AbstractDisperse materials are often characterized by their particle‐size distributions which can be determined by different measuring methods. Frequently the quantitative portions of the particle size classes obtained by way of experiment are considered insufficient and one strives for approximating the measured values by an analytic function. This can be reached, more or less satisfactorily, by power, exponential, or normal distribution functions which may include two, three, or four parameters.Publications as well as common practice have proved again and again that due to very rough or incorrect approximations wrong conclusions are derived from test and operating results, so that it appears expedient to investigate – when such an approximation would actually be advisable and – which function would be expedient in the specific case.Application of distriction functions will always be justified then when there is an actual need for utilizing the advantages connected with an approximation. Evaluation of a great number of measuring results revealed that the two‐parameter functions in most cases only permit an utmost rough approximation which might give rise to wrong conclusions. Carefully classified products, however, mostly require three‐ or four‐parameter functions for making an appropriate approximation.The functions, which appear appropriate for certain types of products, are set out in Table 1.In order to make the tabulated individual results accessible to a practical utilization a systematization was attempted. The outcome is a graph which is meant as an assistance for the practician and which presents the possibility of a quick selection of fu