The legislative purpose of Art. 83 EPC is to ensure that the skilled person can reproduce the invention without his own research or undue experimentation. Experiments are an undue burden if their primary aim is to find the solution to the problem but not if they are carried out merely to determine the numerical limits of a functionally defined range (T 312/88). They should quickly give a reliable picture of how the products can be produced or manufactured (T 475/88). However, it is not necessary for the experimental data filed with the patent in suit to be an exact repetition of the worked examples of the patent, as long as the experimental work can be regarded as being within the scope of the invention (T 674/96).

For experiments to be considered reasonable, the application need not disclose the best and easiest method; a long and complicated route which is nevertheless clearly successful may be considered reasonable (T 412/93).

The board in T 1133/08, faced with a multitude of options for selecting suitable materials, dimensions and procedural parameters which were merely outlined in the part of description relating to embodiments, found that there was no specific information describing in detail at least one way of carrying out the invention claimed. Experiments were needed to solve the problem (i.e. identify parameters and conditions resulting in a sinusoidal profile), and, as established in T 312/88 in conjunction with T 68/85 and T 18/89, such experimentation had to be considered unduly burdensome. If an invention had several variants, it was very important first of all to describe as many as possible in detail, rather than merely outlining them, to show the skilled person that the invention could be carried out in practice across the entire breadth of the claims. Here, not one single way of carrying out the invention was apparent, nor had any subsequently been demonstrated, e.g. on the basis of experiments. The board also analysed T 14/83, contrasting it with T 412/93 (genetic engineering).