Most of the basic knowledge about radical reactions comes from radiation chemical studies in vitro. In view of the rapidly increasing knowledge on radical reaction in vivo, it is important to reconcile the fundamental physico-chemical reaction characteristics of radicals with the need to explain their alleged biological effects. Severe problems in the understanding of their in vivo action remain unsolved. An example is phagocytosis, which seems to be a paradigm of a 'deleterious' radical process. The exact mechanism is not clear; so it is an open question whether the intruder is eventually killed by radicals (like OH) or by endproducts of radical reactions (like H2O2 and/or HOCl). It is even more difficult to understand signalling by radicals: owing to their chemical nature they are 'unspecifically' reacting species--they withdraw or add electrons--and thus their reactions are governed by redox-properties. Since all radicals have different redox characteristics and different molecular shapes, the usual key-and-keyhole picture for molecular interaction does not apply, as there, is no reactive site conceivable which has the property of reacting with radicals 'specifically. Our intent in this article is: (i) to briefly review some fundamental characteristics of in vitro radical reactions, (ii) to extrapolate from this to the conditions in vivo, and (iii) to discuss current hypotheses concerning the redox-regulation of cellular signalling. This leads us to the tentative conclusion that radicals per se must be tolerated by the cell and do not threaten its life, if they stay below a certain concentration limit. The main biological implication of radical-reactions seems to be that the cell derives signals from the balance of oxidative versus reductive processes and that radicals may interact with pathways of intra- and intercellular communication.