With regard to the stability of the NO. radical as a chemical entity, it is without doubt able to serve as an intra- as well as an intercellular messenger. The radical O2-., in contrast, does not seem to be suited to far-range signalling in the vascular system. Its short chemical half-life, which is limited by the presence of various reactive blood constituents to below 50 ms, results in a free diffusion path length of less than 40 microns, i.e. only the distance between just a few cells. While accelerated 'downstream' transport by arterial blood may help to extend the action sphere, there is no possibility for O2-. to serve as a signal in an upstream direction. The estimates presented, however, do not invalidate arguments for a possible role of superoxide anions in intra- or pericellular signalling phenomena. Cross-talk between NO.- and O2-.-dependent signal routes, e.g. by peroxynitrite formation, is unlikely to be a relevant process under the conditions which prevail in the vascular system.