There is now firm evidence that phototherapy has the following effects on bilirubin metabolism in humans with unconjugated hyperbilirubinemia. It rather rapidly converts a substantial fraction of the normal toxic 4Z, 15Z form of bilirubin to the 4Z, 15E form, which probably is less toxic. Simultaneously it enhances the overall excretion of bilirubin by converting it to oxidation products and structural and configurational isomers that are excretable in bile and urine without the need for glucuronidation. We know that these reactions occur in vivo because we have synthesized the compounds involved and have identified them unambiguously in vivo in the tissues of jaundiced babies and rats undergoing phototherapy. It is unlikely that these photobiological effects on bilirubin metabolism and transport are restricted to babies undergoing purposeful phototherapy. All babies are exposed to visible light and all develop hyperbilirubinemia during early life, with many exhibiting jaundice. Because there is no lower intensity threshold for photochemical reactions, it seems probable that the photobiological effects described in this paper occur in most newborns to some degree. Furthermore, similar photoprocesses would be expected to occur in the approximately 2-5% of the population who have the benign condition known as Gilbert's syndrome, which is characterized by chronic mild unconjugated hyperbilirubinemia, particularly when they sunbathe. Clearly, in the particular instance of phototherapy of neonatal jaundice, blue light is therapeutic. In some respects it acts like a drug, almost like the ideal magic bullet, because it is specific for the target molecule and safe. The main limitation of phototherapy is that it is inefficient, a limitation that seems to be imposed by transport processes in the body and the optics of skin rather than by the photochemical reactions on which it depends.