Paper compares various analysis methods suitable for an analysis of point pattern with directionally varying point density. As example was used point pattern from spacing experiment created according to Nelder design in which tree density changes greatly but in smooth manner. Such a pattern could be not analysed by methods assuming homogenous point density e.g. classical estimator of Ripley's K function. More useful are estimators of Ripley's and pair−correlation functions for inhomogeneous point patterns. K2 function, which proved to be efficient for heterogeneous point patterns with distinct changes in the local point density appears not to be efficient for heterogeneous point patterns with smooth changes in density.