At the relatively low dose levels encountered in diagnostic radiology, detrimental effects have not been observed. Based on a linear extrapolation from the available data at high dose, (in excess of 0.3 Gy) a small risk for the occurrence of stochastic effects can, however, not be excluded. In diagnostic radiology the exposure conditions are highly inhomogeneous. When only one type of tissue is exposed, e.g. in the case of mammography, it will be most adequate to quote an average dose to the organ of interest. In nuclear medicine the administered activity, considered by the ICRP as a reference value, can easily be converted into effective dose. For diagnostic radiology with external X ray beams, field parameters such as air kerma free-in-air or the dose-area product (DAP) can be considered as easily measurable quantities. The limitations of the latter parameter are demonstrated on the basis of measurements in paediatric radiology. The DAP increases with age of the children whereas the effective dose remains nearly constant. The absorbed dose at the skin of the patient is not an appropriate quantity for computed tomography. For other techniques this dose value should be complemented by information on beam quality and field dimensions. Although objections have been raised against the use of effective dose, this concept provides a useful means to classify diagnostic procedures in terms of radiation exposure of the patient.