[Problems with the morphometric-stereologic measurement of biomembranes].

In seeking to draw comparisons between surface density Sv estimates of various organelle membranes, it becomes apparent that main classes of data may be recognised. In the 1st (class A) there is a relatively good agreement between the uncorrected Sv values obtained in different research laboratories, e.g. the boundary membranes of mitochondria, lysosomes and nuclei. In the 2nd class (class B) fall estimates for organelles or organelle sub-compartments whose membrane boundardies cannot be everywhere reliably predicted when they form but vague images on the micrograph, e.g. cristae and membranes of the rough and smooth ER, i.e. class B membranes do not fulfil the assumption for morphometric counting that the identification on section is unambiguous (WEIBEL). For these class B membranes Sv estimates are often more widely discrepant for 2 main reasons. The 1st is connected with the difficulties in finding suitable correction factors which correct either for underestimation of membranes (due to membranes oriented obliquely to the axis of the electron beam) or overestimation due to the Holmes effect. The 2nd equally important reason lies in the subjectivity of the analyser who has to decide to count or not to count vague membrane images. In order to show the insecurities of class B membrane Sv estimates when presenting data it is proposed to abandon the 2nd and 3rd decimal. 2 proposals are made to tackle the problems connected with Sv estimates of class B membranes: Firstly a new method of obtaining a correction factor is presented and secondly, a rigorous standardisation of the criteria used for defining membrane profile intersections. Examples of the latter are shown.