[en] The problem of dose specification in external beam therapy with photons and electrons has been dealt with in ICRU Report 29 (1978). This problem arises from the fact that the absorbed dose distribution is usually not uniform in the target volume and that for the purpose of treatment reporting a nominal absorbed dose - which will be called target absorbed dose - has to be selected. When comparing the clinical results obtained between radiotherapy centres, the differences in the reported target absorbed doses which can be introduced by differences in the methods of dose specification often are much larger than the differences related to the dosimetric procedures themselves. This shows the importance of the problem. In this paper, some definitions of terms and concepts currently used in radiotherapy are first recalled: tumour volume, target volume, treatment volume, etc. These definitions have been proposed in ICRU Report 29 for photon and electron beams; they can be extended to any kind of irradiation. For external beam therapy with photons and electrons, the target absorbed dose is defined as the absorbed dose at selected point(s) (specification point(s)) having a meaningful relation to the target volume and/or the irradiation beams. Examples are discussed for typical cases. As far as interstitial and intracavitary therapy is concerned, the problem is more complex and no recommendations have so far been made by the ICRU Commission. A major difficulty arises from the sharp dose gradient as a function of the distance to the sources. The particular case of the treatment of cervix carcinoma is considered and some possible methods of specification are discussed: (1) the indication of the sources (in adequate units) and the duration of the application, (2) the absorbed doses at selected reference points (bladder, rectum, bony structures) and (3) the description of the tissue volume (height, width, thickness) encompassed by a given isodose surface (60Gy). (author)