in cases of overlapping ranges of compounds, a claim to a narrower range as compared with a broader prior art range was always selectively novel if it could be demonstrated that the narrow range was inventive over the broader range. However, the board observed that in the case cited the board had repeatedly emphasised that selection novelty was not different from any other type of novelty under Art. 52 and 54 EPC 1973, so that the proper approach was to consider availability in the light of a particular document. Thus the board found that a claimed group of compounds, essentially resulting from omitting those parts of a larger group of compounds which a skilled person would have immediately considered as being less interesting than the rest, could not be selectively novel. In addition, in the board's opinion, a skilled person would, having regard to these considerations, have seriously contemplated applying the technical teaching of this prior art document in the range of overlap.