[The fitting of various growth functions to the description of tree height growth]

Electronic computing renders possible new methods for determining parameters of functions of stochastic relations. PAUL (1975), for instance, published a FORTRAN-program for iterative non-linear regression for any function having an independent variable. Thus, optimum parameter values of a number of growth functions were determined on the basis of tree height development. Growth functions by RICHARDS (1959, according to v. BERTALANFFY 1941) and by SCHARF (1971) reach highest accuracy of fitting. If systems of growth curves are constructed, the elements of which differ by variation of a single parameter (e.g. height class fan), then the flexibility of multiparametric growth functions gets lost however.