Top heights are being calculated in forest growth research since a long time and are considered a good gauge for the characterisation of a site and for the changes which occur in it. The height development of the trees in a stand is considered the measurement of growth which is least influenced by the stand treatment and therefore min-ors the influences of the site itself and the climate. Top heights are frequently calculated from estimated values from height curves. A result of this procedure is that the standard deviations are reduced. Using research plots in which all the tree heights were measured, the extent of the reduction is determined. There are no large differences between the calculated top height and the "true heights", which were established by the measurement of all the trees although in a few cases they are considerable. They have a tendency to be larger in the uneven-aged stands than in the even-aged ones. On the other hand the standard deviation becomes considerably smaller and this must be taken into consideration with the interpretation of top height differences. In order to avoid this problem, it is suggested that all dominant trees are measured in research.