In T 38/84 (OJ 1984, 368) the board of appeal pointed out that the achievement of a numerically small improvement in a process commercially used on a large scale (here enhanced yield of 0.5%) represented a worthwhile technical problem which should not be disregarded in assessing the inventive step of its solution as claimed (see also T 466/88, T 332/90). In T 155/85 (OJ 1988, 87) the board added that it was correct to say that even small improvements in yield or other industrial characteristics could mean a very relevant improvement in large-scale production, but the improvement had to be significant and therefore above margins of error and normal fluctuations in the field in consequence of other parameters. In T 286/93 the invention related to a process for manufacturing wrapping paper and board. The results for the process had shown that the machine speed and the mechanical quality of the paper obtained had improved by some 3 % vis-à-vis a process in which the order in which aluminium polychloride and cationic starch were added had been reversed. Since a process of this kind was obviously intended for the production of paper on an industrial scale, even a small improvement had to be regarded as significant.