For a number of quantitative electron microscopical techniques it is relevant to obtain an estimate of the thickness t' of the section which, in general, will differ from the actual distance t between the two cuts that generate the section. To estimate t' of ultrathin sections, several techniques have been adopted in the past, both with and without the aid of the electron microscope and additional equipment, which are summarized in an appendix. In the present study five methods have been evaluated experimentally using sections of ten different interference colours: (a) the 'small-fold' technique, (b) the 'electron scattering' method, (c) interference microscopy with (A) the Vicker's M86@ scanning microinterferometer and (B) the Jenoptik Amplival Interphako@ interference microscope and (d) the 're-embedding' method. Reliable, reproducible and comparable results were obtained with the small-fold technique, with the Vickers M86@ scanning microinterferometer and with the electron scattering method. For the last method, standard test lines for the different settings of the electron microscope were developed. The results obtained with the Jenoptik Amplival InterphakoB interference microscope are reproducible, but show a constant difference, i.e. a factor of 1-36, in thickness compared with the other three techniques. The possible cause of this 'systematic error' is discussed. The re-embedding method proved to be more laborious and slightly less reliable than the other techniques.The variation in t' between sections of a particular interference colour (inter-section variation) was found to be larger than the variation in t' within a section (intra-section variation).