Neonatal jaundice is a normal transitional phenomenon after birth, as the newborn mammalian fetus adapts to life outside the womb. Hyperbilirubinemia in the neonate may even have a physiological role in the defense against oxidative stressors, such as oxygen, light and microbe-induced inflammationFall unavoidable in life on our planet. However, some babies have excessive elevations of bilirubin in their blood, and if they exceed the bilirubin-binding capacity of the circulation, mainly determined by albumin, bilirubin can accumulate in other tissues. One tissue that seems to be especially vulnerable is the brain with discrete areas prone to bilirubin-induced injury, for example, the globus pallidus, hippocampus, internal capsule, thalamus and eighth cranial nerve (VIII). The explanation for these regional susceptibilities and the mechanism of the toxicity have so far eluded researchers, although the fact that bilirubin can be toxic under some conditions encountered in the newborn is indisputable. Intriguing hypotheses have been proposed (Brites et al., this issue on pp S8–S13) and converging lines of investigations may eventually yield answers that might inform preventive strategies. One of the most tantalizing issue, perhaps because of its paradoxical character, is that bilirubin, an antioxidant in the usual biological circumstance, can cause or exacerbate oxidation under pathological conditions, such as excessive accumulation in tissues, triggering the production of cytokines, which, depending upon the timing in development and other confounding inflammatory states, such as infection, can alter neuronogenesis, cell-to-cell interactions or even contribute to abnormal apoptosis. What is clear to clinicians is that ‘bad things can happen’ in association with hyperbilirubinemia, even in apparently healthy term infants, if the bilirubin level is high enough (Watchko et al., this issue on pp S14–S19). Recent data suggest that much lower levels can also be associated with harm, such as neurodevelopmental impairment, profound impairment, cerebral palsy and hearing deficits, in the most immature neonates. Nonetheless, a blood bilirubin level, in and of itself, is rarely sufficient to predict toxicity in the individual case or even in a population. Thus, scientists remain focused on trying to predict which infants are most likely to develop excessive bilirubin accumulation in tissues outside the circulation, a phenomenon dependent upon more than the absolute level per se, but also upon albumin-binding capacity and affinity, the unbound or ‘free’ bilirubin (Bf) as well as the transporters that can facilitate bilirubin’s movement out of the brain. Excessive bilirubin production, as in the case of hemolysis (from any cause), is the main driver of load on the system and can often overload it. It is not surprising that hemolysis has been associated empirically with an increased risk of bilirubin-induced injury. However, impaired conjugation such as that associated with Gilbert’s disease can further exacerbate the load, contributing to more Bf, capable of entering the brain tissue. The genetic polymorphisms contributing to overload situations (that is, HO-1 overexpression or uridine diphosphoglucuronate glucuronosyltransferase deficiency) only exacerbate natural exigencies, such as prematurity, infection or acquired causes of heme degradation. Although these topics are not the focus of this supplement, one of the articles does attempt to model using Gunn rat pups the central nervous system’s Bf, probably a critical number to know if we could in the clinical setting (Watchko et al., this issue on pp S14–S19). These data support the contention that it is, indeed, the Bf in the CNS that is toxic. Still, clinicians are left with the rough empirical correlation of toxicity with the blood bilirubin level (in the case of the model 35 mg per 100 ml or 599 mmol l ), and the need to avoid such high levels. Therefore, until such time as the scientists can provide clinicians with more targeted diagnostics and therapeutics, the focus must necessarily be on prevalent practices. Moreover, the supplement senses this purpose well, dealing with safer management of newborn jaundice to prevent severe neonatal hyperbilirubinemia (Bhutani and Johnson, this issue on pp S4–S7), including a report on the Pilot USA Kernicterus Registry (Bhutani et al., this issue on pp S25–S45), the answering of frequently asked questions (Bhutani and Johnson, this issue on pp S20–S24), the presentation of a six-step strategy to prevent severe neonatal hyperbilirubinemia (Bhutani and Johnson, this issue on pp S61–S67), a description of a system-based approach to newborn care that could reduce the risk for severe jaundice (Stark, this issue on pp S53–S57), an argument for glucose-6-phosphate dehydrogenase deficiency screening (Kaplan and Hammerman, this issue on pp S46–S52) and a demonstration project for hyperbilirubinemia management by the Hospital Corporation of America (HCA) (Lazarus and Avchen, this issue on pp S58–S60). Correspondence: Dr DK Stevenson, Department of Pediatrics, Stanford University School of Medicine, Stanford, CA 94305, USA. E-mail: dstevenson@stanford.edu Journal of Perinatology (2009) 29, S2–S3 r 2009 Nature Publishing Group All rights reserved. 0743-8346/09 $32