[Sampling techniques in quantitative morphology (author's transl)].

Abstract

Adequate methods of sampling are by far the most neglected procedures in stereological work. This is true of both theory and practice. Indeed, it is often forgotten that Delesse not only introduced the stereological method, but also specified the necessary conditions for its use. To help make good this deficiency methods of stereological sampling are reviewed. In particular, the ratio estimate is discussed and examples are given. Estimation of the percentages of each cell class in populations of blood cells and cells from vaginal smears have problems in common. In most cases the bias of the ratio estimate is trivial. The formula for approximate calculation of the standard error of the ratio estimate requires more sampling units than most people realize. In addition, in artifical smears, a more even distribution of the cells is recommended.