In assessing the inventive step involved in an invention based on a combination of features, consideration must be given to whether or not the state of the art suggested to a skilled person precisely the combination of features claimed. The fact that an individual feature or a number of features were known individually does not conclusively show the obviousness of a combination (T 37/85, OJ 1988, 86; T 656/93; T 666/93; T 1018/96). The question is not whether the skilled person, with access to the entire prior art, could have made the combination according to the invention, but whether he actually would have done so in expectation of an improvement (T 2/83, OJ 1984, 265; T 713/93; T 223/94; T 406/98). When assessing inventive step in a combination invention, the decisive criterion is not whether individual elements of the combination were known and obvious from prior art, but whether the state of the art would lead a skilled person to this particular overall combination of (possibly already known) features. Were this not so, it would be impossible for a combination consisting exclusively of known individual features to involve an inventive step (T 37/85, OJ 1988, 86; T 388/89; T 717/90; T 869/96).
A mere aggregation of features must be distinguished from a combination invention.
The existence of a combination invention requires that the relationship between the features or groups of features be one of functional reciprocity or that they show a combinative effect beyond the sum of their individual effects. In T 1054/05 the board stated that two features interacted synergistically if their functions were interrelated and lead to an additional effect that went beyond the sum of the effects of each feature taken in isolation. It was not enough that the features solved the same technical problem or that their effects were of the same kind and added up to an increased but otherwise unchanged effect (see also T 926/11, in which it was found there was no synergistic effect between the feature groups). See also in this chapter I.D.9.2.2.
In T 406/98 the board found that as a rule, particularly when large numbers of citations were involved, it was necessary to ask why the skilled person would consider documents in that specific combination, and whether, not knowing the invention, he had reason to do so. In this case, a complete solution to the problem required deliberate selection from a large number of citations.
In T 55/93 the appellant's argument that the alleged invention should have been regarded as a mere aggregation of solutions of two independent partial problems which were not interrelated, was not accepted by the board. This reasoning could only stand if the primary and more general problem were already known and solved by the prior art. In the case in point, not only could the primary problem underlying the contested patent neither be found nor be derived from the prior art documents, but also the claimed features complemented each other. The board stated that the features were functionally linked together, which was the very characteristic of a combination invention. The board stated it was no consequence whether, in a combination invention, possibly all features were already known per se, separately (T 37/85, OJ 1988, 86). It was wrong to select on the basis of a plurality of partial problems to be solved, the respective constructional means used in the apparatus combination, or the steps of the method worded in terms of functional features, which by working together provided a solution to the problem taken as a whole. The non-obviousness of a combination claim turned on the simultaneous application of all its features (T 175/84, OJ 1989, 71). A combination effect was also acknowledged in T 120/88, T 731/94, T 434/95, T 897/95, T 1201/13.