[Methodologic aspects of biomathematical description of growth processes]

The biomathematical processing of growth data aims the separation of the deterministic variation of the growth variable from the superposed random fluctuations. For the quantitative representation of the deterministic component, 3 possibilities are noted (mean values polygon, empirical regression, and least squares fitting of functions). By 2 examples of measured growth data, the suitability of these approaches will be discussed to give reproducible results and to avoid artifacts. The model free recording of the deterministic component by empirical regression is recommended.