Summary Radiation damage in the electron microscope is reviewed in the context of biological X-ray microanalysis. A phenomenological description of the dependency of radiation induced mass loss on electron dose in frozen–hydrated and frozen–dried biological material is given and some semiquantitative data are presented as a rule of thumb for practical X-ray microanalysis. Quantitative imaging, spatial deconvolution and extrapolation to zero dose are discussed as possibilities to minimise radiation effects in low temperature X-ray microanalysis. The consequences for the measurement of local water fractions by low temperature electron microscopy and X-ray microanalysis are deduced. Introduction X-ray microanalysis can be a very powerful tool in physiological research, because of its ability to track with high resolution the distribution of ions in cells and tissues. This requires low temperature techniques for three reasons. First, there is a beneficial effect of low temperatures in terms of radiation damage. Second, the retention of water during the measurement excludes a number of possible redistribution artefacts. Third, retention of water by cooling is required by most of the techniques to measure local water fractions. It was clear from the beginning of biological X-ray microanalysis that its potential could be fully exploited only when the distribution of water as well as the distribution of detectable elements could be measured (see, for instance, Gupta, Hall & Maddrell, 1976).