Cobalt metabolism and toxicology are summarized. The biological functions of cobalt are updated in the light of recent understanding of cobalt interference with the sensing in almost all animal cells of oxygen deficiency (hypoxia). Cobalt (Co(2+)) stabilizes the transcriptional activator hypoxia-inducible factor (HIF) and thus mimics hypoxia and stimulates erythropoietin (Epo) production, but probably also by the same mechanism induces a coordinated up-regulation of a number of adaptive responses to hypoxia, many with potential carcinogenic effects. This means on the other hand that cobalt (Co(2+)) also may have beneficial effects under conditions of tissue hypoxia, and possibly can represent an alternative to hypoxic preconditioning. Cobalt is acutely toxic in larger doses, and in mammalian in vitro test systems cobalt ions and cobalt metal are cytotoxic and induce apoptosis and at higher concentrations necrosis with inflammatory response. Cobalt metal and salts are also genotoxic, mainly caused by oxidative DNA damage by reactive oxygen species, perhaps combined with inhibition of DNA repair. Of note, the evidence for carcinogenicity of cobalt metal and cobalt sulfate is considered sufficient in experimental animals, but is as yet considered inadequate in humans. Interestingly, some of the toxic effects of cobalt (Co(2+)) have recently been proposed to be due to putative inhibition of Ca(2+) entry and Ca(2+)-signaling and competition with Ca(2+) for intracellular Ca(2+)-binding proteins. The tissue partitioning of cobalt (Co(2+)) and its time-dependence after administration of a single dose have been studied in man, but mainly in laboratory animals. Cobalt is accumulated primarily in liver, kidney, pancreas, and heart, with the relative content in skeleton and skeletal muscle increasing with time after cobalt administration. In man the renal excretion is initially rapid but decreasing over the first days, followed by a second, slow phase lasting several weeks, and with a significant long-term retention in tissues for several years. In serum cobalt (Co(2+)) binds to albumin, and the concentration of free, ionized Co(2+) is estimated at 5-12% of the total cobalt concentration. In human red cells the membrane transport pathway for cobalt (Co(2+)) uptake appears to be shared with calcium (Ca(2+)), but with the uptake being essentially irreversible as cobalt is effectively bound in the cytosol and is not itself extruded by the Ca-pump. It is tempting to speculate that this could perhaps also be the case in other animal cells. If this were actually the case, the tissue partitioning and biokinetics of cobalt in cells and tissues would be closely related to the uptake of calcium, with cobalt partitioning primarily into tissues with a high calcium turn-over, and with cobalt accumulation and retention in tissues with a slow turn-over of the cells. The occupational cobalt exposure, e.g. in cobalt processing plants and hard-metal industry is well known and has probably been somewhat reduced in more recent years due to improved work place hygiene. Of note, however, adverse reactions to heart and lung have recently been demonstrated following cobalt exposure near or slightly under the current occupational exposure limit. Over the last decades the use of cobalt-chromium hard-metal alloys in orthopedic joint replacements, in particular in metal-on-metal bearings in hip joint arthroplasty, has created an entirely new source of internal cobalt exposure. Corrosion and wear produce soluble metal ions and metal debris in the form of huge numbers of wear particles in nanometric size, with systemic dissemination through lymph and systemic vascular system. This may cause adverse local reactions in peri-prosthetic soft-tissues, and in addition systemic toxicity. Of note, the metal nanoparticles have been demonstrated to be clearly more toxic than larger, micrometer-sized particles, and this has made the concept of nanotoxicology a crucial, new discipline. As another new potential source of cobalt exposure, suspicion has been raised that cobalt salts may be misused by athletes as an attractive alternative to Epo doping for enhancing aerobic performance. The cobalt toxicity in vitro seems to reside mainly with ionized cobalt. It is tempting to speculate that ionized cobalt is also the primary toxic form for systemic toxicity in vivo. Under this assumption, the relevant parameter for risk assessment would be the time-averaged value for systemic cobalt ion exposure that from a theoretical point of view might be obtained by measuring the cobalt content in red cells, since their cobalt uptake reflects uptake only of free ionized cobalt (Co(2+)), and since the uptake during their 120 days life span is practically irreversible. This clearly calls for future clinical studies in exposed individuals with a systematic comparison of concurrent measurements of cobalt concentration in red cells and in serum.