Blood consists of a fluid (plasma) in which are suspended various cells which make up about 55% of the total volume. Broadly speaking these cells are of five kinds: platelets, red cells (erythrocytes) and three kinds of white cells (granulocytes, lymphocytes and monocytes). The relative sizes and numbers of these cells are shown in Table 1. Strictly speaking platelets and erythrocytes are not cells, since they have no nuclei, but that distinction is of no consequence to this paper. The ability to label these separate cellular components of blood with radionuclides is of immense value in medical research, and subsequently in the diagnosis of disease and in the assessment of therapy, especially with those radionuclides which emit y-rays so that their movement can be followed, after reinjection, by external counting systems. For example, labelled platelets can be used to detect clots in the veins or arteries, and labelled granulocytes can be used to detect deep-seated abscesses. However, when we label cells in order to study their kinetics and bio-distribution in health or disease-we must be aware of the hazards that may lead the labelled cells to behave abnormally-or die. Unfortunately, with the possible exception of the labelling of monocytes by phagocytosis, all the methods of labelling cells so far discovered are “nonspecific”; i.e. in a mixed population of cells all will be labelled, and so to label a specific kind of cell, e.g. lymphocytes, these must first be separated from all the other cells in the blood sample. This separation procedure exposes the cells to the hazard of mechanical damage, though with care and practice this can be minimised. Labelling cannot be achieved without exposing the cells to chemicals which may be toxic and again may damage them. However, this problem too can usually be overcome with care. The third possible cause of damage to labelled cells arises from the radiation absorbed within the cells following the decay of the radioactive label. Of course, if each cell were labelled with only one radioactive atom then once that atom had decayed the cell would no longer be labelled, so any abnormality in its behavior would not be detected. However, in practice (as we shall see) each cell usually contains a very large number of radioactive atoms, and so may remain detectable long after radiation damage has led to erratic behaviour. When planning investigations which use labelled cells, we must therefore give thought not only to the radiation dose to the whole body or to some critical organ (as is customary for in vim radionuclide studies) but also to the radiation dose experienced by the labelled cell itself. So far as an individual labelled cell is concerned. once it is m-injected into the blood stream it will be virtually unaffected by radiation from other cells. but will be affected by short-range radiation originating within itself. In attempting to calculate this radiation dose, we can therefore ignore all the yand Xradiation which contributes so importantly to the total body or critical organ dose, and consider only that radiation which has a range shorter than or comparable to the cell diameter. In practice this means we need consider only the Auger electrons (or, occasionally, conversion electrons of very low energy) associated with the decay process. Whilst the treatment of this subject which follows is superficial compared to that being applied to